This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
1. In how many accidents have you been involved when you were the driver? 2. I often feel fear when driving. strongly disagree < INPUT TYPE=radio NAME=05v2 VALUE=1> < INPUT TYPE=radio NAME=05v2 VALUE=4> < INPUT TYPE=radio NAME=05v2 VALUE=5> strongly agree 3. Are you Male or Female? < BLOCKQUOTE > param('00exp'); open(INFO, "»$path to_datafile/$filename.data"); foreach $key (sort ($query->param)) { $value = $query->param($key);
#filter out "'s and ,'s $value =~ s/\"/\'/g; $value =~ s/,/ /g; if ($value !~ /^pf/) { print INFO "\"$value\" , "; } else {
# filter out items that need to be expanded at submission time pf* if ($value =~ /ApfDate/) {
print INFO "\"$mon/$mday/$year\", "; } if ($value =~ /^pfTime/) {
print INFO "\"$hour : $min: $sec\" , " ; } if ($value =~ /Apf Remote/) {
print INFO "\" " , $query->remote addr(),"\", " ; } if ($value =~
/^pfReferer/)
{
print INFO "\"",$query->referer() , " \ " , " ; } } #print "$key : $value"; } print INFO "\"complete\"\n"; close (INFO);
print $query->redirect ($redirect_to); exit( );
to install another script. The data will arrive in the designated folder on your server as comma separated values (CSV) files that can be easily imported to many statistical and spreadsheet applications. Methods for working with and analyzing data collected in this format are described in detail in Birnbaum (2001a).
HOW TO ANALYZE LOG FILES Scientific investigation relies on the principle of raw data preservation. Raw data need to be saved for scrutiny by other
26. Behavioral Research and Data Collection researchers from the community (American Psychological Association, 2001), for example to aid in reanalysis or meta-analysis. This principle applies to Internet-based research as well, where it can be argued that server log files, properly configured, are the raw data (Reips, 2001a; 2001b, 2002b; Reips & Stieger, 2004). In addition to all data sets from full participation, complete log files from a Web server used in an Internet-based investigation contain the following useful information: 1. Data about potential participants who decide not to participate (e.g., the number of people who see the link to a study, but do not click it) 2. Data about technical conditions during the investigation (i.e., the general Web traffic conditions at the server and the particular conditions of each request by a participant) 3. Data sets for those who provided incomplete information; partial nonresponses may reveal information about user type and potential problems (Bosnjak, 2001; Reips, 2002b) Not reporting the information listed above, for example because one did not save it, carries the great danger of misleading scientists with respect to the main effects of variables (Birnbaum & Mellers, 1989; Reips, 2000, 2002c; Reips & Stieger, 2004). Even when dropout rates are equal in two groups, the observed trend may easily show the opposite of the true effect that one would have found had all data been complete. Because participants find it quite easy to quit Web studies, there can be sizeable attrition in such studies, and this attrition needs to be documented and reported. For more on drop-out analysis, see Frick, Bachtiger, and Reips (2001), Reips (2002b, 2002d), Reips and Stieger (2004), and Knapp and Heidingsfelder (2001). Log files need to be configured so that the information needed will be saved. Often the log file format follows a common predefined format, for example the Webstar log file format. This log file format is used by the Web Experimental Psychology Lab and contains information in the following order: CONNECTIONJD DATE TIME RESULT HOSTNAME URL BYTES_SENT AGENT REFERER TRANSFER.TIME SEARCH_ARGS. A platform-independent interactive Web site that helps Internet researchers in analyzing log files is Scientific LogAnalyzer (Reips & Stieger, 2004, http://genpsylab-logcrunsh.unizh.ch/). It was created to meet the needs of those who collect data on the Internet. Scientific LogAnalyzer provides an option for selecting the predefined format mentioned above, and it also contains procedures to identify and process any type of log file. To match a predefined format, the user may also rearrange columns in the log file before uploading it to Scientific LogAnalyzer, which can easily be done in text editors and spreadsheet programs. Scientific LogAnalyzer has features important to behavioral and social scientists, such as handling of factorial designs, response time analysis, and dropout analysis. Scientific LogAnalyzer was developed to include calculation of response times, flagging of potential multiple submissions, selecting either first or last response from the same IP, marking of predefined IP addresses or domain names, and free definition of session timeout). The program is highly flexible on the input side (unlimited types of log file formats), while strictly keeping the traditional one-case-per-row output format. Other features include
•
485
free definition of log file format; searching and identifying any combination of strings (necessary for organizing conditions in experiment data); computation of approximate response times; a module for analyzing and visualizing dropout; detection of multiple submissions; output in HTML and tab-delimited files, suited for import into statistics software; speedy analysis of large log files; and extensive help from an online manual.
PRETESTING OF INTERNET-BASED RESEARCH Before an experiment is placed on the Web, it is necessary to perform a number of checks to make sure that the study will yield useful data. First, one should be clear on how the data will be analyzed and that the study will answer the question it is supposed to answer. This check is basic and applies to lab research as well as Web research. Those who are unclear on how the data will be analyzed almost never devise a study that can be analyzed. Second, one should conduct checks of the HTML and CGI script to code and save data to ensure that every possible response is properly coded and recorded in the proper place in the data file. One should check that every radio button (in a given item) functions properly and that answers to one question do not overwrite responses to another item. This is one of the advantages of using a Web service like FactorWiz (Birnbaum, 2000b) or WEXTOR (Reips & Neuhaus, 2002); these programs save time by automatically creating safe code. Third, one should test some participants in the lab. Observe them as they read and respond to the materials. Ask them to identify aspects of instructions that are unclear. Internet-based research does not have a lab assistant who can answer questions, so every question must be addressed in advance. Check if people are responding before they have scrolled to see the information that they are supposed to review before responding. One of the first author's students had placed a response device before the material the participant needed to read. During pilot testing, the first author observed a number of people who responded before they had scrolled to make visible what they were supposed to be judging. Analyze the data from the pilot study to see that the coding and analysis will function properly. It is often when analyzing data that students discover problems with their studies. That is why some pilot data should be analyzed before the main study is run. One can discover a lot by observing participants in pilot research. For example, in a study with the randomized response technique (e.g. Musch, Broder, & Klauer, 2001), participants were supposed to toss two coins and then respond "yes" if both coins were heads, "no" if both coins were tails, and to tell the truth otherwise. The purpose of the technique is to allow an experimenter to assess a population statistic without knowing any person's true answer. For example, most people would be embarrassed to admit that they cheated on their income taxes, but with the randomized response technique there is no way to know if "yes" meant that the person did cheat or that the coins were both heads. If people foliow instructions, this method
486
BIRNBAUM AND REIPS
allows the experimenter to subtract 25% "yes" answers (that occurred because of two heads) and 25% "no" (resulting from two tails), and double the remainder to find the correct proportions. For example, if 30% of the group say that they cheated on their taxes, it means that 10% of the population say they cheated. In our pilot test, however, only 1 of 15 participants took out any coins, and she asked first if she should actually follow the instructions. The first author urged his student to add stronger instructions and an extra pair of items at the end of the survey asking if the participant had actually used the coins, and if not, why not. About half said they had not followed the instructions, giving excuses such as "lazy" and "I had nothing to hide." This example should make one very concerned about what happens in studies that are not observed and still more concerned about studies that are launched without pilot testing. It is also important to pretest the materials with different browsers and systems and with a small monitor, to make sure that everyone will see what they are supposed to see. Otherwise you would probably run into what the second author coined "configuration error IV" in Internet-based research—the all-toocommon underestimation of the technical variance inherent in the Internet (Reips, 2002b). Consider adding items to your research instrument to ask about monitor sizes, volume settings, and other settings when you think these variables might make a difference to the results. Such information can be used to partition the sample for separate analysis. Considerations of delivery of psychophysical stimuli and a discussion of when such settings may or may not matter are reviewed by Krantz (2001).
RECRUITING PARTICIPANTS FOR INTERNET-BASED RESEARCH Participants can be recruited by traditional means, such as a course assignments, student participant pools, face-to-face requests, word of mouth, posters, flyers, newspaper advertisements, and so on. We will focus here on Internet methods such as recruitment via Web site, mailing list, online panel, newsgroup, e-mail, listings, and banner ads. Recruitment of Internetbased studies can be made much more effective by using one or several of the techniques that are described by Birnbaum (2001a) and by Reips (2000, 2002a, 2002b, 2002d).
Recruitment via Web Site and Search Engines One natural way to recruit participants is via one's own homepage. However, many personal home pages are rarely visited. Besides, visitors of your homepage may know too much about your research (or read about it on the page) to be suitable (i.e., naive) participants for certain studies. A useful strategy to test whether such self-selection may be biasing your results is the multiple site entry technique (Reips, 2000, 2002c) that will be described later. An institution's homepage, for example a university's, may be a better choice for recruiting large numbers of participants. In addition, an institution's homepage will often convey a more legitimate impression than a personal homepage. Indeed, it will not be easy in most cases to get agreement to announce your study unless a number of people have agreed that it is acceptable to do so. One of the best places for recruitment are institutionalized Web sites for Internet-based experimenting, such as the web experiment list, the Web Experimental Psychology Lab, and the Psychological Research on the Net list by Krantz (1998), who published some of the first Web experiments (Krantz, Ballard, & Scher, 1997; Welch & Krantz, 1996). Some of these Web sites are visited by thousands of potential participants every month (Reips, 2001a, 2001b), and some managers even provide you with a free check of your experiment, before linking it. A link on a Web experiment site may also serve an archiving function as an example for future studies and as a reference in publications. People coming to your study via one of these Web research sites are true volunteers who have already decided that they want to take part in one or more psychology studies and who chose your study from a list of ways to participate. So, the concerns one might have with respect to students who are participating only to fulfill an assignment are relieved with this source of participants. To recruit participants by means of search engines, you can enhance your pages to help people find your study. Suppose you wanted to recruit people interested in psychology. To help those who are looking on the Web for psychology, you can put that word in the title of your site, add it to the early text in your site, and add meta tags including psychology as a key word. Suppose you wanted to recruit people with rare characteristics, such as transvestites. You could include meta tags in your
TABLE 26.5. Use of Metatags to Recruit via Search Engine 221: Collect judgments in a between-subjects design. Psychological Methods, 4, 243249. Birnbaum, M. H. (1999b). Testing critical properties of decision making on the Internet. Psychological Science, 10, 399-407. Birnbaum, M. H. (Ed.). (2000a). Psychological experiments on the Internet. San Diego, CA: Academic Press. Birnbaum, M. H. (2000b). SurveyWiz and FactorWiz: JavaScript Web pages that make HTML forms for research on the Internet. Behavior Research Methods, Instruments, & Computers, 32, 339-346. Birnbaum, M. H. (2001 a). Introduction to behavioral research on the Internet. Upper Saddle River, NJ: Prentice Hall. Birnbaum, M. H. (2001b). A Web-based program of research on decision making. In U.-D. Reips & M. Bosnjak (Eds.), Dimensions of Internet science (pp. 23-55). Lengerich, Germany: Pabst Science. Birnbaum, M. H. (2002). Wahrscheinlichkeitslernen. In D. Janetzko, M. Hildebrand, & H. A. Meyer (Eds.), Das Experimentalpsychologische Praktikum im Labor und WWW [A practical course on psychological experimenting in the laboratory and WWW] (pp. 141-151). Gottingen, Germany: Hogrefe. Birnbaum, M. H. (2004a). Human research and data collection via the Internet. Annual Review of Psychology, 55, 803-832. Birnbaum, M. H. (2004b). Methodological and ethical issues in conducting social psychology research via the Internet. In C. Sansone, C. C. Morf, & A. T. Panter (Eds.), Handbook of methods in social psychology (pp. 359-382). Thousand Oaks, CA: Sage.
Birnbaum, M. H., & Mellers, B. A. (1989). Mediated models for the analysis of confounded variables and self-selected samples. Journal of Educational Statistics, 14, 146-158. Birnbaum, M. H., & Wakcher, S. V (2002). Web-based experiments controlled by JavaScript: An example from probability learning. Behavior Research Methods, Instruments, & Computers, 34, 189-199. Bosnjak, M. (2001). Participation in Non-Restricted Web-Surveys: A Typology and Explanatory Model for Item-Nonresponse. In U. -D. Reips & M. Bosnjak (Eds.), Dimensions of Internet-Science (pp. 193-207). Lengerich, Germany: Pabst Science. Buchanan, T. (2000). Potential of the Internet for personality research. In M. H. Birnbaum (Eds.), Psychological experiments on the Internet (pp. 121-140). San Diego, CA: Academic Press. Buchanan, T. (2001). Online personality assessment. In U.-D. Reips & M. Bosnjak (Eds.), Dimensions of Internet science (pp. 57-74). Lengerich, Germany: Pabst Science. Dillman, D. A., & Bowker, D. K. (2001). The Web questionnaire challenge to survey methodologists. In U.-D. Reips & M. Bosnjak (Eds.), Dimensions of Internet science (pp. 159-178). Lengerich, Germany: Pabst Science. Drake, P. J. W. (2001). Successful aging: Investment in genealogy as a function ofgenerativity, mobility and sense of place. Master of arts thesis, California State University, Fullerton. Eichstaedt, J. (2001). An inaccurate-timing filter for reaction-time measurement by JAVA-applets implementing Internet-based experiments. Behavior Research Methods, Instruments, & Computers, 33,179-186. Francis, G., Neath, I., & Surprenant, A. M. (2000). The cognitive psychology online laboratory. In M. H. Birnbaum (Eds.), Psychological experiments on the Internet (pp. 267-283). San Diego, CA: Academic Press. Frick, A., Bachtiger, M. T, & Reips, U.-D. (2001). Financial incentives, personal information, and drop-out in online studies. In U.-D. Reips & M. Bosnjak (Eds.), Dimensions of Internet science (pp. 209-219). Lengerich, Germany: Pabst Science. Goritz, A. (2004). Apache, MySQL, and PHP for Web surveys. Retrieved March 2, 2004, from www.goeritz.net/ati/download.htm Goritz, A., Reinhold, N., & Batinic, B. (2002). Online panels. In B. Batinic, U.-D. Reips, & M. Bosnjak (Eds.), Online social sciences (pp. 27-47). Gottingen: Hogrefe & Huber. Hardin, C., & Birnbaum, M. H. (1990). Malleability of "ratio" judgments of occupational prestige. American Journal of Psychology, 103, 1-20. Knapp, E, & Heidingsfelder, M. (2001). Drop-out analysis: Effects of the survey design. In U.-D. Reips & M. Bosnjak (Eds.), Dimensions of Internet science (pp. 221-230). Lengerich, Germany: Pabst Science.
26. Behavioral Research and Data Collection
Krantz, J. H. (1998). Psychological research on the Net. Retrieved March 2, 2004, from psych.hanover.edu/research/exponnet.html Krantz, J. H. (2001). Stimulus delivery on the Web: What can be presented when calibration isn't possible? In U.-D. Reips & M. Bosnjak (Eds.), Dimensions of Internet science (pp. 113-130). Lengerich, Germany: Pabst Science. Krantz, J. H., Ballard, J., & Scher, J. (1997). Comparing the results of laboratory and World-Wide Web samples on the determinants of female attractiveness. Behavior Research Methods, Instruments, & Computers, 29, 264-269. Krantz, J. H., & Dalai, R. (2000). Validity of Web-based psychological research. In M. H. Birnbaum (Eds.), Psychological experiments on the Internet (pp. 35-60). San Diego, CA: Academic Press. McGraw, K. O., Tew, M. D., & Williams, J. E. (2000). PsychExps: An on-line psychology laboratory. In M. H. Birnbaum (Eds.), Psychological experiments on the Internet (pp. 219-233). San Diego, CA: Academic Press. Michalak, E. E. & Szabo, A. (1998). Guidelines for Internet research: An update. European Psychologist, 3, 70-75. Musch, J., Broder, A., & Klauer, K. C. (2001). Improving survey research on the World-Wide Web using the randomized response technique. In U.-D. Reips & M. Bosnjak (Eds.), Dimensions of Internet science (pp. 179-192). Lengerich, Germany: Pabst Science. Musch, J., & Klauer, K. C. (2002). Psychological experimenting on the World Wide Web: Investigating context effects in syllogistic reasoning, In B. Batinic, U.-D. Reips, & M. Bosnjak (Eds.), Online social sciences (pp. 181-212). Seattle: Hogrefe & Huber. Musch, J., & Reips, U.-D. (2000). A brief history of Web experimenting, In M. H. Birnbaum (Eds.), Psychological experiments on the Internet (pp. 61-87). San Diego, CA: Academic Press. Parducci, A. (1995). Happiness, pleasure, and judgment. Mahwah, NJ: Lawrence Erlbaum Associates. Reips, U.-D. (1995). The Web experiment method. Retrieved September 1, 1995, fromwww.uni-tuebmgen.de/uni/sii/Ulf/Lab/WWWExp Method.html Reips, U.-D. (1997). Das psychologische Experimentieren im Internet [Psychological experimenting on the Internet], hi B. Batinic (Eds.), Internet fur Psychologen (pp. 245-265). Gottingen, Germany: Hogrefe. Reips, U.-D. (1999). Theorie und Techniken des Web-Experimentierens [Theory and techniques of Web experimenting]. In B. Batinic, A. Werner, L. Graf, & W Bandilla (Eds.), Online Research: Methoden, Anwendungen und Ergebnisse (pp. 277-296). Gottingen, Germany: Hogrefe. Reips, U.-D. (2000). The Web experiment method: Advantages, disadvantages, and solutions, hi M. H. Birnbaum (Eds.), Psychological experiments on the Internet (pp. 89-117). San Diego, CA: Academic Press. Reips, U.-D. (2001a). Merging field and institution: Running a Web laboratory. In U.-D. Reips & M. Bosnjak (Eds.), Dimensions of Internet science (pp. 1-22). Lengerich, Germany: Pabst Science. Reips, U.-D. (2001b). The Web Experimental Psychology Lab: Five years of data collection on the Internet. Behavior Research Methods, Instruments, & Computers, 33, 201–211.
491
Reips, U.-D. (2002a). Context effects in Web surveys. In B. Batinic, U.-D. Reips, & M. Bosnjak (Eds.), Online social sciences (pp. 6979). Seattle: Hogrefe & Huber. Reips, U.-D. (2002b). Internet-based psychological experimenting: Five dos and five don'ts. Social Science Computer Review, 20, 241249. Reips, U.-D. (2002c). Standards for Internet-based experimenting. Experimental Psychology, 49, 243-256. Reips, U.-D. (2002d). Theory and techniques of conducting Web experiments. In B. Batinic, U.-D. Reips, & M. Bosnjak (Eds.), Online social sciences (pp. 219-249). Seattle: Hogrefe & Huber. Reips, U.-D., & Bosnjak, M. (Ed.). (2001). Dimensions of Internet science. Lengerich, Germany: Pabst Science. Reips, U.-D., & ranek, L. (2004). Mitarbeiterbefragungen per Internet oder Papier? Der Einfluss von Anonymitat, Freiwilligkeit und Alter auf das Antwortverhalten [Employee surveys via Internet or paper? The influence of anonymity, voluntariness and age on answering behavior]. Wirtschaftspsychologie, 6(1), 67-83. Reips, U.-D., & Neuhaus, C. (2002). WEXTOR: A Web-based tool for generating and visualizing experimental designs and procedures. Behavior Research Methods, Instruments, and Computers, 34, 234-240. Reips, U.-D., & Stieger, S. (2004). Scientific LogAnalyzer: A Web-based tool for analyses of server log files in psychological research. Behavior Research Methods, Instruments, & Computers, 36(2), 304-311. Schmidt, W. C. (1997a). World-Wide Web survey research: Benefits, potential problems, and solutions. Behavioral Research Methods, Instruments, & Computers, 29, 274-279. Schmidt, W. C. (1997b). World-Wide Web survey research made easy with WWW Survey Assistant. Behavior Research Methods, Instruments, & Computers, 29, 303-304. Schmidt, W. C. (2000). The server-side of psychology Web experiments. In M. H. Birnbaum (Eds.), Psychological experiments on the Internet (pp. 285-310). San Diego, CA: Academic Press. Schmidt, W C. (2003). Server and CGI program installation summary. Retrieved February 20, 2004, from http://ati.fullerton.edu/tfbilly Schmidt, W. C., Hoffman, R., &MacDonald, J. (1997). Operate your own World-Wide Web server. Behavior Research Methods, Instruments, & Computers, 29, 189-193. Schwarz, S., & Reips, U.-D. (2001). CGI versus JavaScript: A Web experiment on the reversed hindsight bias. In U.-D. Reips & M. Bosnjak (Eds.), Dimensions of Internet science (pp. 75-90). Lengerich, Germany: Pabst Science. Smith, M. A., & Leigh, B. (1997). Virtual subjects: Using the Internet as an alternative source of subjects and research environment. Behavior Research Methods, Instruments, & Computers, 29, 496-505. Tuten, T. L., Bosnjak, M., & Bandilla, W. (2000). Banner-advertised Web surveys. Marketing Research, 11(4), 17-21. Welch, N., & Krantz, J. H. (1996). The World-Wide Web as a medium for psychoacoustical demonstrations and experiments: Experience and results. Behavior Research Methods, Instruments, & Computers, 28, 192-196. Williams, J. E., McGraw, K. O., & Tew, M. D. (1999). Undergraduate labs and computers: The case for PsychExps. Behavior Research Methods, Instruments, and Computers, 31, 287-291.
This page intentionally left blank
Part
IX SPECIFIC WEB APPLICATIONS IN INDUSTRIAL SETTINGS
493
This page intentionally left blank
27 WEB-BASED PROGRAMS AND APPLICATIONS FOR BUSINESS Misha W. Vaughan and ]oseph S. Dumas Oracle Corporation
• Focused on getting some kind of work accomplished for a business. For example, some part of a job as an employee, a manager, or a chief executive officer (CEO) is dependent upon using this system. The focus is on Web-based software that is used to run or understand the business and how a company performs relative to its competitors. As another example, if a person works in a pharmaceutical company, a Web-based application that helps him or her keep track of and manage all of the clinical trials in progress is within the scope of this chapter. However, a Web site that provides consumers with information about ongoing clinical trials is not within the scope of this chapter. Other examples of relevant business applications are: • An application that every employee uses to access a company's benefits information • An application that tracks sales leads or marketing campaigns • An application that displays a scorecard of company key performance indicators • An application that a business analyst uses every month to track sales trends by region and product, across all the company executives • Transaction oriented. Users of a Web application will spend most of their time creating, updating, and monitoring objects in the system. This is different from many business-to-consumer Web sites (at least for now) that have a blend of content display with some lightweight transactions. • Presented in the context of a browser, and so may rely on hypertext markup language (HTML), XML, Java plug-ins, or other Web presentation technology for the user interface.
INTRODUCTION This chapter is a discussion of human factors issues for Webbased applications that are designed to run internally on a business intranet, or externally on an extranet, between businesses (e.g., business-to-business, or B-to-B). Other references provide a discussion of human factors for e-commerce applications, that is, Web-based applications designed for business-to-consumer (B-to-C) for example, Amazon.com, Yahoo.com, or consumerto-consumer (C-to-C), for example, Ebay.com. For example, see Najjar (chap. 28, this volume).
A TAXONOMY OF WEB-BASED APPLICATIONS FOR BUSINESS Web applications are an emergent phenomenon. What this means, practically, is that they are ill-defined. Some might define Web applications based on the technology used to implement the experience, for example, thick client, thin client, extensible markup language (XML), and so on. Others use the terms Web services, Web applications, and Web content interchangeably. A detailed review of the literature provided no consensus on what a Web application for business is. We have, therefore, decided to offer one here. We will propose a working definition of Web applications for business based upon their differences from other Web and desktop user experiences and interfaces. Keep in mind that this definition is limited to the current state of affairs and could well change over time. Specifically, Web applications for business are:
This definition is not intended to be exhaustive or comprehensive, but to provide a frame of reference for the problem space of this chapter. 495
496
VAUGHAN AND DUMAS
In addition to the above definition, we have found it helpful to clarify some terms that are often used when discussing Web applications: • Thin client/thick client refers to a continuum of processing capability at the user's machine in a client-server environment. With a thin client application, there are few data and little processing capability at the user's machine; with a thick client application, data and computing capability are at, or transferred to, the user's machine. A pure HTML application would be an example of a thin client. • Zero-footprint is a Web architecture not requiring any software to be installed at the user's machine. This term is often used as a synonym for thin client. • Java applets are dynamically downloaded Java code used to provide processing capability at the user's machine. Java applets make a thin client thick. Human factors of Web applications for business is also an emergent area of study. E-commerce has been the primary focus of human factors work for the Web. Because of the burst of attention that online retail sites attracted in the late 1990s, there was a spike of interest in the usability of these sites (Donoghue, 2002; Vergo, Noronha, Kramer, Lechner, & Coffino, 2002). Again, however, the focus of that activity was on users as consumers of the products of a business rather than on the people who run the business. A detailed review of the literature provided little additional substance that is directly relevant to business applications, and so most of our discussion will be based upon our own industry experience, as well as the in-depth knowledge of our industry peers. We will attempt to address those issues that are most important from a business user and a usability perspective and hence important for anyone actually designing these systems. In general, we see three main factors impacting the human factors of Web applications for business: • Factor 1: The size of a business or company intending to implement Web applications • Factor 2: The types of business application areas (e.g., type of work) intended to be supported • Factor 3: The types of users of these applications We will first present each of these factors as they relate to designing Web applications for businesses. We will follow this discussion with some ideas and suggestions for companies attempting to implement their own Web applications. Finally, we will present a more detailed discussion of particular human factors issues related to designing Web applications. Again, this discussion is intended to provide a user-centered framework from within which one can approach systems designed for the Web in a business context. Factor 1: Size of Business or Company. The scale* of small-(e.g., < 500 employees) to mid-market companies
(e.g., 500-1000 employees) compared to mid- to large-market (e.g., 1000 + employees) companies has a distinctly different impact on Web applications. For a small- to mid-market company, in general: • The size of the database any application is working against will be smaller. This smaller database size typically means faster performance (unless, of course, the company is running on a less expensive or less powerful server), as well as Web applications that will display smaller numbers of items, from employees to inventory (unless a particular company has an inventory with an extremely large number of parts). • The complexity of the business processes will also be lower; smaller companies simply do not have the same organizational complexity as larger companies (e.g., they are less likely to be globally distributed). Software designed for this context supports simpler businesses processes, for example, a simple approval process for purchases or budgets. • The total number of users will be small, for example, under 500 employees. This means that system response time will be less impacted by load than for a large organization. Again, this is true unless the company is running on a less expensive or less powerful server. • With smaller companies, the amount of money and time available to spend on training and education is lower and so software must be necessarily easier to learn. • The amount of money and time available to spend on a Web application implementation is also lower, which means that the backend technology must be simpler to uptake, for example rather than taking 3-9 months to install and configure, it needs to take 1-3 weeks. Along similar lines, a smaller company is less likely to have a dedicated professional to administer and maintain the application. Online services or hosted applications, which are Web applications delivered to a company via the Internet, have sprung up to serve this market. The software is not owned and installed on the buyer's hardware; rather, purchasers are buying the right to access and use a Web application for their company. Some examples of this type of service are Salesforce.com and Netledger (netledger. com). In the mid-market to large-market space, the above issues are also important for Web applications, but for different reasons: • In terms of the size of the databases, one is designing for systems that will possess much larger data sets. For example, a person might be designing a medical inventory system to handle 3 million products. The scope and scale of such large data sets bring unique design challenges. • In a larger organization, a person may be working across multiple databases rather than from a single database. This adds complexity to the backend and may slow performance, which then has to be handled gracefully in the user interface. • Another issue for larger organizations is that the business processes are going to be more complex and often multinational
"The categorizations for size of companies was taken from CIO magazine (cio.com) in March 2003.
27. Web-Based Programs and Applications for Business (e.g., multiple currencies and languages). For example, a purchasing or sales process may require several levels of approval and notification to be tracked and represented by a Web application. • In larger organizations, the total number of users on the system will be higher, by orders of magnitude. This has the ability to severely impact system response times and thus impact the user experience. In addition, a large company may add new divisions or reorganize its structure frequently, making its organizational hierarchy a moving target. • Also, a larger organization is often able to spend more money and time on training and education. Often these companies will provide a customized, well-structured, series of classes that are available on an ongoing basis. The impact here on Web application design is indirect in that developers building Web applications for this audience will assume, correctly or not, that every user will have access to extensive training and so will be less concerned about human factors issues. • A larger organization will have more money and time to spend on implementation of a Web application. They are likely (although not happy) to spend 3-9 months installing and configuring Web applications for their company. They will likely also have access to a technical staff to maintain and administer the application. Two general classes of Web application providers have emerged to service mid- to large-market companies, bestof-breed, and suite providers. Best-of-breed providers offertargeted solutions for a particular problem space (e.g., business intelligence reporting for Web applications). Suite providers offer entire ranges of products, from human resources to sales and marketing software, via a Web application. Examples of companies in this space are SAP, Peoplesoft, Oracle, and Siebel. Factor 2: Types of Applications. Software designed for business contexts could include most desktop applications, for example Microsoft Excel and PowerPoint and ACT! (contact management software). Web applications are designed to work in conjunction with some of these and in competition with others. The strength of Web applications for any sized organization is removing the need to install custom client software on each end user's machine. Users simply have a Web browser and an account; then they can use the software. The following discussion of different types of Web applications is based on one-on-one, in-person interviews conducted with user interface (UI) managers from Oracle. Specifically, we interviewed Kristin Thompson, manager of customer relationship management (CRM) UI; George Hackman, manager of enterprise resource planning (ERP) UI; Jeremy Ashley, manager of business intelligence (BI) UI; Richard Wright, manager of server technology and database UI; and Dan Workman, manager of analytic applications UI. The types of Web applications designed for business contexts currently include: • CRM • ERP • BI
497
• Portals - Web sites that aggregate access to various sources of content (see Emery, Jacko, Sainfort, & Moloney chap. 29, this volume) • Collaboration suites - Web applications that facilitate file sharing, cobrowsing of Web sites, chat, and so on. (see van Tilburg & Briggs, chap. 30, this volume) • Industry-specific applications, for example verticals, such as Pharmaceuticals, health care, and education What will become apparent from this brief discussion is the variety and complexity of tasks and users that business Web applications must support.
Customer Relationship Management CRM can be characterized as software designed to assist with the marketing, sales, and support of a product or service to a company's customers (K. Thompson, personal communication, November 11, 2002). It is often divided into three broad groups: marketing automation, sales automation, and support/service automation. • Marketing automation software supports campaign managers in the planning and implementation of marketing campaigns along multiple channels, for example, direct mail, phone solicitation, e-mail, and events. One human factors challenge of marketing automation software is that not all companies conduct marketing the same way. For example, most companies vary widely in terms of how they arrange the hierarchy of marketing objects that include campaigns, programs, promotions, and events. Trying to design a generic user experience for this domain is challenging. Another human factors issue related to marketing automation is tracking the communication and coordination that happens between different roles. There are many different people involved in putting together a marketing campaign or activity, and software that supports this process must make the sharing of information and approvals easy from a UI perspective. Another human factors challenge for this domain is providing value beyond simply managing the execution of a campaign, that is, assisting with strategic planning as well as analyzing what has been most effective for the next strategy. • Sales automation software supports sales representatives and managers in the tracking of leads, creating quotes and proposals, creating orders, forecasting, and contact management. One human factors challenge of designing for sales automation software is the behavior of salespeople themselves. Contact information is central to a salesperson's success or failure, and in general, salespeople are not inclined to share information much less enter it into an enterprise-wide application. Also, salespeople have highly individualized and established systems for how they handle and manage contacts. The challenge is to identify the value added by using automated systems for salespeople. Currently there is value added to sales management by allowing them to more easily track forecasted sales. However, the information a company wants to collect is often more than what salespeople need, so one has to
498
VAUGHAN AND DUMAS
convince salespeople that this will give them value. Another human factors issue with sales automation software is the mobility of salespeople. This means information and systems must be accessible via multiple devices, from laptops to personal digital assistants to paper. An additional issue related to sales software is the need to surface, or highlight, information from other areas that help salespeople make decisions about their customer relationships, for example, the health of the relationship, the progress of the deal, and the next action they need to take to close the deal. • Support/service automation software is used by customer service representatives and managers in troubleshooting customer problems and issues including repairs, returns, and technical support. One human factors challenge of designing support automation software, similar to a marketing automation challenge, is differences between industries. In the service and support world, industries vary widely in terms of how they provide support and service to their customers. For example, a high-tech software development company will have a different need for support than that of a traditional consumer goods company. Similar to marketing automation software, it is difficult to identify a generic or best practice model that will serve many different industries. Another human factors issue for support automation is that there are two very different types of roles within service. There are people who specialize in a given area versus people who span multiple areas of service. An additional issue is that the service process involves people from many different areas within a company. For example, if you have a problem with your car, there is the person who handles first call, the person who diagnoses the problem, the technician who helps solve the problem, the person who gives feedback to product development about defects, and so on. These people have widely varying levels of skill and experience. One of the biggest challenges to designing Web applications for CRM is the limitations of HTML and the thin client. The CRM user population is highly professional, and in some cases heads down (e.g., call center service and sales representatives); high speed system performance and keyboard-only commands and navigation are important, but difficult to achieve in HTML. Data representation of deep hierarchies and navigation is challenging on the Web because we do not have the levels of navigation available in the desktop (e.g., Windows) environment. Also, the general conception that users have of HTML being synonymous with simple competes with the need for these applications to support complex tasks and power users.
Enterprise Resource Planning ERP software typically encompasses automating and managing the internal operations of a company, for example, the personnel, the finances, the assets, and so on (G. Hackman, personal communication, November 19, 2002). It is most often thought of in three groups: financial management systems (FMS), human resource management systems (HRMS), and supply chain management (SCM) and manufacturing systems.
Financial management systems software is designed to assist with the accounting of the inflow and outflow of money and key assets in a company. Users range from financial analysts and accountants who manage the ledgers to employees and customers who need to check on the status of their expenses or bills. The biggest human factors challenge for FMS software is the need to balance between ease of learning and productivity. For example, an application designed to help employees submit their expenses must be easy for an infrequent traveler to learn as well as efficient for the salesperson who travels frequently. Productivity is a key requirement with the professional users, but ease of learning is critical for the rest of the user population. Another human factors issue is the global nature of FMS systems for many companies. Ideally a company would have one ERP system to manage their global enterprise. The financial applications need to operate as one system, while accommodating widely varying tax systems and accounting rules between countries. Human resources management systems assist human resource (HR) managers and professionals with setting policies within a company; helps line managers with typical human resource functions such as raises, terminations, transfers, and hires; and helps employees with their occasional HR needs such as signing up for benefits. One human factors challenge for HRMS software is the highly sensitive nature of HR policies. For example, entering the correct effective date for a termination is very important. Most HR functions require careful attention to this kind of detail. Another human factors issue, similar to FMS systems, is the global perspective. Ideally a company would have one HR directory that included employees all over the world. However, different countries have widely varying legislation about how much information one can distribute in a directory. In addition, large companies tend to reorganize frequently. The challenge is to design a solution that works for all countries, but is not so vague as to make it unusable. Supply chain management (SCM) and manufacturing software helps users work collaboratively with suppliers to come up with a new product design (e.g., a new design for a tape recorder) and assessing when suppliers can deliver the parts. SCM also includes sourcing, where companies work with suppliers to purchase all the materials they need. It is also procurement, where employees buy what they need, and transportation, which involves moving the materials from the supplier to the manufacturer to the customer. Manufacturing software supports production line efficiency (e.g., scheduling a manufacturing line) and quality (e.g., assessing variances in product quality). The biggest human factors challenge for designing SCM and manufacturing software is similar to other domains in the need to serve both casual users, as well as production and professional users. For example, an employee ordering pens will need a different user experience from a buyer for an automaker ordering tires. Another challenge is the need for visualization, particularly when designing planning activities, for example, supply chain planning. Users in this domain like to visualize the routes and the trade-offs, for example there might be 10 different options for shipping
27. Web-Based Programs and Applications for Business an item, and these need to be displayed and compared. Finally, users working on a manufacturing floor are different in that they often share a single machine or workstation. For the large market, many ERP applications are still in thick client modes due to the limitations of HTML, such as the need for drag-and-drop, easily maintained context, and keyboard type ahead. However, some Web applications do exist such as the casual user (also known as self-service user) products for FMS (e.g., time cards, expense reporting) and for HRMS (e.g., benefits, line manager functions).
Business Intelligence BI software is ultimately geared toward providing key decision makers in a company with the right information about their company at the right time (J. Ashley, personal communication, January 5, 2003). BI software combines data administration tools, analytic tools, and end user BI applications. The technical, or administration, tools allow for the extraction, transformation, and loading of tables from a database into summarized views and into certain formats (e.g., time formats, multiple currencies). These tools also support managing security and accounts. Users of these tools are a small set of highly technical database administrators; there might be between 4 and 10 business intelligence administrators in a company of 10,000+ employees. For this group of administrators, there is no need for HTML as there is no need for ubiquitous access to configuration tools. Analytic software supports both the everyday (i.e., routinized, scripted reporting) and ad hoc (i.e., on an as-needed basis) tabulating, viewing, querying, and calculating of information about a company's operations. The users are business analysts and business professionals. As with the administration tools, this is a technically sophisticated audience for whom HTML is often not sufficiently interactive. However, this group of analysts does create BI reports that are often displayed in HTML for end users. Finally, there are some Web-based BI applications that pull information from these tools for end users to view and manipulate. • BI reports are software products that display summary information from both CRM and ERP applications to support uppermanagement decision making, such as hiring and firing and buying and selling companies. One key human factors issue associated with BI reports is the accuracy of the data displayed. In order for vice presidents to be comfortable making decisions about information in a report, they must trust that they are viewing accurate information. There are three components to this: accuracy of the question, accuracy of the result set, and timeliness of the data or frequency of the data refresh. Accuracy of the question requires a UI design that supports asking the right questions in the right order; for example "Show me sales from California, over $10,000, by product line" will produce a different display of data than "Show me all sales over $10,000, by product line." Accuracy of the result set goes to appropriately matching a visual representation with the given type of results, for example, pie chart, bar graph. Timeliness of the data means
499
that a user interface has to make clear both when the data were last updated and how to update the data if needed. • BI applications are software products that support enterpriselevel decision making on a daily basis by showing a decision maker (a) what the information is, (b) how and why the information is moving a certain way, and (c) what can be done with this information. This is also known as corporate performance measurement, for example, scorecards, intelligence on inventory, transportation, human resources. Users of BI applications are often CEOs, executive vice presidents, vice presidents, and essentially anyone who is responsible for decision making about how a company is run. They typically want to know how much money their company has made to date and how much has been spent to date. This group provides some unique human factors issues that will be discussed later as types of users. In addition to the special needs of these users, BI applications share the same human factors issues as mentioned under BI reports. As mentioned above, the range of Web applications for business intelligence tends to be limited to reports and end user applications. In small- to mid-market companies Microsoft Word and Excel are often used for generating HTML reports. Mid- to large-market companies use Web reports and some BI applications.
Industry-Specific Applications Industry-specific, or vertical, software solutions typically grow from one of two directions, either (a) heavily customizing an existing, generic Web application, (e.g., creating a hospital-specific version of an accounting application) or (b) creating a Web application to solve some industry-specific problem (e.g., student registration for universities and colleges) (G. Hackman, personal communication, November 19, 2002). • Healthcare applications: One example of a vertical Web application domain is health care. Existing Web applications for financial services, human resources, and so on are often heavily customized to suit the health care industry. In addition, new Web applications are being built to assist their specific problems of managing patient data. There are a variety of human factors challenges in trying to create a data management system for patients, which include privacy issues, staff that are typically low on computer literacy, unions, and doctors who have established ways of doing their work. Similar to the challenges of designing for sales forces, the people who really want software automation are not the primary end users. Right now the health care provider organizations desire this automation the most, as it would save them money by removing paper handling and paper storage, avoiding double billing, and so on. • Educational applications: Another example of a vertical solution is educational Web applications. In this case, Web applications are designed to deal with problems specific to universities and colleges, for example, student systems designed to help universities and community colleges automate registration processes. The human factors issues associated with this domain include (a) trying to push the work of registration and course
500
•
VAUGHAN AND DUMAS
administration back to the students and the faculty, that is, moving more toward the self-service model; and (b) globalization and localization issues, for example, in U.S. universities students pick a track and a major but have flexibility in their programs, whereas, in the United Kingdom students pick a track and all the courses are laid out in advance. • Pharmaceutical applications: The pharmaceutical industry has experienced a number of important challenges that new systems are attempting to address. In this highly regulated environment, the demands for rapid data collection and reporting are increasing. The pressure continues to grow for faster drug trials and for rapid and accurate reporting of safety data. The introduction of Web-based applications has been an important part of responding to these challenges. A doctor or nurse who is virtually anywhere in the world can enter or review drug trial data using a Web-based application with any computer with Internet access. These data can be immediately captured in a secure database, reducing errors and providing access to business managers tracking the progress of the trial. Web-based medical terminology dictionaries can be easily accessed to maintain consistent descriptions of adverse medical events for safety reporting. Finally the drug trial submittal process for regulatory agencies is now totally electronic. Creating, editing, storing, and transferring documents over the Web in a secure environment is a top priority for pharmaceutical manufacturers, and, therefore, software manufacturers. Factor 3: Business User Profiles. For the purposes of discussing Web applications for business, we have articulated a set of business user profiles. The profiles focus on four types of users: system administrators, professional users, self-service users and senior executives. Not all of these users make use of Web applications, but their work intersects with Web applications. Much like our discussion regarding types of Web applications for businesses, these user profiles were drawn from our own knowledge as well as that of our peers.
System Administrators These are the most technically proficient users inside companies. Examples are database administrators, system administrators, and Web masters. Rarely are they actual users of Web applications as they are often small in number and require highly interactive applications. More often they are responsible for installing, configuring, or otherwise tailoring Web applications for businesses. Although these users may not have much business expertise, they play a key role in making products available for business end users.
knowledge for financial analysts. They will end up using a particular business application or set of applications with a high degree of frequency, such as hourly, daily, or 'weekly. They use a computer for most of their work day and the majority are considered power users, including their use of the Web. Given this high degree of exposure to business applications, they are often highly computer literate—in their specific set of software. Key characteristics of professional users include: • Are capable of using dense screens and pages • Are comfortable with drilling into hierarchies and progressive disclosure of information • Can handle a large amount of functionality • Can deal with complex, tabular displays • Are comfortable using power user features such as a list box allowing multiple selection by control key, keyboard accelerators, and mnemonics • Can deal with data dependencies. As the complexity of the application increases, it may be more likely that increased data dependencies exist. For example, when an item in a choice list is selected, a portion of the page redraws with dependent information. There may be successive levels or multiple dependencies per page. • Can use application and context switchers allowing them to move between applications, objects, or contexts quickly— depending on what level of the application hierarchy they are viewing. • Are familiar with master and detail concepts, which provide a high density UI with access to an object and its details or many objects and their respective details. • Can personalize frequently used data, information display, or layout to accommodate his or her individual needs.
Self-Service Business Users This group includes the full range of business users, but they use applications that require little specialized knowledge or skill. For example, even a database administrator may need to use a company human resources site to look up health benefits. Such a site is considered a self-service application that anyone in the company should be able to use. In general, these Web applications are designed to accommodate users with only a minimum familiarity with the Web and a minimum of familiarity with the particular business process they are working on.
Senior Executives Professional Users These are the users who have received the most attention in business applications and have been the primary target for new applications. Professional users share a common experience of being highly skilled in a particular business domain, for example, sales knowledge for sales representatives and accounting
These are often the end users of the information generated by administrators and professional users. In the past, executives were the least computer literate of the business user types, but that profile is changing (A. Rich, personal communication, December 20, 2003). More and more executives use their computers daily and want hands-on access to business data. They want to
27. Web-Based Programs and Applications for Business be able to check the accuracy of data, research abnormalities in the data themselves, and send electronic messages rather than use the phone to resolve issues with subordinates. They also want to drill down to greater levels of detail and create graphs to summarize trends and to do their own what-if and sensitivity analyses. Although there are still executives who are hands-off decision makers, senior executives are transitioning passive recipients of business reports to active analysts of high-level trends. They are extremely sensitive to usability issues—they do not tolerate Web applications that are difficult to use.
HOW BUSINESSES CAN DEVELOP WEB-BASED APPLICATIONS Challenges for UI Design. When designing Web applications for businesses, one is often asked to do so in the context of other existing Web applications. Furthermore, the market for medium to large businesses is moving toward an integrated solution to all of a business's computing needs. For example, a number of vendors all build Web applications intended to be used as a suite or family of applications, where end users at a customer location use multiple applications from one company. If the user interfaces share a common look and feel, it means the end users in this situation can transfer their knowledge of the user interface between applications, which increases their satisfaction with the experience, reduces their learning time with new applications, and increases their productivity. From a business standpoint, this reduces costs and time associated with training as well as help desk support. A consistent user experience is achieved by developing a common look and feel across applications. A common look and feel is more than just a uniform color palette; it includes such things as art direction standards, interaction designs for common tasks, standard UI widgets, and rules for page layout. Since the introduction of suites of office productivity tools in the early 1990s, there has been interest in making the user interfaces to applications created by the same company look and feel the same. While interest in this goal remains high, achieving consistency has proven to be difficult. In concept, the objective is simple: allow users to take advantage of past learning by making new products work the same (Stewart & Travis, 2002). But user interface designers in companies with many product offerings face several obstacles to achieving a consistent look and feel: • There is always a legacy of existing applications: At any point in time, there are products in various stages of development and deployment. Given limited resources, only some of these applications can be rewritten to conform to a company style guide at any point in time. Typically, applications that are scheduled to be replaced or to be de-supported are the first on the list of applications that will not be converted to a new standard. In addition there may be applications aimed at small or very specialized markets that it does not make sense to convert. As a consequence, not all of the applications will conform to a standard. • Platforms and technologies evolve: Style guides that work for one type of technology often cannot be applied to a new one
501
without substantial modification. For example, separate style guides may be needed for use on desktop and client server platforms. For Web applications, a guide that works for a heavy or smart technology, such as Java, is a force fit for an HTML-based technology. Consequently, a company may be forced to support more than one user interface style guide. • Creating a company style guide takes a dedicated staff: Creating a company style guide is a major undertaking. The size of the effort depends on the scope of the guide. The style guide may apply only to the top-level menu or tab structure and the visual look at that level. But the quality of usability is usually in the details. Consequently, lower level items such as the size, placement, and labels on buttons may be just as important to usability as higher level components. Stopping at too high a level may limit a style guide's effectiveness. But the labor required to create a guide is always substantial and those resources cannot be used for other human factors activities. The most precious resource in a corporate user interface department is the time of its senior UI designers. Creating a style guide uses up some of those resources. • Maintaining and updating the style guide takes constant attention: As a style guide gets implemented, there is a constant stream of decisions about its scope. Does a style guide rule apply or not apply to a new situation? Development teams are constantly finding exceptions to the rules. Furthermore, the implications of rules often are not clear when they are created. A rule that works well for one task situation may not fit another that, at first, looks similar. Someone, or a team of style guide specialists, must make these decisions about scope and adjust the guide to new situations. A style guide is a living document that is never finished. • Maintaining consistency may mean ignoring better solutions: Sometimes maintaining consistency by enforcing a style guide requires not implementing exceptions that are more usable. The fact is that consistency is a mixed blessing. It allows transfer of training, which is an advantage, but it can stifle creative solutions that are not consistent with the guide. An organization that decides to standardize look and feel has to be willing to live with some of this inefficiency to maintain the standard. We will discuss some examples of companies and organizations that have developed Web application guidelines. Guidelines and Style Guides for Web Applications. The human factors profession has a long history of publishing guidelines of good practice. Perhaps the most well-known early effort was the series of guidelines for human-computer interaction (HCI) published by Sid Smith of the Mitre corporation in the mid-1980s (Smith & Mosier, 1984). That tradition continues with guidelines for Web usability. The purpose of Web usability guidelines is to convey good practice to Web designers from experienced human factors professionals. In general, there are two kinds of guidelines: 1. Design rules, which describe detailed practices, such as "Place tabs that are used for links at the top of the page and ensure that they look like clickable, real-world tabs." 2. Design principles, which provide general guidance that must be tailored to each design case, such as "Clearly articulate the
502
VAUGHAN AND DUMAS
primary goals of the Web site before beginning the design process." Guidelines are most useful when they: • • • •
Communicate established human factors principles Are based on research and backed with usability testing Are specific enough to avoid misinterpretation Are accompanied by visual examples of their intention and application
But general guidelines have their limitations, especially when they are: • • • •
Too vague to apply to a specific design problem Too specific to generalize across applications Conflicting Not based on any empirical data
Examples of confusing and vague guidelines are: • Clutter is preferable to clicking • Do the same as everybody else: If most big Web sites do something in a certain way, then follow along since users will expect things to work the same on your site. (http://www. useit.com/alertbox/991003.html) Furthermore, very few guidelines are focused specifically on business applications for the Web, especially applications that require analysis and visualization of complex data. Guidelines that address business applications still focus on the late-1990s phenomenon of the e-commerce Web sites. There are myriad guidelines in many places: books and journal articles, conference proceedings, professional society sites, and Web pages devoted to Web guidelines. Some representative sources can be found in the Appendix. Also, see Mariage, Vanderdonckt, & Pribeanu (chap. 38, this volume) for an overview of standards and guidelines. As a side note, the Web application guidelines designed by Oracle attempt to bridge the too specific-too vague gap by providing intermediate levels of detail in the form of page templates, page flows, and interface models and patterns (Beier & Vaughan, 2003). An interesting variation that combines the use of guidelines with usability evaluation is an automated usability evaluation. With this technique, a Web page or entire site is compared by a computer program with a list of human factors usability guidelines. WebSAT, the Web Static Analyzer Tool, was created at the National Institute of Standards and Technology (NIST) (Scholtz, Laskowski, & Downey, 1998). It compares a Web page or site, including an application, to two different sets of usability guidelines. WebSAT is available at http://zing.ncsl.nist.gov/WebTools. The Value of Usability Engineering Resources. A company intent on building Web applications for business would do well to make use of the existing usability engineering talent in the software community. Two general approaches to tapping into this talent pool are to develop an internal usability
organization and to engage an outside human factors organization that can assist with the design and evaluation of Web applications. The benefits of an internal organization can include developing a reservoir of knowledge over time of a company's design problems and solutions, developing political clout to help move the company in one design direction, and developing long-term relationships with product teams to educate them on good design. The benefits of an external organization can include their expertise as applied to multiple domains (potentially even an organization's competitors), their availability as needed such as when investment in an internal group is not feasible or when current resources are limited, and their access to a range of experts in usability and unique domains.
Internal Groups For companies interested in starting their own internal usability organizations to design Web applications for businesses, we have provided some profiles that offer details about other organizations' efforts. The following profiles are based on oneon-one phone interviews conducted with representatives from different usability organizations. Specifically, we interviewed Manik Singh and Christian Pantel from Peoplesoft, Janice Rohn of Siebel, Daniel Rosenberg of Oracle, and Keith Elliot of SAP Labs. For a broader history of usability engineering groups in the corporate world (e.g., Kodak, Thompson Consumer Electronics, Hewlett-Packard, Microsoft, American Airlines), see Wiklund (1994). A common theme across these profiles is that, whereas the design challenges with the Web are new, usability engineering processes (e.g., user requirements, field studies, surveys, and usability evaluations) and roles (interaction designers, usability engineers, prototypers, lab support staff) are generally the same. Across all of these organizations, two key components to success were (a) recruiting the right personnel and (b) developing a set of guidelines for their Web applications. One new advantage of the Web from a usability engineering standpoint is the ability to do usability testing with remote participants. Peoplesoft Background. The User Experience group at Peoplesoft is headquartered in Pleasanton, California (C. Pantel, personal communication, January 7, 2003). Peoplesoft has always had a user experience focus. Even when there were only 10 people in the company, there was 1 person who had user experience as an interest area. In 1996, Peoplesoft formally created a small usability team of about four members and built two usability labs. One lab is a standard-sized usability lab with a testing room and participant room. The other lab is conference-room sized and can seat up to 20 people, so that it is suitable for both usability testing and focus groups. They added a portable lab system in 2002. In 1999, they added more personnel with training in human factors and interaction design and have since grown to 31 people. Peoplesoft has over 150 products including CRM, financials, supply chain, human capital management, and People Tools (a development tool/core technology). Their applications are
27. Web-Based Programs and Applications for Business 100% Web-based now, except for their development tools. Part of their success is attributed to the upper-management support from the executive vice president of products and technology. Web Application Guidelines. One of the biggest goals of the Peoplesoft User Experience group has been to establish consistency. To do this, they have developed a set of Web application guidelines that are available internally. The guidelines are occasionally shared with customers who have need of them. All of the HTML UIs are generated out of a central engine; this enforces consistency, as well as produces challenges. For example, as discussed earlier, teams that are innovative or facing competitive pressures will push, or violate, the guidelines. Human Factors Challenges of Web Applications. In Peoplesoft's experience, the unique human factors challenges of Web applications focus on the HTML technology itself. Pure HTML technology imposes design constraints, such as the limited set of controls, or UI widgets, that are available in HTML. There are also design constraints around the client being disconnected from the server. For example, consider a Web application page displaying the total number of units to buy. If a person wishes to recompute that total after a change, the application has to make a trip back to the server to refresh the total. That means that the application has to communicate to the users that the total may not be the freshest data and they may want to update it. Another issue has been that HTML was originally created for document browsing, whereas Peoplesoft is trying to do business transactions. On the Web, every HTML page that downloads to a Web browser is an independent document, and it may look similar to the prior one a user had, but in reality it is a brand new document. The challenge for Web applications is that one needs to design—for the user—as if all the interaction is occurring in one place or document. One new challenge Peoplesoft sees has been the change in user profiles. Prior to the Web, when Peoplesoft still used a Windows client (in the mid-1990s), they had a much more limited user base. Most of their products were used by experts, for example, people who received training on the products and were using the products all day long. As they have moved to the Web, their customers have been able to deploy their products more widely and there are potentially many more end users of the products, with a broader range of skills. For example, they have many more self-service users, that is, average employees in a company who enroll in benefits once a year, or managers who give employees salary increases and manage vacation time. These people typically do not have any training at all. There is also a middle ground of users who use an application as a core part of their job but do not work every day with the application. For example, salespeople using sales force automation software are much more than casual self-service users, but they are not heads-down expert users. Siebel Background. The User Experience Department at Siebel was created in September 1999 and is located at
503
headquarters in San Mateo, California (J. Rohn, personal communication, December 16, 2002). The department has one usability lab at this location. At the peak of the organization, it employed 26 personnel. Siebel currently has over 200 products with both horizontal and vertical product lines. Horizontal products include CRM (e.g., sales, marketing, service, call centers), partner relationship management (PRM), customer-facing products (e.g., online catalogs, online services), and employee relationship management products (e.g., employee intranet content and applications). Verticals are customized versions of the horizontal products for a specific industry, such as energy and consumer goods. Almost 100% of their products are Web applications, except for their development tools products. Web Application Guidelines and Templates. Siebel's User Experience Department has driven a common user experience by developing a set of style guides for both customerfacing and professional Web applications. What proved to be a particularly effective strategy by Siebel's User Experience Department was the use of Web templates. The UE group designed and owns the page, view, and pagelet templates, which are HTML code, that define and constrain the way pages can be built. Internal Siebel developers, Siebel partners, and customers use these templates created by the department. Human Factors Challenges of Web Applications. Based on Siebel's experience, the biggest human factors challenge of Web applications is that the Web and browsers were not created for highly interactive applications; they were created for text. Also, some of the perceived benefits, such as low cost of administration and run-anywhere, turned out to be more complex. Depending on a customer's operating system and browser version, one can build different levels of interactive applications. To solve these problems, they built a core application architecture that would support interactive applications. As Rohn stated, "It was like designing the building where you don't yet know what the materials are" (J. Rohn, personal communication, December 16, 2002). To design in this context, they had to iterate on the design in order to assess what was technically feasible, within a given time frame, and still provide sufficient interactivity. From Siebel's perspective, a benefit of designing for Web applications, compared to Web sites, is that there is often a much more defined user profile.
Oracle Background. The Usability and Interface Design group at Oracle was begun in 1994 and has grown since to about 60 people (D. Rosenberg, personal communication, January 8, 2003). They have labs at their headquarters in Redwood Shores, California, with additional labs in Reading, United Kingdom, and Burlington, Massachusetts. At their headquarters, they have four traditional usability labs, one research lab, one participatory design room, one Internet living room, and one Internet classroom. Oracle currently has over 300 products that include CRM (e.g., sales, marketing, service, contracts), ERP (e.g., financials, manufacturing, human resources, projects), server
504
VAUGHAN AND DUMAS
technologies, databases, business intelligence (e.g., analytics, reporting, planning and simulation), and some vertical applications in health care, educational systems, and Pharmaceuticals. Approximately 70% of their products are Web applications. Web Application Guidelines. Oracle, similar to Peoplesoft and Siebel, had a mission to create consistency by creating a set of Web application user interface guidelines. These guidelines are publicly available on the Web (http://technet.oracle.com/tech/blaf/). When developing the guidelines Oracle found the biggest challenge was not being able to directly map the graphical user interface (GUI) to HTML. For example, not having an indigenous tree control in HTML was problematic. As Rosenberg stated, "It was like having a very limited set of colors in your coloring set, going from 600 colors down to 40" (D. Rosenberg, personal communication, January 8, 2003).
Human Factors Challenges of Web Applications. There are a number of human factors challenges, based on Oracle's experience, when designing and building Web applications. First is the lack of a selection model, that is, the difficulty of selecting an item in an HTML Web page. In traditional GUI applications, the design is based on an object-oriented model of design, or task action grammar, for a finite set of actions applied to a large range of objects. That model falls apart on the Web. For example, due to the difficulty of selecting, UI designers have to use work-arounds such as check boxes and radio buttons in cells of tables. Also, it becomes more difficult to deal with actions that are specific to an object versus actions that are shared by a group of objects. Designing this solution takes up valuable screen real estate in a Web application (e.g., with check boxes, icons, and button bars). In a GUI application, users would simply select all the objects and use a pull-down menu that would indicate whether the actions were available for every selected object. In the GUI world, users did not have to articulate the scope of an action in advance so precisely. The design solutions produce an inherent loss of information density and loss of efficiency. Another design challenge in the Web environment is the inability to validate data on a page and the inability to provide real-time feedback. In traditional GUI applications users are provided with field-level validation; users cannot really proceed in error. Compare that with a Web application in which users can do a lot of work, submit the page, and then find out their page is riddled with errors. Also, the back button becomes the enemy of data validation. In terms of real-time feedback, it becomes difficult to support certain work tasks with HTML, such as a call-center agent who needs to manipulate objects in the UI quickly. If applications use any dynamic HTML to drag-and-drop objects, the actual cost of changing the cursor as the user moves over drop targets is extremely expensive because validation of the drop target is done at the server. One of the design responses to this is that Web applications have become more task oriented. This works fine for smaller, less complex applications, but becomes challenging for larger, more complex ones. The case of more
complex applications, such as handling project planning, is still difficult. The tasks are more indeterminate and open-ended versus step-by-step and process-oriented, for example, creating a spreadsheet versus filling out a Web shopping cart. Another challenge is the set of expectations users currently have for the Web, such as e-commerce or information sites, versus a Web application. For example, the notion of a home page is completely different for Web applications. A Web application homepage is not about branding, image, and what is available to purchase on the site. It is about telling users what their last transaction was and its current state. Another example is users' preconceived notions about when things are saved in the GUI/desktop world versus saved in a Web application with a database. SAP Labs Background. SAP Labs, composed of about 10 full-time employees, is the research and development group for SAP AG (M. Singh & K. Elliot, personal communication, January 13, 2003). Most of the work done by SAP Labs is sent back for further development by SAP AG in Walldorf, Germany. SAP Labs has usability personnel in Bangalore, India, as well as designers in France, Montreal, Canada, and Tokyo, Japan. The User Experience group of SAP Labs was started in 1997, about the same time as SAP Labs, and is located in Palo Alto, California. Their forward-looking, or innovation, projects explore the cutting edge of technology and, after being researched, are eventually handed off to development to become products. They currently have 11 personnel who do work ranging from futurelooking, proof-of-concept projects to managing an internal SAP conference known as Tech Ed to accessibility research. Their major focus is on field studies of customers (e.g., between two and six customers for a project). When working on usability engineering projects, perhaps 80% of their time is spent on field studies with 20% spent on evaluation activities. The flagship product of SAP AG is the R/3 system, which covers mostly ERP applications. R/3 is offered in three different UIs: Java, Windows, and Web-based. R/3 originally had a Windows-only UI, and now the focus is on a Web-based UI, which is a direct rendering of the Windows UI. Web Application Guidelines. The SAP Labs User Experience group developed guidelines for Web accessibility for all of SAP, worldwide. The central usability group in Germany is responsible for developing the general UI design standards for Web applications.
Human Factors Challenges of Web Applications. For the SAP Labs User Experience group, the biggest challenge in designing Web applications has been the global nature of customer requirements. For example, a customer's requirements for a CRM application in Asia are completely different from the United States. The global nature of the requirements is difficult, as well as getting the data from customers in a timely fashion. Another challenge they see is the amount of time it takes to render a page in a Web application. Slow system
27. Web-Based Programs and Applications for Business response times become a human factors challenge and a UI design problem.
•
505
Typical Range of Time for Web Services Projects. The timing of projects is driven by the clients, but 2 weeks to 2 months is typical. A trend over the years is for projects to be conducted over shorter time periods.
External Groups One of the other ways to develop Web applications for businesses is to hire qualified professionals to perform usability services. There are a few larger organizations who do this work and many individual consultants in all parts of the world. We interviewed three of the more prominent independent firms that provide these services. We asked them to provide some information about their services, the challenges they see in delivering Web applications, and what advice they would give to organizations looking to hire an outside vendor. We conducted one-onone interviews with representatives of the following firms that provide usability services to companies creating Web designs for business applications: American institutes of research Background. American Institutes of Research (AIR) (www.air.org/program_areas/usability/usability-set.htm) has about 20 usability professionals located at offices in Concord, Massachusetts, Palo Alto, California, and Washington, D.C. (M. Wiklund, personal communication, January 13, 2003). Usability professionals are primarily located in AIR's Concord office. In Concord, they have two regular labs and one portable lab, while in Palo Alto, they have one usability lab. They also rent other labs and focus group facilities when necessary. Range of Services. including:
AIR provides a full range of services
• User requirements, Web site vision and goals, preferred look and feel • Conceptual modeling, conceptual design, navigation schemes, mock-ups, and screen templates • Interactive prototypes, detailed designs, and UI style guides • Usability testing (including competitive tests), cognitive process analysis, expert reviews • Workshops and support for establishing in-house usability groups Signature Services. AIR considers the services that clients most often ask for as: • Benchmark and competitive usability tests • Design of software and hardware user interfaces for special purpose devices (e.g., consumer electronic devices, medical and scientific equipment, household products) • Development of Web sites that deliver a large body of content to a diverse audience • Assessment of the efficacy of Web-based educational technologies
Challenges AIR Sees in Developing Web Applications for Clients • Ensuring early involvement of usability specialists so that they can have a positive influence on the underlying conceptual model • Scheduling usability tests so that developers have enough time to fix identified problems • Balancing the need to make the application visually and dynamically progressive as well as accessible (508 compliant) and compatible with older browsers
Advice to Firms Considering Hiring an Outside Vendor • Take some time to find a firm that you can trust • Let the firm help you decide what consulting services will be most helpful at a specified price point • Think about conducting iterative activities rather than a oneshot activity as an ultimately more cost-effective way to produce design excellence Nielsen Norman Group Background. The Nielsen Norman (NN) Group (www. nngroup.com) has 12 usability professionals, most located in Fremont, California (K. Coyne, personal communication, January 9, 2003). They rent lab facilities when necessary to run evaluations and group sessions. Range of Services. ing services:
The NN Group provides the follow-
• Design reviews, usability tests, competitive studies, and heuristic evaluations of Web sites • Learn-by-doing workshops and seminars on usability methods • Internal research studies and seminars disseminating information on research topics • Process mentoring, visioneering, and usability guidelines • Providing keynote speakers for meetings and conferences Signature Services. are:
The services clients ask for most
• Usability tests and Web site reviews • Keynotes, seminars and workshops • Research and research dissemination
Typical Range of Time for Web Services Projects. Timing of project is driven by client needs but the range is 3 days to several months.
506
VAUGHAN AND DUMAS
The Challenges the NN Group Sees in Developing Web Applications for Clients • Keep the internal politics of your organization out of the Web design • Listen to the advice you pay for
Advice to Firms Considering Hiring an Outside Vendor • Take some time to find a firm that you can trust • Let the firm help you decide what makes the most sense for you to do, including the range of costs for the services • Make sure you get cost-effective results from the activities— think about iterative services rather than a one-shot activity User-Interface engineering Background. User-Interface Engineering (UIE) (http:// uie.com) has six usability professionals and is located in Bradford, Massachusetts (J. Spool, personal communication, March 14, 2003). They rent lab facilities when necessary to run evaluations and group sessions. Range of Services
Know whether you want hands (contracting services) or brains (consulting services)
THE HUMAN FACTORS ISSUES ASSOCIATED WITH DESIGNING, IMPLEMENTING, AND EVALUATING BUSINESS WEB APPLICATIONS Design Constraints of Business Applications for the Web. Web applications differ from content-oriented Web sites in that they are designed and built to accomplish some specific tasks, rather than to present information or to entertain. Until recently (and in some cases still), these kinds of applications have been the realm of desktop and client-server environments, where a rich, highly interactive user experience is the norm. As organizations try to move these applications to the Web to take advantage of its advantages, they struggle to provide the kind of interactivity that is essential to the user experience. Users are not interested in the limitations of Web technologies, but want the kind of interactive experience that they have come to expect. For business end users and for user interface designers, the advantages of the Web include:
UIE provides the following services:
• Custom research to gather information about the people who use a site and competitive sites • Training to empower developers to create more usable and competitive sites • Consulting services to fill in the gaps in clients' skills Signature Services. The services clients ask for most • Educational services, especially its user interface conferences (http ://uiconf. com) • Organizational skills audits
• Ubiquitous access: Even for highly specialized applications, being able to access them from anywhere is a huge advantage. • Installation and maintenance: Not having to install and maintain applications on local machines is a huge savings of time and resources. • Branding and identity: The Web affords a much greater opportunity to establish a unique look and feel and identity and to connect with the user. • Real estate: In Web applications, users typically go to full Web pages rather than dialogue boxes. Web pages provide much more real estate to work with. This allows content to be organized logically, rather than breaking it up arbitrarily to fit into small dialogue boxes, as well as providing more room for instructional text.
Typical Range of Time for Web Services Projects. A few days for consulting services to 8-14 weeks for custom research.
The Challenges UIE Sees in Developing Web Applications for Clients • Understanding the constraints of the platforms users will have • Layering applications on systems that were not designed for the Web • Realizing that Web use can be killed by the accumulation of many little problems
Advice to Firms Considering Hiring an Outside Vendor • Get referrals for the firms you are considering • Know what you want to do
In transitioning applications to the Web, the complex operations are often easy to provide, but simple interactivity is hard to provide. For example, denning complex multidimensional queries against very large databases may be performed at least as a highly interactive Java component for the same purpose. But when it comes to being able to display the results of these queries, there may be no way in HTML to display the results in a table with fixed row and column headings and a scrolling data body. This very simple control, which users take for granted in desktop GUIs, is simply not available in HTML. Without it, navigating and interpreting a large table is very cumbersome. There are ways to solve this problem (e.g., by embedding Java applets or ActiveX controls), but this defeats a major part of the rationale for moving applications to the Web in the first place. Some of the other things that users take for granted in GUI applications that, at least at present, are given up in Web applications include:
27. Web-Based Programs and Applications for Business • Controls: Built on a mature platform, desktop GUIs can take advantage of a large set of rich controls. Some of the most useful controls that are missing from the Web include combo boxes and scrolling tables. • Menus and toolbars: On the Web, the browser, not the application, owns the menus and toolbars. This takes away some of the best tools for layering information and functions in the application and competes with the application for users' attention. • Direct manipulation: By taking advantage of local processing power, desktop GUIs can support direct manipulation. This does not just mean drag and drop, but also manipulating controls and seeing immediate visual feedback. • Windowing model: Desktop GUIs have a strong windowing model, including windows and dialogue boxes (modal and modeless) that enable layering information and functionality, while maintaining context. • State management: By being able to save local data, write to the registry, and so on, desktop GUIs can do a very good job of maintaining the state of an application, users' preferences, and personal information. • Pop-up windows: One of the important productivity tools in desktop GUIs is pop-up menus, typically accessed through the right mouse button. Business analyst and power uses depend on these menus for quick access to frequently used functions. At present, these tools cannot be used in HTML-based applications. The challenge for the user interface designer of a Web application is to provide analogs for these tools, especially for complex applications accessing databases and displaying large tables and dense graphics. The productivity differences between desktop GUIs and Web applications are strikingly contrasted when one asks professional office workers about accessing their e-mail either on their desktop or through a browser. Most do have e-mail available in both contexts. But when we did an informal survey at Oracle about which one respondents typically use, they said that they almost never use the browser version unless they are away from their office. The Web-based e-mail is available from nearly everywhere, but: • Often it is painfully slow, especially with messages containing graphic images and attachments • It shows only one page of messages at a time and requires a transit to the server to get more messages • It requires several steps and page refreshes to move messages between folders These and other limitations may have slowed down the spread of Web-based applications without Java applets. Human Factors Issues. In general, the design of Web applications for business uses is similar to the design of software intended for other environments. However, there are some human factors issues that apply to designing Web applications for businesses. These issues include:
507
Maintaining context in complex Web applications Save model Page-level buttons, to repeat or not to repeat Searching Complex create tasks, processes within processes The need for integration The pervasiveness of Excel and Word
Maintaining Context in Complex Web Applications The ability to maintain context in Web applications has proven challenging and problematic for Web users (Shubin & Meehan, 1997). As suggested by some of the above discussion, the Web was originally designed for documents, not transactions. Any single page is its own document, and the browser controls allow users to move through those documents. However, in a Web application, designers do not necessarily want a user to conceive of the system in that way. By using the browser controls, rather than the user interface controls that are part of the page design and layout, the user can quickly, and unintentionally, remove him or herself from the intended application flow. Design solutions for attempting to maintain this context include using only one browser window for a whole Web application (i.e., none or very few modal dialogues) or eliminating the browser controls altogether. Another part of the context issues goes to navigating the complex tasks supported by Web business applications. Browsing news online is a relatively straightforward task compared to creating and updating multiple dependent objects in a marketing automation application. The sheer volume of information, in addition to the actions users can take on that information, can quickly overwhelm the user. To deal with contextual navigation without menus, designers have developed visual methods of grouping and chunking information, such as via tabs and sub tabs, side navigation bars, and page footers that replicate tab structures.
Save Model A second complication of these complex tasks and the document model of Web browsers is what happens to the save model used by Web applications. Users have become accustomed to smart GUI applications that do not lose data just because the user opens another application window or navigates to another document. In the world of Web applications, losing data is much easier. At Oracle, an extensive usability study was conducted to understand the user's mental model of saving changes on a Web page or pages. Eighteen users were asked to perform eight tasks with a fictitious e-commerce Web site. They were presented with a variety of page types (i.e., a list of objects, a linear train page, and a page with a side navigation menu) crossed by different save models (i.e., explicit save, implicit save, and implicit save with caching). The primary finding was that users perceived that data on a given page would be saved (i.e., not
508
VAUGHAN AND DUMAS
lost unless explicitly cancelled) (a) when: when hitting a button labeled "Apply/Commit/Finish/Done" and (b) when navigating between steps in a flow. In the case of the first assumption, this meant understanding what users perceived as being "on a page." They included any fields, editable tables, and so on that were represented visually as being part of the same page. It also meant that users did not distinguish between the technical concepts of saving to a temporary location (e.g., a cache) versus committing to a database for a given page. To users, the information was all simply saved. In the second case, this meant understanding what users perceived as a group of related pages belonging to a flow. They did not perceive moving between tabs as related to a flow, but they did perceive visual representations of linear and nonlinear processes such as steps in a wizard as all related to one flow. The same issue of not caring about the difference between caching and committing to a database applies here as well. To solve the save problem, one should provide technical sophistication on the backend, for example, saving to temporary tables and caching, but not require that users understand the subtle, technical nuances. If, within the context of entering information, users have the ability to navigate away from a page or a flow without saving data (i.e., just caching), they must be warned to save their changes before moving on to other tasks. Also, when changes have been saved (i.e., versus simply cached) users should be notified of the changes being saved. This information is crucial feedback. A key component of the save model is consistency. An application should consistently, implicitly or explicitly, save information. In an implicit save model, the application always saves this information and users do not have to worry about losing information. In an explicit save model, users always have to take some action to save information, such as clicking on a Save button. Mixing the two models has the potential to cause confusion and loss of data and should be done only after validating a model with user testing.
Page-level Buttons, to Repeat or Not to Repeat In the world of Web applications, actions a user can take on an object are often displayed on a page with a lot of detail about that object. For example, a sales representative might want to create an order out of a quote by first viewing all the details of that quote and then clicking a button called Process Order. A dilemma arises when the actions (e.g., displayed as buttons) at the top of the page are not visible. At Oracle, a usability test was conducted with 12 users to examine buttons displayed at the bottom of the page and buttons displayed at the top and bottom of the page. Participants were shown a variety of page lengths and asked to complete two typical banking tasks. Across the page types, users were measured on the time from filling out the last field on a page to hitting a page-level button. The final result was a significant effect for page type, that is, on longer pages it took users longer to find buttons, and a significant effect for button location, that is, users performed faster across page types with buttons repeated at the top and bottom of the page. The
buttons at the top of the page often acted as advance reminders for users that they would be able to or need to take an action on the page content.
Searching Searching in Web applications is complex and problematic. Compared to the Internet, where users' expectations have begun to coalesce around simple text fields followed by a list of matching results (e.g., Google.com) or browsing lists of categories (e.g., Yahoo.com), searching across Web applications has become more complex. One issue of complexity that needs to be gracefully revealed in the user interface is that of scope. The scope of a search in a Web application can vary as widely as (a) searching on a list of objects currently being viewed (e.g., a list of employees), (b) searching across all objects available in the Web application (e.g., across a list of employees, a list of benefits, and a list of pay grades), or (c) searching across multiple applications or databases (e.g., find all documents written by employee Smith). An additional issue of complexity is the range of users' skill sets. As mentioned earlier, a Web application will typically have self-service, professional, and business executive user types. While the current expectations of Web searching (e.g., simple text field and result set) might suffice for self-service users, the same expectation will not work for professional users who have become highly skilled in both the content of and finding information in the application. Designing for this diversity of user types can be addressed by creating different classes of user interfaces that progressively disclose more powerful features useful to professional users, such as saving searches, filtering, and sorting (Vaughan & Beier, 2002).
Complex "Create" Tasks, Processes-Within-Processes In the world of Web applications, users are occasionally called on to create multiple objects as part of the same creation process. For example, a call-center representative, while creating an order for a customer, may also have to create a new customer record as well. These objects may or may not be packed with information, and so they vary in the degree of visual complexity. In the world of GUI applications, these layers of process and their relationships to each other (e.g., this order is related to that customer) are revealed with modal dialogue windows. This is not so easily done in Web applications, where context must be aggressively maintained in order not to confuse the user. At Oracle, a detailed study was conducted of end user expectations about processes within processes across a variety of visual representations (e.g., a train station visual with route and stops on the route, a side navigation bar, a horizontal navigation bar). Different combinations of these visuals were tested (e.g., going from a train to a side navigation bar) across five tasks and eight users. The most important findings were (a) that users were able to learn to move between processes and (b) that users perceived a processes-within-processes model
27. Web-Based Programs and Applications for Business as one long, unified process, rather than individual processes launched in relationship to each other.
The Need for Integration Users of Web applications in a business context, especially professional users, may have a need to view information from other applications as part of doing their job. For example, a person approving credit applications for desktop computers may want to view a customer's order history before approving a new line of credit. This ability to view needed information that is often owned by other applications is both a design and a technical challenge. However, it is a necessary challenge in order for users of business applications to gain the full benefit of their customer/organizational/employee data. From a technical standpoint, this means providing application programming interfaces (APIs) that are capable of sharing data with other applications, as well as a product architecture that can share discrete bits of data between job roles. From a human factors standpoint, it means providing a user interface architecture and information design that accommodates the users' need to (a) do high frequency, complex tasks, as well as (b) view lots of contextually related information.
The Pervasiveness of Excel and Word In a business environment, Microsoft's productivity applications Word and Excel are dominant (www.olapreport.com/origins. htm). The presence and importance of these two tools has an impact on users' expectations and needs for Web applications. At a simple level, the prevalence of these tools means that users will need the technical ability to transfer data between Web applications and Word and Excel. From a human factors standpoint, this means users will need a user interface that walks them through the process of exporting data to the desktop that is relatively simple and through the process of importing data that is much more complex. The complexity of the import process is due to the need to integrate the incoming data from Excel or Word into the Web applications data structure. For example, in order to import a list of customer addresses, users need to make sure the addresses are divided into the proper groups and aimed at the right import tables in the Web application. A user would not want to import customer names into the tables where the street addresses are stored. In addition to the need to move data back and forth between Web applications and productivity tools, the prevalence of Word and Excel impact users' expectations of the operation of user interfaces. For example, users who make frequent use of Excel for analytic applications expect an Excel look and feel, as well as Excel-like features in a user interface offering analytic capabilities. For example, Excel users expect to be able to click on a cell in a table to place it in focus or to be able to create and save annotations about the data in a cell. In the case of Web-based e-mail editors that are often part of marketing applications (i.e., for e-mail blasts to a list of customers), users' familiarity with Word leads them to expect both a similar look and function-
509
ality (e.g., to select Bold by clicking a "B" icon) in the text editor. Requirements Gathering and Evaluation Issues With WebBased Business Applications. Zhu, Vu, and Proctor (chap. 18, this volume) discuss evaluating Web usability. Here our focus is on the unique issues that apply to evaluating Web applications for businesses. Specifically, we discuss five key areas: 1. 2. 3. 4. 5.
Requirements gathering for Web applications Understanding business processes Evaluating different types of Web applications for businesses Selecting and recruiting business users Obtaining usability feedback over the Web
Requirements Gathering In general, the usability engineering methods used for requirements gathering of desktop or other applications can be applied to the realm of Web applications. However it is helpful to understand requirements gathering in a Web domain. For more information and examples, see Proctor et al. (2002) and Vaughan, Candland, and Wichansky (2001) for discussions of the following methods applied to Web-based designs: • • • • • • • • • • • • •
Interviews Verbal protocol analysis Group task analysis Narratives and scenarios Critical incident reports Questionnaires Focus groups Wants and needs analysis Observation and contextual inquiry Ethnographic studies User diary Concept sorting Log files
Understanding Business Processes The quality of all of the usability evaluation methods depends on the skill and experience of the usability specialist using them. Evaluating business applications requires a knowledge of the business processes the application is attempting to automate. This knowledge is especially needed to evaluate applications that are used for analysis rather than simple self-service applications. Nielsen's (1992) finding that usability specialists who also have domain knowledge find more usability problems in an expert review than usability specialists without domain knowledge highlights the importance of understanding the work users are trying to accomplish. Without an understanding of how businesses are run or, for example, how budget targets are allocated in an organization, it is difficult to design an effective evaluation
510
•
VAUGHAN AND DUMAS
and to create solutions to usability problems. This is especially true when the design issues involve concepts rather than steps in a process.
Evaluating Different Types of Web Applications A Web application intended for complex business processes requires different procedures for evaluation than evaluating a self-service Web application. In a self-service usability test, the participants are given very little orientation or training. Even a single error by a test participant is an important event. Furthermore, in a self-service application, when a task requires an assist by the test administrator, it usually indicates a design flaw that needs attention. But in more complex applications, especially ones in which end users will receive training and will use frequently, learning the application is often a trial-and-error process spread over weeks or even months. In that context, picking the wrong menu or tab the first time may not indicate a problem requiring fixing. In a more professional application, helping the participant get through a task does not necessarily indicate a serious usability problem. Of course it might, but the meaning of an assist can be different in complex applications. The focus is often on mastering complex ideas within the conceptual design of the application or the user interface. The transition to HTML-based user interfaces puts constraints on the manipulations users can make. Simple actions such as selecting a cell in a table or dragging an object may work differently from what users expect from their experience with a desktop application. Participants often blame the user interface design for these constraints rather than the technology. These constraints need to be considered in an evaluation. They can influence a participant's subjective ratings of usability and contribute to measures such as statements of frustration. Finally, in an organization that implements a user interface style guide, a usability specialist needs to have a thorough understanding of its details and the organization needs to use expert reviews or heuristic evaluations to detect products' deviations from the guidelines. It is typically only in an expert review that violations of a style guide are discovered. Participants in a usability test will not detect them and may miss many inconsistencies that the violations create. Consequently, an organization with a style guide needs usability specialists or a user interface that can use expert reviews and usability tests in an effective combination. Most often, the expert review is done first to clean up obvious usability issues and guideline violations so that the valuable and expensive time of usability testing participants can be used to find problems unique to the target users of the application.
Selecting and Recruiting Business Executives It is a truism that evaluating usability with user-based methods requires recruiting users who are part of the target market for the product. But for many business applications, finding business executives and getting them to commit to participate is
always a challenge. For example, an application based on balanced scorecard technology (Kaplan & Norton, 1996) may be used by CEOs and senior vice presidents. Testing it with financial analysts may be useful, but it will not uncover the subtle usability problems that only executives will find. Senior executives can be found and will take the time to test a new application, but it often takes some ingenuity to get them to commit their time. They want to participate when it is convenient for them, usually in the evening or early in the morning. Having them participate remotely from their own office can help. It may also be necessary to coordinate contacting senior executives with the relevant sales representative in the development organization. Sales representatives are understandably very protective of CEOs of organizations they sell to. There is nothing more awkward than having a sales representatives call a human factors specialist to complain that a CEO will not agree to renew their license because they want the newest version of the product they tested in the usability lab. Setting expectations ahead of time can avoid problems like this one.
Obtaining Usability Feedback Over the Web Before the burst of interest in Web use beginning the mid-1990s, user feedback of non-Web applications tended to focus on usability testing. But Web site designers were uncomfortable with the small number of users typically sampled for user tests. Many Web applications were aimed at broad populations of users and there was a tendency to think that sites were aimed at everyone. These perceived needs lead designers and human factors professionals to think about ways to gather data from larger samples of users and to capture data in real time as users were using the site. Obtaining meaningful data in real time, that is synchronous data gathering, has proven elusive however. Synchronous data that can be captured, such as mouse clicks and pages visited, are often at too low a level to provide meaningful information about usability. Examining site logs (on the Web server) may provide information about the popularity of a site or a page, but not much about its usability. For further information on automated data gathering see Hartson, Castillo, Kelso, Kamler, and Neale (1996) and Perkins (2001). A more recent study has used asynchronous data collection over the Web to increase sample size (Tullis, Flieschman, McNulty, Cianchette, & Bergel, 2002). Using this technique, the participants see two open browser windows: one has the Web site being evaluated in it and the other displays a set of questions that the participants answer about the site. The questions are analogous to tasks in a usability test, and a tinier starts when the participants begin the task and ends when they answer the question. Tullis et al. show that the data they obtained with this technique are comparable in most ways to data obtained from laboratory usability testing. However, the method allowed them to sample many more users than a traditional test. Another way to get feedback on Web usability is to create a Web survey. These surveys are placed on Web sites for visitors to complete. While a business organization could create their own survey, there are several surveys available that professional
27. Web-Based Programs and Applications for Business test developers have created. See Lazar (2002) for a list of these surveys and guidance on using them.
CONCLUSION Current Status of Web Applications for Businesses. The current state of Web applications for businesses reveals many technical constraints on the user experience (e.g., the selection model, maintaining context, the save model) that in many ways are regressions in quality from the desktop or GUI user experience. Some of these constraints will eventually disappear as HTML technology develops further, and some will remain as continuing annoyances to end users (e.g., effective ways to scope a search). What will continue to remain as design constraints are the size of a company, the type of applications or tasks, and the types of users. Future of Web Applications for Businesses. The near future of Web applications will likely mean greater interactivity, but the longer term future is less clear. We now discuss two areas currently being investigated for their potential relationship to Web applications for businesses.
Multimodal Input-Output On the horizon, Web applications will be more integrated into the ubiquitous computing experience. In the late 1990s there was a boom in interest and development work around mobile computing. With the fading of the telecommunications industry at the end of the dot-corn era, that interest has transformed into a more mature interest in multimodal research. Multimodal research is a relatively new problem space for human computer interaction. For a detailed discussion of the domain and the potential research agenda, see Bersen (2002). In general, multimodal research seeks to examine (at least so far) the input-output of media (both text and graphics), acoustics, and haptics, occurring as sequential or parallel processes for end users. Much of the current multimodal research focuses on applications for manufacturing, mapping, and graphs. In terms of business applications, one area of potential is the speech domain (i.e., speech as input and output). Also, as the tablet Personal Computer develops into a mature product, we will likely see enhancements with stylus and handwriting input. What this means for business end users is that they can start using the optimal modality or combination of modalities for their work context, for example, field service personnel using stylus and speech input or health care providers using speech input. In terms of implications for Web applications this means (a) designing a user experience that is some combination of multimodal input and output and (b) designing Web applications to deal with input and output to other modalities (e.g., via synchronization). In some cases, these other applications may not be Web applications at all. The major hurdles to maturation of this technology include the current need for end user training, the lack of data about user performance and preferences, and the trade-offs that will need to be made (e.g., lack of data about what users' error tolerances are).
•
511
Business Intelligence Visualization Another emergent issue for business applications for the Web is the advancement of business intelligence tools and the need for easier to use but powerful analytic tools. Two examples of these tools being explored for Web applications for businesses include network visualization and manipulation (cf., Freeman, 2000; Keim, 2001; Lamping, Rao, & Pirolli, 1995; Van Ham, van de Wetering, & van Wijk, 2001; Yee, Fisher, Dhamija, & Hearst, 2001), as well as easy-to-use ad hoc query building tools (cf., Anick, Brennan, Flynn, & Hanssen, 1990; Jones, 1998; Pane & Myers, 2000; Spence, 2001; Spoerri, 1993; Young & Shneiderman, 1993). Other business areas that can benefit from visualization techniques include Pharmaceuticals via data mining tools, and traffic management based on in-ground sensors. Network visualization. In terms of representing networks visually, the need stems from very specific business problems. In areas such as customer relationship management, the ability to manipulate and explore views of complex customer data will be key for sales representatives and their managers. For example, sales managers may want to view the bottlenecks in building relationships with clients in order to drive more sales. This is a new challenge for Web applications, to give decision makers that ability to comprehend a high-level view of activities and relationships and then rapidly drill down to low-level information. This will provides managers and executives the ability to strategically analyze networks for business intelligence purposes. Some of the specific human factors problems to be solved include the ways of representing networks, finding improved ways of navigating networks, and finding better ways to organize networks in the UI. Also needed is work on representing linkages between graphical representations and timeline based information. End User Query Building. For business executives, the ability to execute complex queries for ad hoc analyses will be key to understanding the health and future of a business. Currently this work is the domain of the highly skilled, computer savvy business analyst. This person usually works for a business executive who wants to ask questions. A critical question for executives is whether we should we make complex queries easier for these end users to construct. One camp says that we should allow executives to construct ad hoc queries and thus get whatever data they want. Another camp says that we should only have highly trained analysts doing this job, and they in turn can create canned queries for executives. From a human factors standpoint, creating an easier to use query-building tool involves empowering the end user with dynamic query previews and hence enabling users to get a better idea of the query before they execute it. This preview function would also help to train users on the functioning of the query tool. Current research in this area is examining the concept of dynamic query previews in conjunction with the traditional concept of Boolean logic, that is, trying to make Boolean logic explicit through visualization, such as using Young and Shneiderman's (1993) filter-flow model.
512
VAUGHAN AND DUMAS
APPENDIX
ACKNOWLEDGMENTS
Web Application Guidelines and Usability Resources Source Category Web sites
Journal articles/ Chapters Books
Resource National Cancer Institute, Web Design and Usability Guidelines http://usability.gov/ guidelines Yale University, School of Medicine Web Style Guide http://info.med.yale.edu/caim/manual Usability Professionals Association http:// www.upassoc.org/html/resources.html* design-style Society for Technical Communication http://www.stcsig.org/usability/topics/ design-for-usability.html Oracle Technology Network, Browser Look and Feel Guidelines http://technet.oracle.com/ tech/blaf/ Lazar, J. (2002). The World Wide Web. In J. Jacko & A. Sears (Eds.), The human-computer interaction handbook (pp. 714-730). Mahwah, NJ: Lawrence Erlbaum Associates. Nielsen, J. (1999). Designing Web usability: The practice of simplicity. Indianapolis, IN: New Riders. Lynch, P., & Horton, S. (2002). Web style guide: Basic design principles for creating web sites. New Haven, CT: Yale University Press.
We would like to thank Janice Rohn, formerly director of user experience at Siebel, Christian Pantel, director of user experience at Peoplesoft, Manik Singh, usability and accessibility specialist, and Keith Elliot, director of the accessibility competence center and manager of the user experience group at SAP Labs, and Daniel Rosenberg, vice president of usability and interface design at Oracle who provided the internal organization profiles, as well as their help in commenting on and reviewing this chap-
ter. Thanks also go to Michael Wiklund of American Institutes of Research, Kara Pernice Coyne of the Nielsen Norman Group, and Jared Spool of User Interface Engineering who provided the external vendor profiles. We would like to thank Kristin Thompson, manager of CRM UI; George Hackman, manager of ERP UI; Dr. Richard Wright, manager of server and tools UI; Dan Workman, manager of analytic tools; and Jeremy Ashley, manager of business intelligence UI, from Oracle's Usability and Interface Design group, for providing key insights into the different application domains. Joseph H. Goldberg, principal research scientist at Oracle Corporation, and Delia Grenville, senior research scientist at Oracle Corporation, provided key insights into the future of Web application user interface issues. Many thanks also go to Anna Wichansky for providing thoughtful and detailed review and commentary on our drafts.
References Anick, P. G., Brennan,J. D., Flynn, R. A., &Hanssen, D. R. (1990). A direct manipulation interface for Boolean information retrieval via natural language query. Proceedings of ACM SIGIR 1990 Conference (pp. 135-150). New York: ACM Press. Beier, B., & Vaughan, M. W. (2003). The bull's-eye: A framework for Web application user interface design guidelines. Proceedings of CHI 2003: Human Factors in Computing Systems (pp. 489-496). New York: Association for Computing Machinery. Bersen, N. O. (2002). Multimodality in language and speech systems— from theory to design support tool. In B. Granstrom et al. (Eds.), Multimodality in language and speech systems (pp. 93-148). Netherlands: Kluwer Academic. Donoghue, K. (2002). Builtfor use: Driving profitability thorough the user experience. New York: McGraw-Hill. Freeman, L. C. (2000). Visualizing social networks. Journal of Social Structure, 1(1). Available: www2.heinz.cmu.edu/project/ INSNA/joss/vsn.html. Hartson, H. R., Castillo, J. C., Kelso, J., Kamler, J., & Neale, W C. (1996). Remote evaluation: The network as an extension of the usability laboratory. Proceedings of CHI 1996: Human Factors in Computing Systems (pp. 228-235). New York: Association for Computing Machinery. Jones, S. (1998). Graphical query specification and dynamic result previews for a digital library. Proceedings of the 11th Annual ACM Symposium on User Interface Software and Technology (pp. 143151). New York: ACM Press. Kaplan, R. S., & Norton, D. P. (1996). The balanced scorecard: Trans-
lating strategy into action. Cambridge, MA: Harvard Business School. Keim, D. A. (2001). Visual exploration of large data sets. Communications of the ACM, 44(8), 39-44. Lamping, J., Rao, R., & Pirolli, p. (1995). A focus + content technique based on hyperbolic geometry for viewing large hierarchies. Proceedings of CHI 1995: Human Factors and Computing Systems (pp. 477-483). New York: ACM. Lazar, J. (2002). The World Wide Web. In J. Jacko & A. Sears. (Eds.), The human-Computer interaction handbook (pp. 714-730), Mahwah, NJ: Lawrence Erlbaum Associates. Lynch, P., & Horton, S. (2002). Web style guide: Basic design principles for creating Web sites (2nd ed.). New Haven, CT: Yale University Press. Nielsen, J. (1992). Finding usability problems through heuristic evaluation. Proceedings of CHI 1992: Human Factors in Computing Systems (pp. 373-380). New York: Association for Computing Machinery. Nielsen, J. (1999). Designing Web usability: The practice of simplicity. Indianapolis, IN: New Riders. Pane, J. F., & Myers, B. A. (2000). Improving user performance on Boolean queries. In G. Szwillus & T. Turner (Eds.). CHI 2000 Extended Abstracts: Conference on Human Factors in Computing Systems (pp. 269-270). New York: ACM Press. Perkins, R. (2001). Remote usability evaluation over the Internet. In R. Branaghan (Ed.), Design by people for people: Essays on usability (pp. 153-162). Chicago: Usability Professional's Association.
27. Web-Based Programs and Applications for Business Proctor, R. W., Vu, K. L., Salvendy, G., Degen, H., Fang, X., Flach, J. M., Gott, S. P., Herrmann, D., Kromer, H., Lightner, N. J., Lubin, K., Najjar, L., Reeves, L., Rudorfer, A., Stanney, K., Stephanidis, C., Strybel, T. Z., Vaughan, M., Wang, H., Weber, H., Yang, Y., & Zhu, W. (2002). Content preparation and management for web design: Eliciting, structuring, searching, and displaying information. International Journal of Human-Computer Interaction, 14, 25-92. Scholtz, J., Laskowski, S., & Downey, L. (1998). Developing usability tools and techniques for designing and testing web sites. Proceedings of the Human Factors and the Web [Online] Available at: www.research.att.com/conf/hfweb. Shubin, H., & Meehan, M. M. (1997). Navigation in Web applications. Interactions, 4(6), 13-17. Smith, S., & Mosier, J. (1984). Design guidelines for user-system interface software (Report ESD-TR-84-190). Bedford, MA: MITRE Corp. Spence, R. (2001). Information visualization. New York: AddisonWesley. Spoerri, A. (1993). InfoCrystal: A visual tool for information retrieval and management. Proceedings of the ACM Conference on Information and Knowledge Management (pp. 11-20). New York: ACM Press. Stewart, T., & Travis, D. (2002). Guidelines, standards, and style guides. In J. Jacko, & A. Sears (Eds.), The human-computer interaction handbook (pp. 991-1005). Mahwah, NJ: Lawrence Erlbaum Associates. Tullis, T, Flieschman, S., McNulty, M., Cianchette, C., & Bergel, M. (2002). An empirical comparison of lab and remote usability testing of Web sites. Paper presented at the meeting of the Usability Professionals' Association., Orlando, FL.
•
513
Van Ham, F., van de Wetering, H., & van Wijk, J. J. (2001). Visualization of state transition graphs. Proceedings of 2001 Information Visualization (pp. 59-66). Los Alamitos, CA: IEEE Computer Society. Vaughan, M. W., & Beier, B. (2002). Designing search UIs for a diversity of users: The case of Oracle's search UI guideline. Proceedings of the Human Factors and Ergonomics Society 46th Annual Meeting (pp. 628-630). Santa Monica, CA: Human Factors and Ergonomics Society. Vaughan, M. W., Candland, K. M., & Wichansky, A. M. (2001). Information architecture of a customer web application: Blending content and transactions. In M. J. Smith, G. Salvendy, D. Harris, & R. J. Koubek (Eds.), Usability and interface design: Cognitive engineering, intelligent agents, and virtual reality (pp. 833-837). Mahwah, NJ: Lawrence Erlbaum Associates. Vergo, J., Noronha, S., Kramer, J., Lechner, J., & Cofino, T. (2002). E-commerce interface design. In J. Jacko & A. Sears (Eds.), The human-computer interaction handbook (pp. 758-771). Mahwah, NJ: Lawrence Erlbaum Associates. Wiklund, M. E. (1994). Usability in practice: How companies develop user- friendly products. New York: Academic Press Professional. Yee, K., Fisher, D., Dhamija, R., & Hearst, M. (2001). Animated exploration of dynamic graphs with radial layout. Proceedings of 2001 Information Visualization (pp. 43-50). Los Alamitos, CA: IEEE Computer Society. Young, D., & Shneiderman, B. (1993). A graphical filter/flow representation of Boolean queries: A prototype implementation and evaluation. Journal of the American Society of Information Science, 44(6), 327-339.
28 DESIGNING E-COMMERCE USER INTERFACES Lawrence J. Najjar BMC Software, Inc.
INTRODUCTION e-Commerce is growing rapidly. In the United States, online retail sales increased an estimated 40% from last year (eMarketer, 2003a). Online customers are not buying more. Instead, they are switching some of their purchasing from off-line stores and catalogs to online stores (Kapadia & Moore, 2002). This is especially true for holiday shopping. In the United States, online sales are now about 5% of total retail sales and increasing. Growth is also occurring internationally. For example, Canadian e-commerce sales increased 59% (eMarketer, 2003b) and European e-commerce sales may increase as much as 67% this year (weaver, 2003). The United Nations believes that 1 in 10 persons worldwide has Internet access and predicts that 18% of all purchases by firms and individuals will be made online by 2006 (UNCTAD, 2002; U.S. Census Bureau, 2003). By providing quick, convenient access to products, product information, and prices, the Web is turning products into commodities. The ease of use of e-commerce sites is a way to differentiate a site, increase market share, and enhance a brand (Manning, McCarthy, & Souza, 1998). Seventy-nine percent of users named easy navigation as the most important characteristic of an e-commerce site (Lake, 2000). However, e-commerce sites lose up to 50% of potential online sales because users cannot find what they want (Cohen & Thompson, 1999; Seminerio, 1998). Poor navigation and slow downloads helped make 83% of shoppers leave a site (Thompson, 1999) and up to 78% of shoppers abandon their online shopping carts (BizRate, 2000; ePaynews, 2003b). When an e-commerce site is easy to use, sales can increase. For example, after improving usability, IBM obtained a 400% increase in sales on IBM.com (Battey, 1999; Tedeschi, 1999) and Digital Equipment Corporation reported an 80% increase in revenue (Wixon & Jones, 1992). By making the products easier to access, removing unnecessary graphics, and making product 514
information easier to scan, Liz Claiborne's Elisabeth.com tripled the rate at which lookers became buyers (Tedeschi, 2002a). Thirty-five percent of online consumers said they would buy more products if the products were presented better on ecommerce sites (Vigoroso, 200 Ib). Competing sites are only a click away. An e-commerce site that is easy to use can build consumer loyalty (Najjar, 1999). Loyalty is essential because two of every three sales are made to consumers who know exactly from whom they want to buy online (ActivMedia Research, 2000). Also, once users start buying online, they increase the amount of their purchases each year (ePaynews, 2003a, 2003b; Hansell, 2002).
Process A good design process can be more important than a good designer. A good design process is driven by the needs and preferences of users, identifies and controls requirements so design objectives stay consistent, involves users and clients early and throughout the process, and uses repeating design—user feedback—improve design cycles to maximize usability (Mayhew, chap. 19, this volume). To design efficiently, start with high-level design and work to low-level design, use tools that allow you to make design changes quickly and easily, and communicate your designs using media that allow users and clients to immediately understand the user interface (Najjar, 2002). The following efficient steps for user interface design evolved from many years of performing commercial user interface design work under significant time pressures. These steps emphasize getting high-quality usability work completed in a short period of time. This work is typically performed by usability engineers, not by visual designers or programmers. 1. Develop a business strategy—Define the objective for the site (e.g., generate revenue, drive consumers to the
28. Designing E-Commerce User Interfaces brick-and-mortar store, promote the brand). Determine how the e-commerce site will be different from and better than competing sites. Because the strategy drives the design, document the strategy and get it approved by the clients. 2.Define the users—Identify the users, their objectives, contexts (e.g., home vs. work), computers, display sizes, display resolution, browsers, and connection speeds. Design to meet the needs of the users. 3. Define the functional requirements—Identify and prioritize the functions users want on the site. Use tools such as focus groups, interviews (including interviews with your clients), competitive assessments, and contextual inquiries to gather functional requirements. For example, "Registration" may be a functional requirement. Prioritize the functions using criteria that include value to the user, differentiation from competitors' sites, and ease of implementation. Include some functions that may get low priority, but are essential (e.g., Contact Us). If there are too many functions for the planned project schedule, move some of the functions into later projects. 4. Write use cases—Break down the functions into more detailed user tasks. For example, break down the "Registration" function into use cases for "Show confirmation," "Edit registration," "Remove registration," and "Show error message." Prioritize the use cases with criteria that include user priority, differentiation from competitors, and ease of implementation. Working with engineers, identify systems (e.g., databases) that are involved with each task. Use this information to scope the project and to move some use cases into future projects. 5. Develop site structure diagrams—Using a product such as Microsoft Visio, draw diagrams with boxes and arrows that show how the sections and subsections of the application are organized and named. Review and change the diagrams based on feedback from the clients. 6.Build interactive mock-ups—To allow clients to give you feedback on the user interface design, develop an interactive, hypertext markup language (HTML) mock-up that shows the organization of the application, the functions on each page, and how each function works (Najjar, 2000). Show functionality. Show entry fields, buttons, drop-down menus, hyperlinks, confirmation windows, and error messages. Do not show graphics (Fuccella & Pizzolato, 1999) because they distract reviewers from the functional design and take too long to create and change. Do not take the time to create robust, complete, professional HTML. Instead, use an HTML authoring shell such as Macromedia Dreamweaver. Show fake data. Do not connect the HTML to actual databases. The purpose of the interactive mock-up is to show clients how the site may work, not how it may look. Form follows function (American Heritage Editors, 2000). Define the functional user interface, then design the visual interface to support it. For example, to identify important functions, use location, size, contrast, color, and shape to get users' attention. Conduct iterative design reviews of every representative page of the mock-up with the clients. Put the interactive mockup behind a firewall so remote clients can access it during teleconference design reviews, examine it in detail, and show it to their colleagues. Once the clients approve the first two or three
515
pages of the mock-up, the visual designers can create their visual compositions. Figure 28.1 shows a sample interactive mock-up for a generic e-commerce site home page. The mock-up shows the organization of the site and interactive user functions. It does not include graphics. The user interface designer can improve the interactive mock-up quickly, easily, and cheaply. 7.Test the usability of the interactive mock-ups—Add graphics to several typical, important, or hard-to-use pathways in the mock-up. Get representative users and test the usability of the mock-up. Take this user feedback and improve the interactive mock-up. Ideally, perform this test-redesign cycle several times before programmers write a single line of code. That way, the user interface design drives the programming. 8. Write functional design specifications—To allow the front-end programmers to do their work, capture an image of each page in the interactive mock-up and place it into a Microsoft Word document. List each of the controls on the page. Then describe how each control works (e.g., default state, available choices, changes that occur if users are registered, changes that occur when users performs each action, error messages). Work with your programmers to make sure you give them the information they need to bring the design to life. Because you quickly and cheaply iterated the design based on feedback from users and clients, the programmers should be able to do their work right the first time. This technique is very efficient and should prevent the programmers from redoing their work. 9.Perform user acceptance tests—Perform a usability test of the final version of the site. Connect the site to databases. Look carefully at download times and the alignment of images and text. Because you already iterated the design with representative users and the clients, this test should go well. Make any needed changes to the final code, then "go live." Other steps you might want to add to the process include evaluations of competitor's sites and a usability heuristic evaluation or usability expert review of the existing site. After the site is live, review site metrics (e.g., where users exit the site) to identify opportunities for improvement.
Design Designing the user interface for an e-commerce site is very challenging. E-commerce sites must accommodate nearly all users, include a significant amount of user interactivity, and still be easy to use. Major sections to design include the overall page format (see Tullis, Catani, Chadwick-Dias, & Cianchette, chap. 7, this volume), navigation, catalog, registration, personalization, checkout, and customer service. Also, design for multiple languages and users with disabilities (see Choong Plocher, & Rau, chap. 16, and Vanderheiden, chap. 15, this volume). Page Format. Because download time is the biggest problem with the Web (Graphics, Visualization, and Usability Center, 1998), design your pages to download quickly. Limit
516
•
NAJJAR
FIGURE 28.1. Sample interactive mock-up for generic e-commerce site home page.
your use of graphics and optimize them for the Web. Do not use plug-ins like Flash (Nielsen, 2000; Ragus, 2000b, 2000c). Do not use Java. Design your pages to download in less than 10 seconds (Miller, 1968; Nielsen, 1994, 1997) over a 56K modem because that is how most people access the Web from home (e.g., Harris Interactive, 2003; Orr, 2002; Romero, 2002). To do this, limit the size of each page to 40K or less (Lamers, personal communication, February 27, 1996, Sacharow & Mooradian, 1999; Sullivan, 1998). Horizontal scrolling is annoying and makes users work too hard. Except for a product comparison tool, never require users to scroll horizontally. Avoid forcing users to scroll vertically on the home page. However, it is acceptable to put closely related information (e.g., product details, checkout fields) on a vertically scrollable page (Sacharow & Mooradian, 1999). Put important information "above the fold" (above the vertical scroll line) so users can see it immediately. To make it easy for users to interact with the Web site, format the pages so user interface elements are in familiar locations (Bernard, 2001). As shown in Fig. 28.2, put the return-to-Home hyperlink in the top, left corner. To make it serve double-duty as a branding element, use the company's logo as the hyperlink. Put global navigation controls for the major sections of the site across the top of the page (Pastore, 1999). Place local navigation
controls that work inside each major section along the left side of the page. On the right side or the bottom, left side of each page, put controls (e.g., promotions) that take users off the site. Locate the search entry field near the top of the page, on the left side, below the global navigation controls. Avoid using banner advertisements, even at the top of each page, because banner ads do not work very well (Benway & Lane, 1998). Users need to know what is in their shopping carts. Except for the shopping cart page and checkout pages, try to put a shopping cart summary on each page. In the shopping cart summary, show the quantity of each item, a short hyperlinked product name for each item, the price, and a cost subtotal (Chaparro, 2001; Pastore, 1999; Ragus, 2000a). Also include a link to the complete shopping cart and a link to check out. As shown in Fig. 28.2, put the shopping cart summary on the left side of each page, below the search entry field and local navigation. To easily accommodate registered members, include signin entry fields on the home page and a sign-in hyperlink on every page. Provide a welcome notice (e.g., "Welcome back, john.doe") inconspicuously near the top of the home page so registered users know they are recognized and are receiving member benefits such as express checkout. Be sure to put a "Contact Us" hyperlink on every page, perhaps at the bottom.
28. Designing E-Commerce User Interfaces
517
FIGURE 28.2. Sample interior page for e-commerce site. A good way to get new, loyal users is to put at the bottom of each page a link to a simple referral form in which users need to enter only sender and receiver e-mail addresses (e.g., ImageExchange.com's "Tell a Friend"; Reichheld & Schefter, 2000). E-commerce users are very concerned about providing their personal information online. To reduce this concern, provide links to the privacy and security policies at the bottom of every page, but also show the links more obviously where users are entering private information in the content area of the registration and checkout pages (Stanley, McCarthy, & Sharrard, 2000). Design the user interface to encourage purchases. On the home page and the first page of each major section, include a promotion for a product that will appeal to most of your visitors. Select a moderately priced item that was reviewed very favorably by consumers. Show the product name, image, short description, price, a "Put in Shopping Cart" button, and a link to the complete product description page. Navigation. To make it easy for users to move through your site, design navigation that is simple, intuitive, and obvious. Put the navigation controls in the same locations on each page. Use navigation to tell users where they are, how they got there,
and where else they can go (Fleming, 1998). This is especially helpful to users who arrive at the page not from the home page, but via a search or hyperlink. Provide "breadcrumb" navigation on the site (Rogers & Chaparro, 2003). Breadcrumbs are small, hyperlinked page titles at the top of each page, usually above the title of the current page. These hyperlinks show the page titles users came through to get to the current page (e.g., "Home > Men's Apparel > Shirts"). Like the breadcrumbs dropped by Hansel and Gretel (Grimm & Grimm, 1999), the breadcrumb navigation controls allow users to easily retrace their steps. To make the site more inviting, provide up to seven intuitive names for the major sections of the site. Familiar names will make it easier for users to quickly browse through the site to a desired product. These names serve as the global navigation controls. Design the navigation so users can browse to any product in five clicks or less (Tracy, 2000). It may be better to provide more category names at each level (a broad design) than to provide more levels to click through (a deep design; Selingo, 2000). Also provide specialized browse functions to meet user needs. These specialized browse functions can include product ensembles (e.g., EddieBauer.com "Coordinating Sets"), suggested gifts, holiday specials (e.g., RedEnvelope.com special
518
NAJJAR
FIGURE 28.3. Sample search results page.
occasions), discounted products, and brands (e.g., Target.com's "Shop by Brand"). Search is an extremely important navigation technique. In one study (eMarketer, 2001), 43% of users said that search was the most important online shopping feature. Unfortunately, many e-commerce sites do not design, maintain, or even evaluate their search functions (Hagen, Manning, & Paul, 2000). One study (Gunderson, 2000) found that 42% of e-commerce sites had inadequate search functions. Another study (Nielsen & Tahir, 2001) found that 36% of the time users could not find what they were looking for when they used the search function on 25 e-commerce sites. To improve search, use meta-tag tools (e.g., Dynabase and Spectra), thesauruses, alternate spellings, and database search engines (e.g., EasyAsk, Fact City, iPhrase, Mercado, and Requisite Technology; Guernsey, 2001). For example, the advanced search function on TowerRecords.com allows users to misspell search words (e.g., enter "beetles" and get "Beatles") and to
enter various information (e.g., title, cast/crew, film category). The Tower Records search engine doubled the rate at which users made purchases (Guernsey, 2001). Allow users to search by product name, product category, brand, model/item number, and price (Consumer Reports, 2003). To make the site convenient for users, always try to get the most relevant hits in the first page of search results. As Fig. 28.3 shows, on the search results page display the searched-for keywords and allow users to perform another search, refine the search results, and sort the search results using helpful product attributes such as price or size (e.g., NetGrocer.com).
Catalog. Make it easy for users to see products. Never require users to register to see the product catalog. Avoid requiring users to select a city or enter a ZIP code to see a product catalog (e.g., Lowes.com). Instead, try to let users get directly into the product catalog without performing extra steps. To make it easy
28. Designing E-Commerce User Interfaces for users to browse for products, organize the product catalog the way users expect the products to be organized (e.g., organize clothing by gender). Because users only look at the first two or three pages of a list of products, allow users to narrow down a long list of products by using filtering tools (Nielsen & Tahir, 2001). For example, allow users to reduce a list of shoes by showing only a particular shoe size or style. On each product page, show a small image of the product and provide a link to a larger image. This technique reduces download times for product pages and increases buying rates (Millard, 2002). To avoid frustrating users, show only products that are in stock. If you cannot do this, clearly identify the color or product versions that are out of stock. If you cannot check a product's inventory status until users try to put it into the shopping cart, tell users the product is out of stock, when it will be in stock, suggest an alternate product (e.g., same size shirt in different color), and offer to send an e-mail note to users when the product is available. To prevent user confusion, never allow users to put out-of-stock products into the shopping cart. When users put a product into the shopping cart, update the cart summary on the page and provide a dialog window with a message such as "You successfully put the product into your shopping cart." In case users prefer to make a purchase later, provide a link on the product page to move the product into a wish list. Also, to improve sales, provide a link that allows users to e-mail the page to someone else who may purchase the product. Show one related cross-sell and one more expensive "up-sell" product on each product page. To give users the information they need to make a purchase decision, try to show a shipping cost for the product (e.g., Costco.com). Because 34% of users say a product comparison tool increases the chances they will buy from a site (eMarketer, 2001), allow users to select and compare products side by side on important, differentiating features (e.g., eBags.com's compare selected items tool in Fig. 28.4). Do not limit the number of products that can be compared. Provide links from the comparison back to each product's detail page. Make it easy for users to remove products from the comparison. To make the site more helpful, interesting, and "sticky," consider providing expert and customer reviews. Allow users to enter comments on products. Finally, because gift certificates are very popular, include them in the product catalog. Allow customers to e-mail gift certificates and to mail plastic stored value gift cards. Gift certificates can be very profitable because about 15% of gift certificates do not get redeemed (Tedeschi, 2002c).
Registration. The more streamlined the registration process, the more likely users will register and buy (Agrawal, Arjona, & Lemmens, 2001). Even when they are on it, most users will not read or complete the registration page, and one-half the users leave on each succeeding registration page (Sacharow & Mooradian, 1999). To reduce the number of user entries and to make it easier for users to remember their sign-in names, require users to enter only an e-mail address and a password. Ask users to make explicit clicks on check boxes to give permission to send e-mail notifications (e.g., sales, new products; Charron,
519
Bass, O'Connor, & Aldort, 1998) and to leave a cookie (e.g., "Remember me when I return"). Provide a link to the privacy and security policies. Also, because users have trouble entering their e-mail addresses correctly (Rehman, 2000), provide a sample, correct e-mail address ("For example: [email protected]"). Gather other user information several ways. Instead of requiring users to register before checking out, let unregistered users enter shipping and billing information during checkout. At the end of checkout, tell users the benefits of registering (e.g., quicker future checkouts, personalization, order history, wish list, sale notifications), ask users to register, retain the checkout information for registration, and ask users to provide only a password (Nielsen, 1999). To get more demographic and marketing information later, pop up optional, quick, one-question, multiple-choice surveys (e.g., gender, preferred products) when registered users revisit the site or when users complete the checkout process. Space out the surveys by several weeks or months. Give users control of their personal information. Allow users to edit the registration information and to unregister. For security reasons, never show the entire credit card number; show only asterisks and the last four digits. Personalization. Personalization takes the user's registration information, purchase history, and browsing history, then tailors the user interface. Personalization is very powerful. It can provide a more compelling user experience, reduce an overwhelming number of choices (e.g., reduce news stories to only categories that interest the user), and lead to higher conversion rates, more repeat visits, greater loyalty, and stronger brands (Agrawal et al., 2001; Cooperstein, Delhagen, Aber, & Levin, 1999; Reichheld & Schefter, 2000; Viant, 2000). A customer must buy from a site for two or three years for the site to recoup the cost of acquiring the customer (Reichheld & Schefter, 2000). Personalization can help make this happen. Personalize the news stories, promotions, tools, and other elements of the user interface (e.g., My Yahoo! at my.yahoo.com). One site (BirkenstockExpress.com) lets users see the shoes that are available in the users' size, price range, style, color, or material. Another site (LandsEnd.com) lets users build virtual mannequins for trying on clothes and getting style advice. Figure 28.5 shows the Lands' End personalized virtual model. Also, to make them more effective and less annoying, personalize the e-mail notifications that you send to registered users. Send notifications for only products and services that interest each registered user. Consider providing a way for users to be part of a community (e.g., Wine.com's "Wine Clubs"). This feature may improve user loyalty, user time on the site, and site revenue. After users join an online community, offer tailored information, tools, and message boards. Checkout. On the shopping cart page, provide users all the information they need to complete the purchase. Show hyperlinked product names so users can link back to see product details. Show entry fields with quantities that users can change. Display the price of each product and provide a drop-down menu of shipping choices with delivery periods and costs (e.g.,
520
•
NAJJAR
FIGURE 28.4. Product comparison tool for eBags.com.
"FedEx Ground, 5 business days, U.S. $5.00, Continental U.S. only"). Users want cost estimates before entering their credit card numbers. Fifty-six percent of users stopped checkout when they saw high shipping costs at the end of checkout (Hill, 2001). So, compute and display an order subtotal (including accurate shipping costs, plus taxes if the user is registered) at the bottom of the prices column. Allow users to control their shopping carts. Provide links for removing products, moving products into the user's wish list, returning to the prior shopping page, checking out, and a
button for refreshing the cart page. If registered members leave the site without checking out, automatically save their shopping cart contents for up to 90 days (e.g., Amazon.com). Figure 28.6 shows the BarnesandNoble.com shopping cart. In one study (Rehman, 2000), checkout was the reason 40% of users failed to complete an online order. Twenty-seven percent of users abandoned an order because the site required them to complete cumbersome forms (Sacharow & Mooradian, 1999). So, as described in the "Registration" section, do not require users to register before checking out (Rehman, 2000). Try to put all the checkout fields on a single, vertically scrollable
28. Designing E-Commerce User Interfaces
•
521
FIGURE 28.5. Personalized virtual model from LandsEnd.com. (2003 © Lands' End Inc. Used with permission. Copyright - My Virtual Model Inc., 2003.) page. This reduces user confusion and makes it easier to change the order. To allow users to review and change their selected products, provide a link from the checkout fields page back to the shopping cart page. Provide entry fields so registered members can conveniently sign in. If the users are signed-in members, automaticallyfillin the checkout fields using the members' registration information. If the billing address is the same as the shipping address, do not require new users to enter the same information again in the billing address fields. Instead, allow users to click a button near the billing address for "Use shipping address for billing address." Then, to keep the user interface very simple and obvious, refresh the page and show the shipping information in the billing address fields. If the e-commerce company also has a brick-and-mortar store, provide a check box so customers on the e-commerce site can elect to pick up online orders at the nearest store (e.g., Sears.com, CircuitCity.com; Tedeschi, 2002b). When users make an error or leave empty a required field, show the checkout fields page, put an obvious error message at the top (e.g., "Ooops! We had a problem processing this page. Will you please try to fix the fields that are marked with a red
'Problem?'"), then above each field that had an error show red "Problem" text and text that briefly explains what the trouble was and how to fix it. Make it easy to take users' money. Provide easy, safe, and reliable ways to pay. Design checkout to accept several payment methods (e.g., credit cards, debit cards, gift certificates, stored value gift cards). Twenty percent of users said they stopped an online purchase because they felt the site was not secure (Hill, 2001). Get the site's privacy and security certified by consumer groups, such as TRUSTe or BBBOnLine, and show their logos on the checkout page. Also, provide links that promote consumer protection features such as privacy policy, security protection, a no-questions-asked return policy, delivery guarantees, and customer service e-mail response time guarantees (Agrawal et al., 2001; Rhodes, 1998). After users enter their checkout information, provide a complete, read-only order summary. If users want to make a change, provide a "Change Order" button that takes users back to the single, vertically scrollable, editable, checkout information page. To avoid forcing users to retype information, retain the
522
NAJJAR
FIGURE 28.6. Shopping cart from BarnesandNoble.com. information users entered on the checkout information page. After the system accepts the order, provide an order confirmation that includes the order number, instructions for canceling the order, a hyperlink for tracking the order and shipment, and customer service information (e.g., e-mail, telephone number; Ragus, 2000a). Also, as described earlier, ask unregistered users if they would like to register. Do not show promotions during
checkout because they take users away from the checkout process and interrupt the process of paying. If a user is signed in to the site, and the user wants to change the billing or shipping addresses in the user registration, require the user to enter the user's password. Otherwise, when the user leaves the computer signed-in to the site (e.g., via cookie), someone else could make purchases with the registered member's
28. Designing E-Commerce User Interfaces payment information and send the purchases to a different address. Customer Service. Customer service on e-commerce sites is often inadequate. Fifty-five percent of e-commerce sites do not provide online customer service, such as e-mail, message boards, and live chat (Martin, 2002; Vigoroso, 2002). Twenty-five percent of customer service centers could not answer shopper questions (Gunderson, 2000). Thirty-seven percent of Fortune 100 companies did not answer requests for information submitted via their Web sites (TechWeb News, 2002). The average response time to a customer query at 50 sites was 12 hours (Hirsch, 2002). Clearly, there is a need for Web sites to improve online customer service. Customer service is especially important for ecommerce sites, where a poor customer service experience can quickly drive customers to competitors' sites. So, design the site to provide outstanding customer service. On the "Contact Us" page, provide a toll-free telephone number, e-mail address, facsimile number, and mailing address. The best support is available 24 hours a day, seven days a week, in real time (Agrawal et al., 2001; InternetWeek, 2002). So, try to provide real-time chat with customer service (e.g., Godiva.com Live Assistance). Real-time chat allows modem users to stay connected to the site while getting help from customer service. Also provide FAQs (questions and answers to frequently asked questions) on topics such as checkout, privacy, security, and returns. Forty-two percent of users said they would buy more online if the return process were easier (Broadbase, 2001). If the online site has brick-and-mortar stores, allow customers to return online purchases to the stores. With each shipped order, provide a preprinted shipping label (and possibly a mailing envelope) so it is easy for users to mail a return (e.g., Nordstrom.com). Allow users to cancel an unshipped order. To keep customers updated, send e-mail notifications when the shipping status changes. Provide helpful tools such as wish lists, gift registries, and online order histories. If you want customers to complete a satisfaction questionnaire, present the questionnaire at the end of the purchase process, when customers are most ready to provide feedback (Hirsch, 2002). Access for Users with Disabilities. More than one half a billion people are disabled (United Nations, 2002). In the United States, 20% of the population experience some form of disability (McNeil, 1997; U.S. Census Bureau, 2000). Almost everyone will experience a temporary disability due to illness, accident, or circumstance (Martin, 2002; McNeil, 1997). The Americans with Disabilities Act of 1990 (U.S. Department of Justice, 1990; 28 C.F.R. Sec. 36.303; 28 C.F.R. Sec. 35.160) and section 508 of the Rehabilitation Act Amendments of 1998 (General Services Administration, 2003) require that electronic and information technology (e.g., Web pages and software applications) for U.S. federal sites be accessible to people with disabilities, unless doing so would impose an undue burden on the department or agency. Many state and city governments have similar laws (Williams, 2001). Other countries and international organizations also encourage or require accessibility of Web sites (Commonwealth of Australia, 2003;
•
523
Department of the Taoiseach, 1999; European Union, 2003; International Organization for Standardization, 2003; New Zealand E-government, 2003; Treasury Board of Canada, 2002). In the United States, accessibility is not required for private, e-commerce Web sites (Schwartz, 2002). However, accessibility is good business. By designing your site to be accessible (see, e.g., Stephanidis & Akoumianakisc, chap. 13, and Vanderheiden, chap. 15, this volume), you increase the number of people who can purchase from the site. Plus, the simple designs required for accessibility often improve ease of use and work well on other devices such as Web-enabled cellular telephones and personal digital assistants. To improve accessibility, use relative (rather than absolute) font sizes (with style sheets, use percentages of the default font size rather than exact font size values), provide equivalent alternatives to visual and auditory content (e.g., use ALT or LONGDESC tags to describe the meaning of an image; Nielsen, 1996; Vigoroso, 200Ic). Use client-side image maps with ALT tags for each image area. Do not use server-side image maps because they send only image coordinates, not ALT text. To accommodate users who are blind or color-deficient and printers and displays that are monochrome, do not rely on color alone to differentiate objects (Najjar, 1990). Ideally, use highcontrast black text on a white background. Do not use font style or size as the only way to communicate information. To avoid forcing users of screen readers to repeatedly hear the same major navigation controls on every page, put on every page a hidden hypertext link with an ALT tag that is "Jump to Main Content" (Roblimo, 1999). Associate the hypertext link with an image, such as a single, background-colored pixel, that appears near the top of every page. Link to the beginning of the unique content in each page. Sighted users will not notice the image or the ALT tag, but screen readers will read it for users with visual challenges. Screen readers, such as Freedom Scientific's JAWS (http:// www.freedomscientific.com/fs_products/software_jaws.asp), use the alternative text for images and read tables from left to right one complete cell at a time. This means that users with visual challenges have to listen to all the navigation control names in the left column of a layout table before getting to the main content in the middle column. To allow the screen reader to more quickly get to the content area of a layout table, leave empty the first cells in the left navigation control column and the right promotions column. Use a single column in the content area. This design technique causes the screen reader to jump to the main content in the center cell of the layout table. For information tables, use row and column headers (TH tag). Do not use multiple logical levels of row or column headers. If you must use multiple logical levels, then assign ID attributes to each column and row header, then for each cell use the HEADER attribute to call the column and row ID attributes for each cell (Thatcher, 2001 a). To help users understand the purpose of a table, use the CAPTION tag to identify the table title and the SUMMARY tag to provide a summary of a complex table. For frames, provide descriptive NAME and TITLE tags for each frame because some accessibility technologies read only frame NAME tags and others read only frame TITLE tags.
524
NAJJAR
Allow users with cognitive challenges to stop automatic blinking, moving objects, text scrolling, page updates, or other elements that change quickly over time (e.g., allow users to stop text that moves across a page; Solomon, 2000; Trace Research & Development Center, 2003; Vanderheiden, chap. 15, this volume; W3C, 1999). To allow users with motor challenges to interact with your site, do not require users to make fine cursor movements (e.g., selecting small parts of an imagemap; Nielsen, 1996). Make all functions available via keyboard only. Carefully use JavaScript, Flash, and Java so the information they provide is accessible, even when users turn off scripting and Flash (Sun, 2003b; Thatcher, 200 Ib). For Java, make sure users can tab to objects in a logical order, provide helpful alternative object names rather than taking the Java defaults, and do not hard-code colors, font styles, font sizes, lines, borders, or shadow thickness (Feigenbaum, 2002; Sun, 2003a). Check the accessibility of your site. Turn off images, scripting, and style sheets, then interact with the site using only the keyboard. Use Lynx (a text-only brower available at http://lynx. browser.org) and IBM Home Page Reader (get a 30-day trial at ibm.com). Validate the site with Bobby (http://bobby.watchfire. com/bobby/html/en/index.jsp), the Wave (http://www.temple, edu/instituteondisabilities/piat/wave/), Lift (http://www.usa blenet.com/), or some of the other free tools. Evaluate the accessibility of the user interface via checklists and tests with representative persons with disabilities (Fain & Folds, 2002).
brand identity and stability, widen your potential client base, and reduce customer service costs. Because translated text may take up 40% to 200% of the space of English text (eTranslate, 2000a, 2000b), create a site that is built dynamically from databases, design page widths to be flexible (specify table cell widths in percentages rather than pixels), and avoid putting text on graphics because you have to redesign the graphics. Do not use slang, colloquialisms, bold, or italics because they translate poorly. Be careful with colors (e.g., black is the color for mourning in the United States, but it is white in Japan, so white text could disturb Japanese users). Avoid using images that could be offensive (e.g., the "thumbs up" image is positive in the United States, but offensive in Latin America). Use the "accept-language" parameter so the browser shows the page in the user's preferred language. Because users may arrive at a page via a search or link without going through the home page, provide a way to change the language on every page. Do not show choices using country flags because users may speak a language that is not the dominant language for the country. Instead, show the language choice in the language (e.g., "Espanol," "Francais"; eTranslate, 2000a). If your site is large and changes frequently, it may be more appropriate and easier to develop a separate site for each language (Vigoroso, 2001a). Most important, perform international usability tests of the localized versions of your sites.
CONCLUSIONS Multiple Languages. Internet users are four times more likely to buy from sites that support their preferred language (Vickers, 2000). However, more than 70% of the current 534 million Web sites are in English, whereas only 8% of the world's population (Vickers, 2000) and less than one half of current Internet users are English speakers (Global Reach, 2003; Vigoroso, 200la). United States-based e-commerce sites currently dominate the Web, but by 2004, 50% of all online sales will take place outside the United States (Schmitt, Cooperstein, Tong, & Li, 2000). So, create sites that are globalized and localized. A globalized site is culturally and technically neutral. Do not use jargon, symbols, or abbreviations. Do not use streaming media because users in countries that have narrow bandwidth, such as China, will not be able to use this technology effectively. A localized site meets the cultural (e.g., rules for formality, use of colors, symbols, and graphics), linguistic (e.g., language, format of dates, direction of text), technical requirements, and business practices (e.g., local laws, regulations, shipping, currency, taxes, payment, and copyrights) of a specific location (Choong, Plocher, & Rau, chap. 16, this volume). When you create a localized site, you can create stronger customer loyalty, improve
To be successful, e-commerce user interfaces must be very easy to use. To improve usability, follow an efficient, iterative, usercentered design process. Use proven detailed designs for browsing and searching the product catalog, registering, personalizing, and checking out of an e-commerce site. Design for accessibility and for multiple languages.
ACKNOWLEDGMENTS Lawrence Najjar is a user interface designer with more then 20 years of experience. He worked on AOL's corporate information site, Home Depot's online store, NASCAR.com, and the U.S. air traffic control system. Lawrence has a Ph.D. in engineering psychology from the Georgia Institute of Technology. Portions of this chapter appeared in Najjar, L. J. (2001). E-commerce user interface design for the Web. In M. J. Smith, G. Salvendy, D. Harris, & R. J. Koubek (Eds.), Usability evaluation and interface design (Volume 1 of the Proceedings of HCI International 2001; pp. 843-847). Mahwah, NJ: Lawrence Erlbaum Associates.
References ActivMedia Research. (2000). Consumer loyalty online research series [On-line]. Available: http://www.activmediaresearch.com/ consumerJoyalty_online.html
Agrawal, V, Arjona, L. D., & Lemmens, R. (2001). E-performance: The path to rational exuberance. The McKinsey Quarterly, 1. Available: http://www.mckinsey.com/
28. Designing E-Commerce User Interfaces American Heritage Editors. (2000). American Heritage Dictionary of the English Language (4th ed.). [Entry for Henry Louis Sullivan.] Boston: Houghton Mifflin. Available: http://www.bartleby.com/ 6l/74/S0877400.html BatteyJ. (1999, April 19). IBM's redesign results in a kinder, simpler Web site. Infaworld [On-line]. Available: http://www.infoworld.com/cgibin/displayNew.pl?/hotsites/hotextra990419.htm Benway, J. P, & Lane, D. M. (1998, December). Banner blindness: Web searchers often miss "obvious" links. Internetworking [Online], 1.3. Available: http://www.internettg.org/newsletter/dec98/ banner_blindness.html Bernard, M. (2001, Winter). Developing schemas for the location of common Web objects. Usability News [On-line], 3.1. Wichita, KS: Software Usability Research Laboratory, Wichita State University. Available: http://wsupsy.psy.twsu.edu/surl/usabilitynews/3W/ web_object.htm BizRate. (2000, October 23). 78% of Online shoppers abandon shopping carts according to BizRate survey. BizRate press release [On-line]. Available: http://www.bizrate.com/content/press/release. xpml?rel=88 Broadbase. (2001, February 20). Online shopping returns survey shows some improvement, but overall customer care still lacking. Broadbase Software press release [On-line]. Available: http://www.kana. com / cgi-bin/ kana/ kana/ jsp/newsevents/press_pop. jsp?MainDoc= /cgi-bin/kana/kana/jsp/newsevents/archive/PressReleases2001/ 2001 _0220_survey.jsp Chaparro, B. S. (2001, December). Top ten mistakes of shopping cart design. Internetworking [On-line], 4.1. Available: http://www. internettg.org/newsletter/dec01/article_chaparro.html Charron, C, Bass, B., O'Connor, C., & Aldort, J. (1998, July). The Forrester report: Making users pay [On-line]. Available: http://www. forrester.com/. Cohen, J., & Thompson, M. J. (1999, February). Mass appeal. The Standard [On-line]. Available: http://www.thestandard.com/article/ display/0,1151,4927,00.html Commonwealth of Australia. (2003). National Office for the Information Economy. Accessibility [On-line]. Available: http:// www.govon-line.gov.au/projects/egovernment/better%5Fpractice/ accessibility.htm Consumer Reports. (2003). E-ratings: A guide to online shopping, services, and information: What we look for [On-line]. Available: http: //www.consumerreports.org/main / detailv2.jsp?CONTENT% 3C%3Ecnt_id = 871&FOLDER%3C%3Efolder_id=735&bmUID= 1043273020029 Cooperstein, D. M., Delhagen, K., Aber, A., & Levin, K. (1999, June). Making Net shoppers loyal. Forrester Report: Interviews [On-line]. Available: http://www.forrester.com/ Department of the Taoiseach. (1999, October). Report of the Interdepartmental Group—Recommended guidelines for public sector organisations [On-line]. Available: http://www.taoiseach.gov.ie/ viewitem.asp?id=291& lang=ENG eMarketer. (2001, March 12). Turning shoppers on(line). eMarketer [On-line]. Available: http://www.emarketer.com/estatnews/estats/ ecommerce_b2c/20010312_pwc_search_shop.html eMarketer. (2003a, January 3). BizRate estimates e-retail sales hit 17.44B in Q4. eMarketer [On-line]. Available: http://www.emarketer.com/ news/article.php? 1001946 eMarketer. (2003b, April 30). US B2C e-commerce to top $90 billion in 2003. eMarketer [On-line]. Available: http://www.emarketer. com/news/article.php?1002207 ePaynews. (2003a). Statistics for electronic transactions. Average transaction size by year started using Internet (ActivMedia Research). Total B2C revenues for US, Europe & Asia, 1999-2003
•
525
(Fortune Magazine). Comparative estimates of Latin eCommerce, 2000-2003. Pure online transaction sales, 2000-2002 [On-line]. Available: http://www.epaynews.com/statistics/transactions.html ePaynews. (2003b). Statistics for online purchases. Predicted online retail sales worldwide, Q4 2002 (GartnerG2). US online retail revenues, 2001-2005 (eMarketer). B2C eCommerce forecast for Europe, 2001-2005 (AMR Research). Percentage of abandoned shopping carts, 2001 (eMarketer). Proportion of abandoned transactions at Web storefronts (Datamonitor). Number of online US and worldwide purchasers, 1998-2003 (Donaldson, Lufkin, &Jenrette). European online market in 2001 to equal USD 3.23 billion (Datamonitor) [On-line]. Available: http://www.epaynews.com/statistics/ purchases.html eTranslate. (2000a, April). 16 Steps toward a truly global Web site: A primer on preparing Web sites for translation and localization [On-line]. Available: http://www.etranslate.com/en/know/know. T3.html eTranslate. (2000b, May). Globalization: Web design overview. [Online] . Available: http://www.etranslate.com/en/know/know.T21 .html European Union. (2003). EUROPA-European Commission-Information providers' guide—Rule no. 7: Accessibility [On-line]. Available: http://europa.eu.int/comm/ipg/rule7/rule7_en.htm*description Fain, W. B., & Folds, D. J. (2002). An approach to measuring accessibility of office automation equipment. In Proceedings of the Human Factors and Ergonomics Society 46th annual meeting (pp. 19821986). Santa Monica, CA: Human Factors and Ergonomics Society. Feigenbaum, B. A. (2002, October 1). Coding for accessibility—Use JFC/Swing to build accessibility into your Java applications [Online]. Available: http://www-106.ibm.com/developerworks/java/ library/j-access/ Fleming, J. (1998). Web navigation: Designing the user experience. Sebastopol, CA: O'Reilly. Fuccella, J., & Pizzolato, J. (1999, March). Separating content from visuals in Web site design. Internetworking [On-line]. Available: http://www.internettg.org/newsletter/mar99/wireframe.html General Services Administration. (2003). 1998 amendment to section 508 of the Rehabilitation Act [On-line]. Available: http://www. section508.gov/index.cfm?FuseAction=Content& ID=l4 Global Reach. (2003). Global Internet statistics (by language) [Online]. Available: http://global-reach.biz/globstats/index.php3 Graphics, Visualization, and Usability Center. (1998). GVU's 9th WWW user survey [On-line]. Atlanta: Georgia Institute of Technology, College of Computing, Graphics Visualization, and Usability Center. Available: http://www.gvu.gatech.edu/user-surveys/survey1998-04 Grimm, J., & Grimm, W. (1999). Complete fairy tales of the brothers Grimm. Minneapolis: Econo-Clad. Guernsey, L. (2001, February 28). Revving up the search engines to keep the e-aisles clear. The New York Times, p. 10. Gunderson, A. (2000, September 4). Toil and trouble: Online shopping is still a muddle. Fortune [On-line]. Available: http://www.fortune. com/fortune/personalfortune/articles/0,15114,37024,00.html Hagen, P. R., Manning, H., & Paul, Y. (2000, June). Must search stink? Forrester Report [On-line]. Available: http://www.forrester.com/ ER/Research/Report/0,1338,94l2,00.html Hansell, S. (2002, December 3). Net shopping begins surge for holidays. The New York Times, p. Cl. Harris Interactive. (2003, February 5). Those with Internet access to continue to grow but at a slower rate. Harris Interactive [On-line]. Available: http://www.harrisinteractive.com/news/allnewsbydate. asp?NewsID=579 Hill, A. (2001, February 12). Top 5 reasons your customers abandon their shopping carts (and what you can do about it). ZDNet
526
NAJJAR
[On-line]. Available: http://www.zdnet.com/ecommerce/stories/ main/0,10475,2677306,00.html Hirsch, L. (2002, September 10). E-tail customer service: Finally working? E-commerce Times [On-line]. Available: http://www. ecommercetimes.com/perl/story/19353.html International Organization for Standardization. (2003). Ergonomics of human-system interaction—Guidance on accessibility for human-computer interfaces (ISO/TS 16071:2003) [On-line]. Available: http://www.iso.ch/iso/en/CatalogueDetailPage.Catalogue Detail?CSNUMBER=30858& ICS1=13& ICS2=180& ICS3= InternetWeek. (2002, October 29). Many Fortune 100s ignore Web inquiries. InternetWeek, TechWeb News [On-line]. Available: http://www.internetwk.com/webDevANW20021029S0011 Kapadia, R., & Moore, A. (2002, December 19). In dismal retail season, online shopping shines. Technology—Reuters Internet Report, Yahoo! News. Lake, D. (2000, April 17). Navigation: An e-commerce must. The Standard [On-line]. Available: http://www.thestandard.com/research/ metrics/display/0,2799,14110,00.html Manning, H., McCarthy, J. C, & Souza, R. K. (1998). Interactive technology strategies: Why most Web sites fail [On-line] 3(7). Available: http ://www.forrester. com/ Martin, M. (2002, September 17). Accessibility breakthroughs broaden Web horizons. Sci.NewsFactor.com. Available: http://sci. newsfactor.com/perl/story/194l7.html McNeil, J. M. (1997). Americans with disabilities: 1994-95. U.S. Department of Commerce, Economics, and Statistics Administration [Online]. Available: http://www.census.gOv/prod/3/97pubs/p70-6l.pdf Millard, E. (2002, July 18). Bare-bones e-commerce versus bells and whistles. E-commerce Times [On-line]. Available: http://www. ecommercetimes.com/perl/story/18646.htnil Miller, R. B. (1968). Response time in man-computer conversational transactions. Proceedings of American Federation of Information Processing Societies Fall Joint Computer Conference, 33, 267277. Najjar, L. J. (1990). Using color effectively (IBM TR52.0018) [Online]. Atlanta: IBM Corporation. Available: http://www.lawrencenajjar.com/papers/Using_color_effectively.html Najjar, L. J. (1999, June). Beyond Web usability. Internetworking, 2.2 [On-line]. Available: http://www.internettg.org/newsletter/jun99/ beyond_web_usability.html Najjar, L. J. (2000, December). Conceptual user interface: A new tool for designing e-commerce user interfaces. Internetworking, 3.3 [On-line]. Available: http://www.internettg.org/newsletter/decOO/ article_cui.html Najjar, L. J. (2002). An efficient Web user interface design process for commercial clients [On-line]. Available: http://www.lawrencenajjar.com/papers/An_efficient_Web_user_interface_design_process. html New Zealand E-government. (2003). Web guidelines [On-line]. Available: http://www.e-goverriment.govt.nz/web-guidelines/ Nielsen, J. (1994). Response times: The three important limits [Online]. Available: http://www.useit.com/papers/responsetime.html Nielsen, J. (1996). Accessible design for users with disabilities [Online]. Available: http://www.useit.com/alertbox/96lO.html Nielsen, J. (1997). The need for speed [On-line]. Available: http://www. useit.com/alertbox/9703a.html Nielsen, J. (1999). Web research: Believe the data [On-line]. Available: http://www.useit.com/alertbox/990711 .html Nielsen, J. (2000). Flash: 99% bad [On-line]. Available: http://www. useit.com/alertbox/20001029.html Nielsen, J., & Tahir, M. (2001, February). Building sites with depth. In New Architect Magazine (formerly Webtechniques). Available:
http://www.newarchitectmag.com/documents/s=4570/newl0136 37059/index.html Orr, A. (2002, October 2). NetTrends: Many still happy living without broadband. Technology—Reuters Internet Report. Yahoo! News. Pastore, M. (1999, February 25). Online stores lacking. http://www. clickz.com/stats/markets/retailing/article.php/154021 Ragus, D. (2000a). Best practices for designing shopping cart and checkout interfaces [On-line]. Available: http://www.dack.com/ web/shopping_cart.html Ragus, D. (2000b). Flash is evil [On-line]. Available: http://www.dack. com/web/flash_evil.html Ragus, D. (2000c). Flash vs. HTML: A usability test [On-line]. Available: http://www.dack.conVweb/flashVhtml/ Rehman, A. (2000, November 22). Effective e-checkout design. ZDNet/ Creative Good [On-line]. Available: http://techupdate.zdnet.com/ techupdate/stories/main/0,l4l79,2638874-l,00.html Reichheld, F. E, & Schefter, P. (2000, July-August). E-loyalty: Your secret weapon on the Web. Harvard Business Review, 105113. Rhodes, J. S. (1998, December 29). 8 Quick tips for a more usable ecommerce Web site. Webword [On-line]. Available: http://webword. com/moving/8quick.html Roblimo. (1999, December 10). Interview: Answers about blind computer use. Slashdot [On-line]. Available: http://slashdot.org/article. pl?sid=99/l2/09/1342224& mode=thread Rogers, B. L., & Chaparro, B. (2003). Breadcrumb navigation: Further investigation of usage. Usability News, 5(2) [On-line]. Available: http: //psychology. wichita.edu/surl/usabiUtynews/52/breadcrumb. htm Romero, S. (2002, December 5). Price is limiting demand for broadband. The New York Times, p. C5. Sacharow, A., & Mooradian, M. (1999, March). Navigation: Toward intuitive movement and improved usability. Jupiter Communications [On-line]. Available: http://www.jup.com Schmitt, E., Cooperstein, D. M., Tong J., & li, C. (2000). The multilingual site blueprint. Forrester Report [On-line]. Available: http:// www.forrester.com Schwartz, J. (2002, October 22). Judge rules on Web sites for the disabled. The New York Times, Technology Briefing, p. CIO. Selingo, J. (2000, August 3). A message to Web designers: If it ain't broke, don't fix it. The New York Times, p. Dll. Seminerio, M. (1998, September 10). Study: One in three experienced surfers find online shopping difficult. Inter@ctive Week [On-line]. Available: http://zdnet.com.com/2100-l l-512120.html Solomon, K. (2000, July 3). Disability divide. The Standard [On-line]. Available: http://www.ksolomon.com/disadiv.htm Stanley, J., McCarthy, J. C., & Sharrard, J. (2000, May). The Internet's privacy migraine [On-line]. Available: http://www.forrester.com/ Sullivan, T. (1998). The need for speed. Site optimization strategies. All things Web [On-line]. AvaUable: http://www.pantos.org/ atw/35305.html Sun. (2003a). Accessibilty quick reference guide [On-line]. Available: http://www.sun/com/access/developers/access.quick.ref.html Sun. (2003b). Java accessibility quick tips [On-line]. Available: http://www.sun.com/access/developers/tips.html TechWeb News. (2002, November 5). Many Fortune 100s ignore Web inquiries. InternetWeek [On-line]. Available: http://www. internetweek.com/webDev/INW20021029S0011 Tedeschi, B. (1999, August 30). Good Web site design can lead to healthy sales. The New York Times e-commerce report [On-line]. Available: http://www.nytimes.com/library/tech/99/08/cyber/commerce/ 30commerce.html
28. Designing E-Commerce User Interfaces
Tedeschi, B. (2002a, November 18). E-tailers are putting the multimedia tinsel back on their shopping sites this holiday season. The New York Times e-commerce report, p. C6. Tedeschi, B. (2002b, December 23). Online sales are up for holiday, but the question is just how much. The New York Times, pp. Cl, C7. Tedeschi, B. (2002c, December 30). There is more cheer in the virtual realm as online gift certificates soar. The New York Times e-commerce report, C5. Thatcher, J. (200la). Web accessibility for Section 508—Review of Section 508 standards for Web Accessibility [On-line]. Available: http://www.jimthatcher.com/webcoursec.htm Thatcher, J. (2001b). Web accessibility for Section 508—Review of section 508 standards for Web Accessibility [On-line]. Available: http://www.jimthatcher.com/webcourse4.htm Thompson, M. J. (1999, August 9). How to frustrate Web surfers. Industry Standard [On-line]. Available: http://www.thestandard. com/metrics/display/0,1283,956,00.html Trace Research & Development Center, College of Engineering, University of Wisconsin-Madison. (2003). Designing more usable Web sites [On-line]. Available: http://www.trace.wisc.edu/world/ web/index.html Tracy, B. (2000, August 16). Easy net navigation is mandatory— Viewpoint: Online users happy to skip frills for meat and potatoes. Advertising Age, p. 38. Treasury Board of Canada. (2002). CLF—Accessibility—Standards and guidelines [On-line]. Available: http://www.cio-dpi.gc.ca/clfupe/1/la.e.asp UNCTAD (United Nations Conference on Trade and Development). (2002). E-commerce and development report 2002 [On-line]. Available: http://www.unctad.org/en/docs//sdteecb2sum_en.pdf United Nations. (2002). The UN and persons with disabilities: United Nations commitment to advancement of the status of persons with disabilities [On-line]. Available: http://www.un.org/esa/ socdev/enable/disun.htm U.S. Census Bureau. (2000). Disability status of persons (SIPP) [Online]. Available: http://www.census.gov/hhes/www/disable/sipp/ disstathtml
527
U.S. Census Bureau. (2003). Total midyear population for the world: 1950-2050 [On-line]. Available: http://www.census.gov/ipc/www/ worldpop.html U.S. Department of Justice. (1990). Americans with Disabilities Act of 1990 [On-line]. Available: http://www.ada.gov/pubs/ada.txt Viant. (2000). Experience architecture: The strategy and design of ecommercepersonalization. Viant Innovation Center Project. Available: [email protected] Vickers, B. (2000, November 22). Firms push to get multilingual on the Web. The Wall Street Journal, p. B11A. Vigoroso, M. W (200la, November 2). E-tailers: Globalize with caution. E-commerce Times [On-line]. Available: http://www. ecommercetimes.com/perl/story/l4546.html Vigoroso, M. W. (2001b, November 14). The bottom line in Web design: Know your customer. E-commerce Times [On-line]. Available: http://www.ecommercetimes.com/perl/story/l4738.html Vigoroso, M. W (2001c, December 21). Blind online shoppers deserve attention. E-commerce Times [On-line]. Available: http://www. ecommercetimes.com/perl/story/15453-html Vigoroso, M. W. (2002, January 28). Report: Miles to go before customers rule online. E-commerce Times [On-line]. Available: http:// www.ecommercetirnes.com/perl/story/l6023.html W3C. (1999, May 5). Web content accessibility guidelines 1.0 [On-line]. Available: http://www.w3.org/TR/WAI-WEBCONTENT/-Guidelines Weaver, J. (2003, June 20). European e-commerce on the rise. MSNBC News [On-line]. Available: http://ytlcommunity.com/commnews/ shownews. asp?newsid=6725 Williams, J. (2001, September 7). Making Uncle Sam accessible—and accountable. Business Week Online [On-line]. Available: http:// www. businessweek. com/smallbiz/content /sep2001 /sb2001097_ 766.htm Wixon, D., & Jones, S. (1992). Usability for fun and profit: A case study of the design of DEC RALLY version 2. Internal report, Digital Equipment Corporation. Cited in Karat, C., A business case approach to usability cost justification. In Bias, R. G., & Mayhew, D. J. (1994). Cost-justifying usability. San Diego: Academic Press.
29: INTRANETS AND INTRAORGANIZATIONAL COMMUNICATION V. Kathlene Emery ]ulie A. ]acko Francois Sainfort Kevin P. Moloney Georgia Institute of Technology
INTRODUCTION Intranets capitalize on World Wide Web technology within the boundaries of an organization. Among the various types of information technologies used within organizations, intranets are poised to make the largest impact, especially in terms of organizational learning. Since their infancy (circa 1995), intranet technologies have evolved from being viewed as internal information publication and posting tools, to more sophisticated organizational knowledge management and learning platforms. The power of intranet technology is illustrated by its ability to restructure and fortify internal communications throughout an organization. Interconnectivity of people and organizational entities strengthens the communications network to promote innovation, collaborative work, and organizational learning. The human element is essential to the success of an intranet, as the system's users contribute the knowledge that is necessary for the dynamic expansion of the company's learning organization. In turn, information technology (IT) tools can facilitate or induce changes in an individual's work, as well as impact the productivity, effectiveness, quality, and job satisfaction within an organization. With respect to the human-technology relationship, instability may lead to undesirable outcomes, ultimately transforming the intranet technology into an ineffective
528
(and costly) IT solution (Nielsen, 1997). From a human factors perspective, the integration of IT -within an organizational structure must consider the attributes of its users, rules, goals, and structures within the context of the organization. At the onset of 1996, popular business/technology press, such as Business Week, touted the advent of intranets (with statements such as "Here comes the Intranet" and "No questions, intranets are coming"). In the spring of 1996, CNet noted the emergence of a new penchant in the technical community for corporate intranet applications, rather than consumer-based applications. These "internal World Wide Webs" or "corporate Internets" were poised to take over the role of groupware applications and revolutionize the way organizations were doing business. These predictions of the forthcoming impact of intranet technology have come to fruition, albeit within a limited capacity. This chapter focuses closely on the interplay between intranets and intraorganizational communication and how an organization can exploit the benefits of these information technologies. Specifically, we will integrate the concepts of intraorganizational communication and intranet technologies to examine how these concepts can augment or impede organizational growth. We will then discuss how principals of organizational structure and user-centered design relate to the planning, design, and support of intranets for intraorganizational communication.
29. Intranets and Intraorganizational Communication
KEY CONCEPTS IN ORGANIZATIONAL LEARNING The quest for organizational learning has prompted significant reform within organizations. Akin to intranet, the topic of organizational learning has received much attention in the popular management and IT media. Unlike intranets, however, organizational learning is accompanied by a substantial amount of theoretical and academic research. This section presents an overview and extraction of key concepts associated with organizational learning that influence the development of intraorganizational communication and the role of intranets within organizations. The concept of organizational learning has been prominent in the organizational theory literature since at least the early 1960s. It appeared several decades ago in the developmental stages of organizational behavioral theory (e.g., see Cyert & March, 1963). Although most organizational theorists and practitioners currently agree that organizational learning is an essential factor in the evolution of organizations, there is still no universally accepted definition for this term. Miner and Mezias (1996) noted that, despite being quite sizable, the literature base on organizational learning suffers from a lack of precise definition, measurement, and estimation. Definitions of organizational learning can be broadly categorized into one of two overall perspectives—the behavioral perspective or the cognitive perspective. The behavioral perspective states that an entity learns if, through its processing of information, experience, and knowledge, the range and nature of the entity's potential and realized behaviors are changed (e.g., see Cangelosi & Dill, 1965; Levitt & March, 1988). Therefore, the behavioral tradition views learning as the result of change in behavior, with learning taking place through the modification of an organization's programs, goals, decision rules, or routines. However, the cognitive tradition views organizations mainly as interpretive systems. Hence, learning is considered to be a process of revising assumptions through reflection and through continuous interpretation of the environment (e.g., see Argyris & Schon, 1978; Daft & Weick, 1984; Senge, 1990). March and Olsen (1975), recognizing the importance of bounded rationality in adaptation and organizational learning, further proposed that organizational intelligence is built on two fundamental processes. The first is a process of "rational calculation, by which expectations about future consequences are used to choose among current alternatives The second process is learning from experience" (pp. 147-148). Argyris (1982) denned organizational learning as a process of detection and correction of errors, and proposed that organizations learn through individuals acting as agents of learning. He also noted that an organization, itself, operates as an ecological system, which defines the environment in which individuals' learning activities take place, and hence can influence the process and nature of learning. Levitt and March (1988) viewed the organizational learning process as a result of encoding inferences from history into routines that can guide behaviors. This definition is built on three observations. First, behavior in an organization is based on routines, whereby individuals in organizations behave and
•
529
make decisions based on a process of matching procedures to recognized situations, rather than through using logic of consequentiality of intention. Second, organizational actions are history dependent, based on the interpretation of the past rather than anticipations of the future. Finally, organizations are target oriented, meaning that behavior mainly depends on relations between observed outcomes and aspirations for these outcomes. Weick (1991) argued that the defining properties of learning reside in different responses being generated from the same stimulus. He suggested that few organizations have such properties. In fact, he suggested that most organizations are not built to learn—rather, they are designed to produce the same response to different stimuli, a pattern that may conflict with the very essence of learning. He concluded that organizations do not learn in this traditional, intended way. Miner (1990) further noted that most organizational learning models assume that aspirations and targets precede and drive experimentation. He also demonstrated that organizations can change and adapt over time without top-down, goal-driven variation, and proposed an evolutionary model to organizational learning. Therefore, although organizational learning can be intended and designed, it can also be unintended and evolutionary. In fact, experts sometimes make a distinction between "organizational learning" and "learning organizations." Although some authors do not make this distinction, or consider that organizational learning refers to the process of learning and learning organization refers to the structure in which learning takes place, others see important distinctions. An important characteristic of learning organizations is the adoption of "generative" or "double-loop" learning, "adaptive" instead of "single-loop" learning (Argyris, 1982; Senge, 1990). Double-loop learning focuses on solving problems in the present without examining the appropriateness of current learning behaviors. Therefore, learning organizations continuously learn and unlearn (Hedberg, 1976), designing and redesigning themselves through experimentation and feedback. In a comprehensive synthesis of organizational learning, Huber (1991) emphasized that learning is not necessarily conscious or intentional, does not always increase the learner's effectiveness, and does not always result in observable changes in behavior. In his review, rather than defining organizational learning, he adopted a more operational perspective and defined four important concepts of organizational learning that provide a useful framework for designing and improving organizational learning activities: • Knowledge acquisition: The process by which knowledge is obtained. • Information distribution: The process by which information is shared. • Information interpretation: The process by which distributed information is given one or more interpretations. • Organizational memory: The means by which knowledge is accumulated and stored for the future. Huber (1991) asserted that the construct of organizational memory is central to the idea of organizational learning because the basic processes that contribute to the occurrence-recurrence
530
EMERY ET AL.
(knowledge acquisition), breadth (information distribution), and depth-richness (information interpretation) of organizational learning depend on organizational memory. Therefore, these four constructs are linked to one another as shown in Fig. 29.1. In a more recent text on organizational learning, Argote (1999) focused on "factors explaining organizational learning curves and the persistence and transfer of productivity gains acquired through experience." She reported that organizational learning might explain significant performance variations that are evident at the firm level of analysis. She identified three broad categories of organizational factors that appear to influence the rate at which organizations learn and their subsequent productivity: (a) proficiency of individuals; (b) technology; and (c) an organization's routines, structures, and processes. Consequently, the concepts of organizational memory, knowledge, and information are central to organizational learning. Organizational memory can be thought of as the repositories for knowledge acquired through experience and other means. As described by Cross and Baird (2000), such repositories consist of (a) the minds of individual employees; (b) the relationships between employees; (c) paper, electronic databases, and information; (d) work processes and technologies; and (e) products or services offered. Therefore, organizational memory takes many forms, from memory "embrained" in individuals to memory embedded in technologies, structures, and routines. Intuitively, the transmission of the many forms of organizational memory merits further attention.
FIGURE 29.1. Four constructs of organizational learning (Huber, 1991).
According to Hills (1998), capturing knowledge and expertise so a larger community can reuse the information is the foundation of the knowledge management infrastructure. To this end, intraorganizational communication and its supportive tools need to fit well with previously existing people, processes, and content to encourage organizational growth and expansion. To characterize and understand intraorganizational communication, the classification of shareable knowledge is valuable. Lam (2000) proposed a typology of knowledge that integrates organizational and cognitive dimensions of knowledge. The organizational dimension describes knowledge as individual or collective. The cognitive dimension categorizes knowledge as explicit or tacit. Explicit knowledge is highly codifiable, whereas the codification of tacit knowledge is more problematic. Explicit knowledge is more easily abstracted, aggregated, stored, and shared because it is structured and generated through logical deduction. Tacit knowledge, obtained through practical experiences, is more intuitive, implicit, and unarticulated. Tacit knowledge is difficult to aggregate because it is mostly personal and contextual. Together, the organizational and cognitive dimensions define four categories of knowledge, as suggested first by H. M. Collins (1993). Table 29.1 itemizes the four categories of knowledge with attributes of organizational dimension and cognitive dimension, and the distinctive qualities of each. Lam (2000) further suggested that each organization contains a mixture of knowledge types, but that the relative importance of knowledge types can differ. Some organizations might be dominated by embrained knowledge, where other organizations might be dominated by encoded knowledge. Lam asserted that different types of knowledge correspond to different organizational forms and structures. As a result, the ability with which different organizations can harness and use knowledge, whether explicit or tacit, individual or collective, may vary greatly. As a result, the ways in which organizations learn also vary greatly. The domains and forms of knowledge, explicit to an organization, may bear great weight on its organizational learning and growth. For example, although organizations dealing primarily with explicit, highly structured knowledge adopt formal structures of control and coordination, other organizations dealing with tacit knowledge may tend to be decentralized. Lam (2000) proposed that each knowledge type is associated with an idealtypical organizational form with specific characteristics.
TABLE 29.1. Four Categories of Knowledge Collins (1993). Knowledge Category
Organizational Dimension
Cognitive Dimension
Embrained
Individual
Explicit
Encoded Embodied
Collective Individual
Explicit Tacit
Embedded
Collective
Tacit
Defining Characteristics Formal, abstract, or theoretical knowledge, dependent on an individual's cognitive skills and abilities Information: Knowledge conveyed by signs and symbols Practical, action oriented, dependent on a person's practical experience Organizationally accepted routines and norms
Note. From "The Structure of Knowledge," by H. M. Collins, 1993, Social Research, 60(1), p. 000. Copyright 1993 by Adapetd with permission.
29. Intranets and Intraorganizational Communication
531
TABLE 29.2. Types of Knowledge and the Associated Ideal-Typical Organizational Structure, as Proposed by Lam (2000). Knowledge Type
Organizational Characteristics
Professional bureaucracy and embrained knowledge
• • • • •
The knowledge structure is individualistic, functionally segmented, and hierarchical. The sharing across functional boundaries is limited. Formal demarcation of job boundaries inhibits the transfer of tacit knowledge. Power and status inhibit interaction and the sharing of knowledge. Learning focus tends to be narrow and constrained within the boundary of formal specialist knowledge.
Machine bureaucracy and encoded knowledge
• • • • • •
Key organizing principles are specialization, standardization, and control. The knowledge agents reside in a formal managerial hierarchy. High reliance on management information systems for knowledge aggregation. The knowledge structure is fragmented, collective, functionally segmented, and hierarchical. The organization seeks to minimize the role of tacit knowledge. Learning occurs by correction through performance monitoring.
Operating adhocracy and embodied knowledge
• • • •
Highly organic form of organization with little standardization of knowledge or work process. Knowledge structure is individualistic, but collaborative, diverse, varied, and organic. Tend to learn through interaction, trial and error, and experimentation. Organization capable of divergent thinking and creative problem solving.
I-form organization and embedded knowledge
• Driven by knowledge and embedded in operating routines, team relationships, and shared culture. • Attempts to combine efficiency of a bureaucracy with flexibility and team dynamics of an adhocracy. • Key knowledge agent is neither autonomous individual expert nor the controlling managerial hierarchy, but the semiautonomous cross-functional team. • Learning occurs through shared work experiences and joint problem solving.
Following the work of Nonaka's and Takeuchi's (1995) documenting the Japanese-type organization, Lam (2000) referred to the "J-form" as the organizational form associated with embedded knowledge. Here, the "innovative form" or I-form is referred to, which was initially developed by Sainfort (1987). Sainfort defined an I-form organization as one that has innovation as a core cultural value and capitalizes on organizational knowledge through continuous problem solving. This overview of knowledge types and their associated ideal-typical organizational structure is outlined in Table 29.2. Information technologies, such as intranets, can give an organization the means to support four knowledge types listed in Table 29.2. Although specific applications will be discussed later in the chapter, it is important to first be aware of these categories of knowledge. The organizational characteristics associated with a knowledge type can have great bearing on the capacity of an organization to successfully incorporate different types of information technology in addition to the amount of transition and training time that will be necessary. Although information technologies can help an organization evolve to different structures, the existing knowledge types and characteristics are deep rooted and should be taken into consideration in the planning stages of organizational intranets. Although the four constructs of organizational learning, presented in Table 29.2, are applicable to all types of knowledge, the actual processes of acquiring, storing, sharing, interpreting,
and using the knowledge can vary greatly by knowledge type and organizational form. Figure 29.2 presents this separation of knowledge type and flow according to organization type. In terms of information technology, although organizations who are classified in the upper two quadrants of Fig. 29.2 seem most readily able to incorporate technology to store and communicate explicit knowledge, those organizations in the lower two quadrants who effectively implement IT can experience significant returns in terms of organizational learning. An organization's knowledge structure and form are compellingly correlated to the communication infrastructure. Therefore, tools and technologies used to augment communications should be purposefully chosen for an organization and integrated into an organization for effective coordination with existing structures.
Intraorganizational Communication Communication of an organization's knowledge is a missioncritical component of their learning. In the information age, knowledge is an organization's foremost asset. The capital of a company includes not only physical assets, but also economic value is allocated to the employees' intellectual and social abilities (Carlson, 2000; Drucker, 1988). The sustainability and longevity of an organization have evolved to be highly contingent on the efficient and effective flow of information. In
532
EMERY ET AL.
FIGURE 29.2. Knowledge, organizational learning, and organizational form. Note. Adapted from Lam (2000) and Huber (1991). TABLE 29.3. Internal Forms of Organizational Communication Communication Type Formal broadcasts
Informal intercolleague information exchange
Common Forms
Communication Source(s) • Executives • Corporate Communications Department • Intraorganizational departments • Employees • Colleagues
an interview with Business Week Online in 1998, Jack Welch stated that "[a]n organization's ability to learn, and translate that learning into action rapidly, is the ultimate competitive business advantage." Literature from organizational theory from the last four decades has flooded the gates of popular business culture with terms such as "organizational learning," "knowledge management," and "learning organization," among others. Organizational learning and knowledge management are critical tools for corporations of the twenty-first century and should not be overlooked. The 1990s was a time during which organizations that strove to maintain a viable market presence restructured (sometimes drastically) the way they worked. Organizations transitioned
• • • • •
Announcements Memos Newsletters Bulletins Companywide e-mail
• • • •
E-mail Memos Telephone calls Newsgroups
Content • Policies • Values • Formal event announcements • Processes • Organization charts • Gossip • Best practices • Problem Solution • Practical job knowledge
away from a Taylor-reminiscent, or top-down model of workflow, to a more process-oriented, nonlinear model of workflow. The top-down flow of information from management to employees does not match well to the paradigms for organizational learning. The ability for an organization, at any level, to (a) remove physical and organizational imposed boundaries, (b) acquire and distribute information, and (c) produce new information, enables its successful evolution as a learning organization. Internal communication, according to Hills (1997), can take several forms. Two intuitive groupings of communications emerge: formal broadcasts and intercolleague information exchange and collaboration. Table 293 summarizes each
29. Intranets and Intraorganizational Communication
533
FIGURE 29.3. The linear flow of intraorganizational communication; each symbol represents a different piece of information. Information is often altered by members as it moves down the hierarchy.
classification, information sources, common formats, and content. As organizations have evolved to emulate models of organizational learning, the flow of internal information communication changes with respect to the number of available sources and formats. This freedom of information flow that characterizes organizational learning effectively enables knowledge distribution and acquisition improvements. Traditionally, information flow within an organization was linear. The greater part of an organization's information was shared among a select group of individuals, and information between departments was typically kept separate. Information trickled down, pushed onto employees through the organizational hierarchy. The individuals who regulated information flow have often been referred to as gatekeepers (Gonzalez, 1998; Koehler, Dupper, Scaff, Reitberger, & Paxon, 1998). Gatekeepers powerfully influence the information that is available in decision-making situations, thereby directly influencing organizational learning. Most communication in traditional organizations was also point to point (Koehler et al., 1998); people communicated information person to person or over the phone. Wider messages were distributed through channels to regulate recipients (e.g., memos). The IT explosion and organizational learning
revolution have clearly changed the communication paradigm. This has primarily been achieved through the expansion of possible channels for information transmission. Koehler et al. (1998) summarized three traditional roles of individuals in intraorganizational communication: the gatekeeper, the filter, and the grapevine. Figure 29.3 presents of a hypothetical organization chart. The chart depicts a traditional flow of information through an organization, where each symbol corresponds to a piece of organizational knowledge that particular individual possesses. Figure 29.3 illustrates the linear flow of the organizational knowledge, where gatekeepers control information flow, filters alter information, and grapevines informally share information. In particular, notice how each piece of information (or "symbol") tends to dwell in one classification/department of the organization and is distributed in a point-to-point manner. In this type of hierarchy, crossorganizational communication is rare, especially between hierarchies (Gonzalez, 1998). Information moves up and down within a hierarchical structure in a pattern sometimes called a stovepipe, or silo. Information may be accurately transmitted, but may also be inaccurately conveyed through the insertion or extraction of facts by different entities in the hierarchy. The absence of checks and balances
534
EMERY ET AL.
to determine information redundancy and/or accuracy between classifications in an organization makes it difficult to determine the utility of information to other departments. Once (and if) the gatekeeper releases information, it often travels through "filters" (Koehler et al., 1998). Filters often put a spin on information as it passes through them. The sugarcoating of intentionally strong messages can cause those at the end of the communication chain to see a different picture of things than was intended. Grapevines, as shown in Fig. 29.3, represent a more complex communication network that informally spread information with the underlying goal of knowledge verification. Entities in the organizations are sometimes left out of knowledge sharing (intentionally and unintentionally) by gatekeepers and filters. As a result, grapevines often transmit a distorted picture of the actual information to each other. Without intraorganizational communication channels, the information transmitted is often inaccurate, untimely, and counterproductive to organizational goals because grapevines are limited in their information and verification sources. Barriers to intraorganizational communication can restrict an organization's potential for success in today's information age. Cost, globalization, time, information expirations, and information source platforms are all barriers to intraorganizational communication. Another barrier is an individual's willingness to contribute to the flow of information. In an age where information and knowledge are seen as capital investments, information ownership can give individuals a sense of power and job security, which they are reluctant to relinquish. These challenges to open communication within an organization remind us that an intranet application that enables online publishing will not automatically return the desired results in terms of organizational growth. IT such as intranets is only a means to success. The groundwork of new models of intraorganizational knowledge sharing must first be put into operation. An infrastructure that encourages, supports, and documents knowledge sharing across the entire organization, and creates organizational memory, can better facilitate organizational learning (Harvey, Palmer, & Speier, 1998). Organizations need to effectively reinvest in their human resources (Senge, 1990), not just IT resources, because the employees are a core element to the learning organization. Technology has the ability to both enhance existing organizational structures and facilitate new relationships and information flow (Riggins & Rhee, 1998). Riggins and Rhee expanded this taxonomy to illustrate aspects of work affected by the facilitation and enhancement (e.g., integration and upgrade) of technology. Table 29.4 provides a summary of this expansion. The human element in these technology facilitations and enhancements should not be disregarded. Carlson (2000) summarized four human-centered concepts for evolving organizations to recognize: The value of individual initiative and insight. The need for ethicality, responsibility, and commitment. The essential nature of teamwork, both formally and informally structured. The breakdown of hierarchical order through innovations such as matrix management, (p. 31)
TABLE 29.4. The Role of Technology in Organizational Communication Technology Enhanced Improve coordination with internal business units Promote the efficient exchange of information
Technology Facilitated Coordination of work and information across a variety of applications Promotes overall organization competitiveness with external forces
The plausible impacts of an IT, such as an intranet, on the evolutionary process of organizational learning merits judiciary implementation of the technology. An understanding of the hardware and software components associated with intranets can guide effective decisions and impact long-term effectiveness.
DEFINING INTRANETS An intranet uses Internet technologies and applications from the World Wide Web within an organization. An intranet is a corporate network designed around Internet metaphors, protocols, and technology. Fundamentally, an intranet integrates several information systems that support various functional areas of an organization that would otherwise be functionally incompatible due to differences in technological platforms. Although the components used to define intranets in this section may appear to be older in terms of technology, they should be regarded as core components, critical to the underlying functionality of Intranets. An intranet presents an organization with the functionality and benefits of the World Wide Web, with the added advantages of separation and security away from the rest of the Internet. Anyone outside the organization is restricted to access of the intranet through the security of a firewall. People within the organization are given a channel through which to tap into the Internet. Figure 29.4 illustrates this association. Intranets mimic several aspects of the Internet, but with the security of a firewall and restricted access to organization-only members. An intranet is a distributed hypermedia system within an organization that retains the capability to enable people to access information, communicate with each other, collaborate in a geographically distributed manner, share knowledge, and learn from each other with innovative modes of communication. Technically speaking, this distributed hypermedia system could be any private transmission control protocol that supports intranet applications, including Web protocols [hypertext transfer protocol (HTTP)] and others. Specifically, three web standards [uniform resource locators (URLs), HTTP, and hypertext markup language (HTML)] enable an intranet to simplify locating, displaying, sharing, and accessing information. As long as an operating system has a browser, users across different departments can access the same information and applications, independent of their individual respective platform. At conception, intranets differed from other IT tools and shared applications because an intranet has the ability to integrate across different servers, improving the capacity of the
29. Intranets and Intraorganizational Communication
•
535
TABLE 29.5. Types of Networks Type of Network Sneaker Peer LAN MAN WAN
FIGURE 29.4. Intranet and Internet configuration. information creator/information user relationship. Instances of intranet use in organizations may range from the simple to the complex, depending on an organization's resources. It challenged programs such as Lotus Notes' key market in corporate information and message systems to support workflow. An intranet's Web-based cross-platform and cross-vendor compatibility were much-needed changes from existing products and systems (Dix, Finlay, Abowd, & Beal, 1998). Intranets enable normalized information accessibility among separate departments. Engineering, marketing/creative, and administrative departments can have comparable access to information, independent of the differences between operating system platforms (e.g., UNIX, Macintosh, or Windows). Intranets facilitate intraorganizational communication without requiring individuals to depart from the tools they prefer to use in their work, which could compromise efficiency, productivity, and worker satisfaction. Several authors have identified the building blocks of an intranet (Gonzalez, 1998; Hills, 1997; Rosen, 1997). Two of the most commonly cited components include networks and Web technologies. As some aspects of these components are beyond the technical scope of this text, we provide a brief overview of each. The following section summarizes each element and explains key attributes with respect to intranet technologies.
Networks Networks have been described as the most complex component of technology implementation (Hills, 1997). Networks are the core of intranet and Internet communication because they dictate how information systems are interconnected. There are several ways via which networks can be connected, and innovative products to access networks are continually being developed and released. For this reason, we will not delve into the technological options for networks. Rather, we will highlight attributes of networks that can have an impact on the implementation of intranets and intraorganizational communication. Different network technologies make use of different protocols for information transfer within networks. One example
Description Walk information storage devices (e.g., disks, CDs, media cards) from computer to computer Two or more computers share the same software and peripherals Servers, computers, and peripherals organized for efficient information transmission Campus or citywide network of LANs Network of several LANs
of a type of network technology is an Ethernet. Factors that impact the network efficacy within the context of an organization's intranet include speed of transmission, network capacity, and localization of networks. Greer (1998) identified five types of networks in his work: sneaker, peer, local area network (LAN), metropolitan area network (MAN), and wide area network (WAN). These networks span the technical spectrum from simple to complex. Table 29.5 clarifies these forms as interpreted by Greer (1998). Crucial factors for networks with respect to intranets are ease of use, scalability, and interoperability between different areas of an organization. System responsiveness is one technical attribute that can impact user interactions. In networked computing environments such as the Internet or intranets, every user action may be followed by a delay as information is transmitted across the network. Research has demonstrated that the length, variability, and predictability of delays can affect the way users interact with such systems, and in the context of the Internet, have demonstrated that network delays can also significantly affect perceptions (e.g., Borella, Sears, & Jacko, 1997; Jacko, Sears, & Borella, 2000; Sears, Jacko, & Borella, 1997a, 1997b; Sears, Jacko, & Dubach, 2000). Therefore, it is critical that the delays users experience be carefully considered as researchers investigate the issues that affect the usability of distributed multimedia documents accessed via an intranet (Sears & Jacko, 2000) because these delays can impact organizational productivity and user satisfaction.
TCP/IP Transmission control protocol/Internet protocol (TCP/IP) is a principal protocol in the Internet because it transmits information across networks. The flow of information is managed from place to place in the network (Gonzalez, 1998; Hills, 1997; Rosen, 1997). TCP/IP enables computers between departments to use the types of browsers and servers used to access the World Wide Web-based toolkits.
World Wide Web Toolkit The key element that distinguishes intranets from other organizational IT tools is the implementation of World Wide Web technologies and protocols that reside at its core. As previously mentioned, the intranet can benefit from the tools of the World
536
EMERY ET AL.
Wide Web. Hyperlinks, search engines, site maps, and information repositories are all instruments successfully carried over from the Internet to intranets. Servers and browsers are recognized as two mission-critical components for the start-up of an internal Web (Hills, 1997).
SERVERS Servers are the hubs of the intranet. The number of servers required for an intranet depends on the amount of information to be held in the system. Hills (1997) reported that companies have implemented between 1 and 100 servers and, in unique cases, more than 2,000 servers to help support their intranet. An important distinction is that servers use HTTP. HTTP is the standard communication protocol between servers and browsers (Rosen, 1997), and is another contributing factor to the consistency of Web applications in intraorganizational communication.
BROWSERS Web browser platforms serve as the software to access and display information. This universal interface is a tool for access to all organizational information. In many organizations, specific browser platforms, brands, and versions used (e.g., Internet Explorer 6.0) are well documented so intranet pages can be designed to maximize effective usability throughout the entire organization. The objectives of information access and editing capabilities are independent from the information's source and format. A browser requests information from the servers and displays pages to the user in a graphical user interface. Browsers, along with HTTP, ensure users will see the same view of the information as it is translated from the server. Browser selection should be of utmost importance to an organization, as browsers are fundamental tools for intranet platform independence. Browsers impact scalability, usability, and usefulness of an organization's intranet. Although only two browsers have really permeated the marketplace (Netscape Navigator and Microsoft Internet Explorer), the decision still merits careful consideration. An incorrect match for an organization's needs could impact efficiency, employee productivity, employee satisfaction, and quality. Guengerich, Grahm, Miller, and McDonald (1996) related five dimensions to assess a browser's ability to meet organizational (corporate) requirements: (a) performance capacities; (b) multimedia support (e.g., sound, video, or plug-ins); (c) which computer languages are supported; (d) usability of the browser's functions; and (e) vendor support (Guengerich etal., 1996).
sophisticated toolkit for the support and enrichment of knowledge within an organization. The extent of intranet technologies implemented in an organization and their use are interrelated with the growth of that organization, its learning activities, and knowledge sharing. Figure 29.5 illustrates this influential relationship. This section introduces and expounds on these uses, in terms of function exclusively. The evolution and interaction of an intranet amidst layers of intraorganizational communication are discussed subsequent to the discussion of organizational learning. An intranet has a high capacity to force a company to focus on core information technology strategies and enforce strict technical and procedural standards at beginning stages. Traditionally, an intranet has been studied with respect to five fundamental work applications (Dascan, 1997; Greer, 1998; Hills, 1996): 1. E-mail: In terms of intraorganizational communication, comprises both person-to-person exchanges and person-to-group broadcasts. 2. File sharing: May be present as a person-to-person or a person-to-group interchange of files. 3. Directories: In a functional sense directories are comprised of organizational information and are built to facilitate user access to the information. 4. Searches: The search functionality is one of the most important components of an organizational intranet. In a working environment, where time is money, a user's ability to find what they want, when they want it, the way they expect it can directly affect the bottom-line finances. Intranet search should reflect available knowledge and location of the information. 5. Network management: Network management deals with the maintenance and modification of intranet infrastructures for the modification and compilation of different organizational and work components. A more categorical view of intranet applications lends itself to a three-dimensional classification: (a) information creation, (b) information dispersion, and (c) information manipulation (evolution). Furthermore, each of these activities may be executed in one of three ways: 1. Asynchronous -or- Synchronous • Are related work functions (especially information activities) performed in a serial or parallel manner? 2. Collaborative -or- Independent • Is the information/work activity assembled by an independent organization member or by group(s) of organization members in collaboration? 3. Information Creator Initiated -or- Information User Initiated • Is the user of the information pushing or pulling information and knowledge through the intranet/organizational system?
Intranet Applications The extent of functionality gleaned from intranet use by an organization ranges from basic document management to a more
An intranet-mediated function may be capable of a range of these attributes and functionalities, depending on the nature of the users, work, and existing organizational infrastructure.
29. Intranets and Intraorganizational Communication
537
FIGURE 29.5. The extent to which intranets change the nature of organizations today with respect to organizational learning (Jacko et al., 2002).
due to differences in the organizational structure as they develop into a learning organization. This list of applications in Table 29.6 has been assimilated from popular literature as well as research. Innovative strategic uses of intranet technologies continue to enter into the domain. For this reason, we refrain from detailing the specifics of each application, but instead lay emphasis upon the far-reaching impacts intranet technologies can create within an organization. From a human factors perspective, each application is likely to have a unique set of users, contexts, rules, and goals for which it needs to be optimized. Consider, with each intranetempowered activity listed, how it could potentially be classified within the organization, for this dictates the amount of planning, integration, and management. This chapter proceeds next to a discussion of the ways the application of intranet technologies evolve in an organization and the impacts on intraorganizational communications. FIGURE 29.6. Classifying functionality of intranet tools. Figure 29.6 illustrates the three dimensions along which a function of intranet technology may contrast under different contexts of use. The boundaries within which functionality are defined represent the range of activities in support of that work. Dimensions associated more with organizational learning will be positioned in the area of the diagram with synchronous, collaborative, information-seeking characteristics. The application of intranet tools to an organization can serve multiple functions, or just one. The flexibility of the functionality will reflect the associated changes in the organization, whether it is due to differences between working projects or
Associated Information Technologies The capacity of intranets can be better understood through an examination of closely related (and often interrelated) information technologies. These information technologies include Internet, extranets, and portals. Both portals and extranet are important aspects of intraorganizational communication and intranets. A detailed account of these forms of IT is beyond the scope of this text. However, a macroscopic understanding of what each entails, as well as their relation to and association with intranets, each other (when applicable), and intraorganizational communications, lends additional insight into the role intranets play with these fundamental information technology structures.
538
EMERY ET AL.
TABLE 29.6. Examples of Intranet-Empowered Activity • • • •
Company news distribution Newsletters Press releases Corporate policies
• Company goals • Employee and community activities • Employee recognition • Organizational charts
• • • •
Departmental information Project management Workflow management Knowledge repositories
• • • •
Just-in-time training Quality information Product/pricing information Shipping and inventory data
• Regulatory compliance status • Employee locator
• Sales reports
• Skills directory • Job posting
• Claims processing • Benefits enrollment
• • • •
• Materials ordering
Stock prices Discussion groups Employee surveys Multipoint authoring of documents • Distributed learning/training
• Corporate policies and procedures • Health and safety regulations • Sales automation
TABLE 29.7. Intranets Compared With the Internet Dimension of Comparison
Intranets
Internet
User
Users have knowledge of organizational aspects such as structure, terminology, and circumstances.
Work
Tasks are related to daily work and involve complex applications. Information exists in multiple forms (e.g., progress reports, human resource information). Technology is more uniform and controlled to allow richer forms of media and to allow cross-platform compatibility.
Users are customers who possess less knowledge and have a less investment in the organization (e.g., time and effort). Tasks involve primarily information retrieval. Tasks involve primarily information retrieval. Information is primarily market oriented.
Information
Technology
Differentiating Intranets From the Internet Although intranets and the Internet share the same general attributes, their context of use and designated users deviate, which means that implementation and utilization support of each is significantly different. As introduced in the network section of this chapter, bandwidth, or system delays imposed by information quantity can affect users' perception of a system, as well as their productivity in the organization. Technological nuances aside, the type of content and fundamental applications differ between the Internet and intranets (Scheepers & Damsgaard, 1997). It is especially important not to draw absolute parallels between acceptance and use of an intranet versus acceptance and use of intranet technologies (Scheepers & Damsgaard, 1997). The extent of utilization ultimately lies in the hands of the individuals within the organization, no matter the level technical sophistication an application may capacitate. An intranet is an internal information system that employees might be encouraged (or even required) to use to accomplish their regular work functions. This is different from the Internet, where users may neither work within the scope of one organization, nor share common characteristics, and where users ultimately choose whether they want to visit specific Web sites
Technology capabilities arrange from low to high measures taken to ensure performance on multiple computer platforms and software versions.
(Lazar, 2001). An assortment of specific organizational needs and requirements emerge with the integration of an intranet. An organization's intranet is unique because it narrows the scope of potential users. Also, the functional capacities of an intranet are typically far more complex and context specialized than those needed of the Internet. The specific needs and differences for an organization's intranet and Internet are compared and contrasted in Table 29.7. The fundamental differences between intranets and the Internet are observed on four dimensions: (a) the nature of the user, (b) the nature of work, (c) the nature of information, and (d) the nature of the technology. In consideration of user-centered design principles, the level of difference between intranet and Internet applications signifies the necessity for distinct approaches to the design and implementation of each. Despite the similarity in technologies used to build intranets and the Internet, the boundaries of use and goals for implementation create discord between the two. The fundamentals for motivating end users to use the intranet solutions provided to them differ innately from marketing an Internet site for external use. The quality of work and knowledge sharing produced by an intranet site relies heavily on employee awareness, comfort, and productivity with the tools.
29. Intranets and Intraorganizational Communication Although marketing tools and practices are typically applied to Internet site launch, human resources and strategic management guide intranet launch and evolution within an organization. In the context of Internet sites, market research determines when, where, and how a launch should proceed in order to realize a high return on investment (Rosen, 1997). The best strategies for the sale of intranet to its end users have been identified to include (a) executive support, (b) employee awareness and involvement with development (c) training and education, and (d) continued management and support with evolving intranet use. Intranet designers must capitalize on their ability to characterize and understand users' needs and working environments.
Intranets and Extranets It is rare to read about intranets without some mention of extranets. An extranet allows portions of an organization's intranet to be accessed by outside parties, vendors, and customers. Although there are several overlapping attributes, discrimination of intranet and extranet is accomplished through consideration of the location of the users with respect to the network firewall (Riggins & Rhee, 1998). Intranets and extranets both capitalize on the independent protocol provided by Internet applications, which could cause issues with interorganizational compatibility. The placement of extranets are relative to the location of the user and with respect to the origination's firewall. The large break between intranet and Internet is crossed by intranet technologies accessed from Internet sites. Typical information gained through an extranet would include order status, vendor lists, billing processes, and account information. It is a blend of the public Internet and closed intranet, incorporating fundamentals of each (Nielsen, 1997, 1999). Extranets can optimize communications to strengthen relationships with customers, suppliers, and partners for potential quality improvements and cost savings. In terms of the potential for costs savings, Riggins and Rhee (1998) assert that extranets can improve coordination with existing trading partners, as well as market to reach new customers. Interactive communication of select internal information to privileged business partners may produce gains in productivity and for an organization's just-in-time processes (Koehler et al., 1998). Extranets are added value to secure business-to-business information sharing and transactions (Hope, 2001). Riggins and Rhee (1998) asserted that extranets associate an organization's Internet and intranet infrastructures. It is anticipated that this connection would extend the meaning of electronic commerce beyond point-of-sale applications. Extranets share some design requirements of intranets, with a few identifying attributes. This is indisputably attributed to extranets' placement as a customer relation tool. The goals and the underlying nature of the work to be accomplished via extranet do, in fact, contrast with those of intranet. Hope (2001) listed in her work, the top 10 benefits that companies expect intranets to yield: 1. Enhance competitiveness or create strategic advantage. 2. Enable easier access to information.
539
3. Provide new products or service to customers. 4. Increase the flexibility of information requests. 5. Improve customer relations. 6. Enhance the credibility and prestige of the organization. 7. Provide better products or services to customers. 8. Increase the volume of information output. 9. Align well with stated organizational goals. 10. Enable the organization to respond more quickly to change. Items 2-6 and 10 are specific to customer relations and service quality. This list illustrates that the public relations aspect of the intranet is substantial. One area not directly addressed in Hope's work is the importance of extranets in the facilitation or relationships with vendors, and other important business partners. Extranets provide an important bridge to the world outside the organization, in attempt to produce a more streamlined organization.
Intranets and Portals The literal meaning for portal is a door or gateway. In the technical context, the term "portal" is nebulous. A few years younger than the concept of intranet (information on portals pre-1998 is sparse), the concept of portals began as a simple concept. Portals were principally viewed as homepages that presented web information applications and access to their use. Customization of a portal by an organization member enables the user to have, readily accessible, their most commonly used tools. A presentday example of a portal is MyYahoo, an Internet portal that allows users to choose which resources to present on their start page. Similar to the concept of intranet, portals are associated with several terms, between which boundaries of distinct definitions are weak. Knowledge portals, enterprise portals, information portals, intranet portals, Internet portals, Web portals, and so on—these terms have been used synonymously. They have also been used to define the different infrastructures of portals in the limited research on portals, especially in the popular technology periodicals and newsgroups. This is due to the high rate of development of portal creation applications by software companies and the slower rate of research in this area. The names associated with portals are more likely a product of marketing, and not research. At a less granular level of analysis, portals can be viewed as a gateway to information resources. Two differing, yet overlapping, research classifications of portals are presented within. Tkach (1999) identified three basic typologies of portals: intranet Web portals, information portals, and knowledge portals. Table 29.8 differentiates between the three different types of portals. At the turn of the twenty-first century, portals were viewed as the latest and greatest strategic IT move, the key to organizational Utopia. No longer just a catalog of Web resources, portals are presently integrated with organization-specific applications, such as enterprise resource planning, along with other IT resources internal and external to an organization. More recently, Collins (2001) divided portals into two classes: corporate portals and enterprise portals. A corporate portal builds an infrastructure for needs within an organization,
540
EMERY ET AL.
Intranet Web portals
Information portals Knowledge portals
Provide links to all internal content providers and some external success providers. Present a uniformed, common look for access through system log in. Make available all facilities of information catalog and collaborative facilities, management tools, and a knowledge repository.
Note. From "Knowledge Portals," by D. Tkach, 1999. Available: http://www.4. ibm.com/sof1ware/data/km/advances/kportals.html. Copyright 1999 by Adapted with permission.
and an enterprise portal spans the corporate portal as well as customers and vendors external to the originations. The differentiation between corporate portal and enterprise portal is analogous to intranet and extranet, respectively. The corporate portal assembles navigation services in one main location made available for employees to find information, launch applications, interact with corporate data, identify collaborators, share knowledge, and make decisions. One definition of a corporate portal is: A browser-based application that allows knowledge workers to gain access to, collaborate with, make decisions, and take action on a wide variety of business-related information regardless of the employee's virtual location or departmental affiliations, the location of the information, or the format in which the information is stored. (Collins, 2001, p. 7)
Vendors create enterprise portals that allow end users to customize information access to their preferred internal and external resources. The intranet portal is a layer within the infrastructure of an organization (Collins, 2003). The functional role of an intranet portal may encompass six features: 1. Retrieve information from corporate IT systems, and present the results according to the roles, specific tasks, and preferences of individual employees. 2. Present employees with information relevant to their daily tasks without making them search for it. 3. Gather information about each employee, facilitating communication between the people who need information and the people who can supply the information. 4. Allow employees to act on the information presented in the desktop without requiring them to switch to a different system or interface for the purpose of sharing the information and collaboration with other employees. 5. Present a desktop interface through a Web browser that requires minimal technical training. 6. Support multiple business processes for a single department, a single process across multiple departments, or multiple processes across multiple departments. (Collins, 2001, P. 55) Portals take the concept of intranet one stage further, by allowing individuals to integrate their personalized information
and organizational information resources. Gains from portal use can be realized in terms of better decision-making abilities, improved organizational understanding of terminologies and information infrastructures, and more accessible intraorganization information and resource retrieval; direct links to reports, analysis, and queries; and personalized access to content for each employee/worker. A need for a corporate portal is demonstrated when there are too many applications that the user has to start up and switch between. Portals are becoming a key facilitator of IT, following in the footsteps of intranet technologies.
INTRANETS AND INTRAORGANIZATIONAL COMMUNICATION Immediate impacts of intranet use can be observed in the introduction of new communication mediums. The way people within the organization communicate can be altered through the implementation of intranets and Web-based technologies. Yet, despite these direct, more tangible returns, an intranet can translate into more indirect changes in business infrastructures. An intranet, while propagating information through an organization in the long term triggers a metamorphosis of the way work is accomplished and organizational goals are met (Koehler et al., 1998). Figure 29.7 provides a schematic view of several of the drivers for the use of intranets and similar information technologies within an organization. In addition, Fig. 29.7 also illustrates a number of positive outcomes that may result from the implementation of these technologies to the organizational structure. Huber (1991) recognized that IT could play a critical role in supporting, storing, organizing, and accessing organizational memory. Malhotra (1996) went on to suggest IT could also help in knowledge acquisition (e.g., market research systems, competitive intelligence systems, scenario planning tools, search tools), information and knowledge distribution (e.g., groupware tools, e-mail, bulletin boards, Web casting, e-training). Gonzalez (1998) pointed out that Web-based technologies "can have a profound, long-lasting impact on organizations and can improve companies' chances to compete" (p. 102). The same message was iterated in books, news articles, and Web sites throughout the 1990s—"technology can help your organization achieve anything." Gonzalez (1998) also asserted that the business model for IT and its positive changes can be characterized by a set of drivers and outcomes in the organization. The catalyst between the two is Web-based communication technologies. Figure 29.8 is based on Gonzalez's drivers and outcomes, and illustrates these relationships. Intranets, as with other IT solutions, initiate new ways of intraorganizational communication. Unlike many other information technologies, however, intranets are can be molded from a generic solution into a support tool for a specified management strategy (Damsgaard & Scheepers, 1999). Another term for intranet is organizational Internet (Scheepers & Damsgaard,
29. Intranets and Intraorganizational Communication
•
541
FIGURE 29.7. Interaction of Web-based communication tools in organizational change (Gonzalez, 1998).
1997), implying Web technologies in a closed, defined system. This gives structure to a more complex, dynamic dimension for the flow of information between creator and user (Telleen, 1997). Figure 29.8 depicts a departure from the traditional corporate view of communication introduced earlier in Fig. 29.3. This is a representation of the dispersion of information supported by an interconnected communication structure of a learning organization. Information is shared between departments at several hierarchical levels of the organization. The removal of organizational barriers stimulates better-informed decision making, and better development of organizational knowledge and learning. Communication of information and knowledge abandon the traditional linear structure for a network-based structure (Gonzalez, 1998; Koehler et al., 1998). Information channels throughout the organization reflect both content and structure, as opposed to organizational power or hierarchy.
Roles of the individual within this model of organizational communication are less intrusive on the knowledge management process. Individuals often take on multiple roles, as they become both information-seeking entities and informationproviding entities, or gatekeepers. In fact, new, vital communication roles emerge: Web master and content providers (Koehler et al., 1998). The Web master may work alone or in conjunction with staff, coordinating the structure of the intranet, hands on. Content providers are anyone in the organization who have knowledge to share. The drive for organizational learning and dispersion of information drives this model. The information flows within a more complex network of relationships, moving away from the linear flow of traditional organizations. Strategic management and IT are often cited as the keys to unlock the potentials of informed organizations to achieve this level of communication.
542
•
EMERY ET AL.
Communication tends to be networked.
FIGURE 29.8. The nonlinear flow of intraorganizational communication. Each block in the organization chart represents one employee. Each symbol represents knowledge in the organization.
Intranet technologies better enable users to putt the information they want from the technology, instead of waiting for the information to be placed in front of them. Equally, intranets better enable users to push information, via the use of technology, instead of relying on the other, less efficient, less reliable models of communication. Furthermore, when access is only internal, a company has the potential for better control of information security than that of paper documents. Intranets hold great promise for organizational learning, for along with nonlinear flow of knowledge, intranets enable real-time exchanges of knowledge that users have the ability to shape to meet their working and information needs. An intranet relates to learning organizations, not only because it is a powerful communication medium, but also because it has the potential to serve as a knowledge base. The intranet more easily captures and handles unstructured and implicit knowledge, a benefit over other IT systems. Intranets' usage and organizational learning are interrelated; as each expands
within an organization, the development of one influences the other (Curry & Stancich, 2000; Duane & Finnegan, 2000; Harvey et al., 1998; Jacko et al., 2002).
Intranet and Organizational Growth Several authors have discussed the waves of intranet implementation within an organization (Curry & Stancich, 2000; Damsgaard & Scheepers, 1999; Duane & Finnegan, 2000; Harvey et al., 1998; Jacko et al., 2002). The consensus in the literature is that intranet utilization and applications evolve from communication tools into communication infrastructures. An organization initiates its intranet adoption with a limited number of applications and can increase the technology's usefulness as they activate further intranet functionality (Damsgaard & Scheepers, 1999). The different phases of intranet utility start with simple tools (posting static documents) and evolve into more complex so-
29. Intranets and Intraorganizational Communication
543
TABLE 29.9. Six Stages of Intranet Growth Stage Introduction Customized growth Process and systems integration External value chain integration Institutional absorption
Description Explore the use of an intranet and educate other department of its potential benefits. Acquire information, publish it in an organized and structured way, and provide rapid access to information via a search engine. Integrate an intranet with computer-based systems and network applications to integrate work processes. Integrate the external supply and distribution benefits of the strategic alliance, and share information. Institutionalize an intranet in the organization and continue to explore ways of capturing sharing, storing, and managing tacit information.
NOK. basea on uuane ana nnnegan uuuuj.
lutions, or comprehensive toolkits. Organizations, in the latter stages of intranet implementation, realize the malleability of the technology (Damsgaard & Scheepers, 1999), enabling the spectrum of functions. Stages of intranet development have been identified to be • • • •
Publishing of static information Informal, asynchronous collaboration applications Transaction-oriented applications Formal collaboration applications
Several, more specific "stage models" of intranet growth have been generated in the literature, with significant overlap between them. In their model, Duane and Finnegan (2000) developed and validated their six-stage model of intranet growth. The authors traced the evolution of intranet adoption in a large corporate organization, Hewlett Packard in Ireland. Six stages of growth emerged: introduction, customized growth, collaborative interactions, process and systems integrations, external value chain integration, and institutional absorption. Table 29.9 further describes this growth process. Through this growth process, organizational infrastructures and intranet applications change concurrently with respect to abilities and requirements. Similar to intranets, organizations and organizational learning evolve in phases and components, and the intranet can be used to support the different components. First, consider Huber's (1991) classification of organizational learning as introduced in Fig. 29.1. Activities supported by an intranet can draw parallels between the classifications of knowledge acquisition, information distribution, and information interpretation, providing a repository of sorts for the organizational memory to reside. Cangelosi and Dill (1965) asserted that organizations exhibit adaptive behavior over time. Hence, in their view, organizational learning must be viewed as a series of interactions between adaptation at the individual or subgroup level and adaptation
Phase 1 Learning: Stimulate learning within functional units Intranet: Dissemination of existing documents, and user-initiated sharing of tacit and embedded knowledge Phase 2 Learning: Expansion of learning between organizational hierarchies Intranet: Use of collaborative technologies such as chat rooms and group e-mails to support knowledge-sharing dialogues. Formulation of standards and techniques to support efficient information retrieval Note. Based on Harvery, palmer & Speier (1998).
at the organizational level. They suggested that adaptation occurs as the result of three kinds of stress: discomfort stress (related to the complexity and uncertainty of environment), performance stress (related to perceptions of past successes and failures, outcomes of past decisions, and aspirations levels and expectations), and disjunctive stress (divergence and conflict in individual behaviors). It assumes that learning is sporadic and stepwise rather than continuous and gradual. Harvey et al. (1998) explained that intranets support four phases of intraorganizational learning. Table 29.10 expounds on these phases, which include stimulation, expansion, extension, and encourage. The significant interplay between intranets, organizational learning, and intraorganizational communication influences workflow, employees, and organization quality. Each should be viewed as a mission-critical element of the organization. When observed collectively, these tools and ideologies impact the organization in ways that may not always be positive. Table 29.11 represents an extraction from the literature, representative of work conducted in involving intranet technologies, from 1996 to the present. The nature and scope of this work range from simple to complex and represent the wide range of functionality explored. This is shown in the left-hand column of the table, where intranet functionality evolves from document production and management to knowledge management.
COSTS AND BENEFITS OF INTRANET UTILIZATION The introduction of an intranet, with the necessary support, can spark transformation and generate benefits that have an impact at several levels within an organization. A systematic examination of benefits and costs associated with intranet utilization is useful to create awareness of their far-reaching impacts on intraorganizational communication. Intuitively, a high return on investment is one of the main motivators for adopting intranet technologies. However, it should be emphasized that a focus on cost as justification for the use of a new technology has, many times, not been an indicator of
544
EMERY ET AL.
TABLE 29.11. Research Aimed at Extending the Role and Functionality of Intranets Intranet Functionality Unify dispersed computer-based information systems in an organization into one rich system Document/information management and retrieval (e.g., forms, news, policies, phone directories, product specifications, pricing) for internal and external users Web search and file transfer protocol (FTP): access to organizational memory: job descriptions, employee benefits, operating manuals, organizational policies for search Data mining and data access processes; search and manipulate data without leaving a homepage Real-time transfer of information exchanges between individuals and groups, internal and external Well-established technology to solve intraorganizational information-sharing problems. Easy and straightforward implementation E-mail and workgroup support; integrated functions, maintenance, grouping, sorting, calendaring/scheduling, eliminate geographic limitations, including external partners Transparent interface to e-mail, file transfer, and discussion groups Decision support and decentralized decision making Interactive programs and learning labs where users manipulate systems dynamics models All phases of the training and learning process for end users; at any time, any place, from any location Collaborative design, concurrent engineering, and workflow support The added value of knowledge management activities and Web applications Intranet as a component of strategic management and organizational learning Evolutionary stages of the adoption of intranet technology Strategic management of organizational knowledge through portals
Reference Baines (1996); Scheepers & Damsgaard (1997); Hazel & Jones (2001) Coleman (1997); Gonzalez (1998); Lai (2001); Nielsen (1999); Scheepers & Damsgaard (1997); Zhang & Chen (1997) Zhang Zhang & Chen (1997); Nambisan & Wang (1999)
Young (2000); Zang & Chen (1997) Lai (200 1 ); Nambisan & Wang (1999) Dix, Finlay, Abowd, & Beale (1998); Scheepers & Damsgaard (1997)
Coleman (1997); Greer (1998); Riggins & Rhee (1998); Zhang & Chen (1997) Scheepers & Damsgaard (1997) Suresh (1998) Zhang & Chen (1997); Bullinger, Mueller, & Kempf (2001) Greer (1998); Mahapatra & Lai (1999) Akoumianakis & Stephanidis (2001); Dix et al. (1998); Scheepers & Damsgaard (1997); Gill (2001); Martin (1999) Rademacher (1999); Gonzalez (1998); Martin (1999) Curry & Stancich (2000); Riggins & Rhee (1998); Koehler et al. (1998); Martin (1999) Damsgaard & Scheepers (1999); Duane & Finnegan (2000); Harvey et al. (1998); Lai (2001) Collins (2001, 2003); Ji & Salvendy (2001); Nielsen (2000)
Note. From "Intranets and Oragnizational Learning: A Research and Development Agenda," by J. A. Jacko, G. Salvendy, F. Sainfort, V. K. Emery, et al., 2002, }ournal of Human-Computer Interaction, 14(1), p. 102. Copyright 2002 by Lawrence Erlbaum Associates. Adapted with permission.
success (Duffy & Salvendy, 1999). Benefits expected from IT implementation, including productivity, managed change, and enhanced human abilities (Carlson, 2000), could be influenced in positive ways with intranet technologies (Jacko & Duffy, 1997). Benefits, however, are not without their costs. These costs are financial in nature, as well as in terms of resource expenditure, process redevelopment, and employee training and retraining, among others.
Productivity Gains Versus Information Overload Platform independence and cross-functional operations are two benefits of implementing intranets. This independence unifies "islands" of information that may have otherwise remained separated by technology, time, or physical space (Scheepers & Damsgaard, 1997; Zhang & Chen, 1997). The unification of these islands results in more informed actions on the part of the user. Still, new freedoms that are granted to users by these technological applications (e.g., publishing, editing, developing, and
exchanging information) could also have destructive consequences if left completely unmonitored. For an intranet to support successful knowledge sharing, the information needs to be timely, up-to-date, maintainable, and cost effective (Curry & Stancich, 2000). There is a point of diminishing returns, when the knowledge available for retrieval has to be filtered for relevance, appropriateness, and timeliness to fulfill the user's task. For example, it is imperative to closely monitor information volume. If information builds up in uncontrollable quantities, valuable information may not be retrievable, user frustration and dissatisfaction with the system may increase, and use of the intranet may decline (Koehler et al., 1998). Many individual group or department sites within an organization become grouped in different portals by an organizational IT group. IT groups may not have the facilities to maintain all these individual sites to standard. The IT group usually receives the blame when these sites are not functional, incorrect, or outdated. Often, sites are abandoned when divisions become obsolete (a common occurrence in the last few years), leaving no one around who is accountable for removing them from the server. In addition, it is difficult to upgrade and make
29. Intranets and Intraorganizational Communication changes to portals that experience high traffic. When a user interface is changed, the users often demand a "mirror" site that maintains the old style because they are reluctant to accept the change. Coordination of a system for accessible archives should be determined by the Web master and dictated by the uses of this information. An additional solution is the enforcement of standards for information and knowledge on the intranet. A standard specifies required, recommended, and optional design elements (Nielsen, 2000) for navigational structure to page layout, icons, and logos. Nielsen identified several guidelines for standards. • Every intranet page should have a search button. (Robust search functionality should support the search button.) • Provide illustrated examples that fully comply with standards. • Provide checklists for verifying standards are met. • Provide consistent support to all questions about standards. • Actively enforce and support the use of standards through "evangelism programs." • Maintain changes to standards as needed. • Comply with standard usability design standards, bringing attention to discrepancies, when these digressions are necessary. • Support development tools and templates to facilitate compliance. • Post standards to the intranet, with appropriate search function and hypertext rules. • Printed standards should have a sound index for quick reference. A disclaimer should accompany these standards because, as an organization grows, the user interfaces of intranet applications become more difficult to standardize and maintain. The root of the impediments to intranet standardization parallels issues with Internet standardization: too much development, too fast. Developers and IT specialists likely become decentralized and are dispersed throughout the organization. In addition, the standardization of intranet pages may be given less priority than the standardization of externally facing Web pages (e.g., extranet, Internet). Although, in fact, a lack of continuity between intranet pages could negatively impact the bottom line of productivity. It should also be noted, however, that too much control and standardization can undermine the intranet's ability to attract information content providers, as well as suppress the content provider's creativity. Ultimately, less information will be available, which will result in fewer interested users and eventually smother productivity gains. Alternately, allowing for continued "organic growth" of the information system, with a hands-off approach, could lead to a mass amount of information, interconnected in unproductive, confusing, or nonimplicit ways. This also leads to diminished productivity and a negative user perception of the tool. In their case study of a South African phone company's implementation of intranet technology,
545
Scheepers and Rose (2001) pointed to the necessity of balancing control and individual ownership through strategic management.
Employee Empowerment Versus Strategic Management Intranet technologies allow for a varied "locus of control" for different activities, individuals, and contexts (Mahapatra & Lai, 1999). Users are becoming empowered and beginning to track their own information needs (Harvey, Palmer, & Speier, 1997). One compelling reason for the implementation of intranet technology is this motivation of employee initiative. This empowerment has potent effects, often creating a stronger sense of community among many individuals and groups within the organization (Greer, 1998). Duane and Finnegan (2000) reported the effects of an intranet on employee empowerment, which illustrated the achievement of improved satisfaction and productivity by means of the implementation of an organizational intranet. Although employees can share more knowledge with the intranet, the transition of many internal services becoming more "self-service" in nature, means less face time spent with people outside of a given department and increased social isolation. Many groups, such as corporate travel, have been virtually replaced by online Web sites and forms. There are often critical exchanges that can be lost with the migration to Web sites and forms. Zuboff (1988) painted a graphic picture of the psychological consequences of simplification, isolation, and computermediation of clerical work. In a study of two offices migrating to computer-based clerical work, Zuboff found clerks complaining of physical and mental discomfort (eye strain, nervous exhaustion, physical strain, irritability, enervation, sedentariness, back pain, short tempers and intolerance, and a host of other concerns). Automation removed the need for "bodily presence in the service of interpersonal exchange and collaboration now required their bodily presence in the service of routine interaction with a machine." The resultant sentiments of the users could be prohibitive to bottom-line productivity. In the sense that this technology will be used for human resources planning and learning within the organization, it seems relevant that some lessons from manufacturing can be carried into this discussion. In manufacturing and product development, it is clear that one needs to consider some organizational aspects, human aspects, and social aspects to gain the expected benefits of a new technology (Duffy & Salvendy, 2000). Furthermore, it must be understood how the organization's required structure supports the goal. Strategic management is the most commonly cited factor of successful intranet adoption for intraorganizational communication (Damsgaard & Scheepers, 1999; Duane & Finnegan, 2000; Gonzalez, 1998; Koepler et al., 1998). Damsgaard and Scheepers (1999) cautioned against the use of intranet technologies as a change agent because the intranet tends to replicate existing structures and may morph into a barrier for further change. Instead, they cautioned, the intranet should support the adoption of a new management strategy (Damsgaard & Scheepers,
546
EMERY ET AL.
1999). Intranet technology and management strategies could work against each other if not planned for effectively. Active management is a mission-critical component to optimizing intranet development efforts. Although strategic planning is an important preintranet adoption and development tool (Wachter & Gupta, 1997), intranet-related management has been experienced to intensify in the latter stages of intranet evolution. The intranet, at this stage, is a critical component to workflow (Duane & Finnegan, 2000). Strategic management must address issues of intranet growth, data ownership, data content, and intranet work group coordination (Lai, 2001). Scheepers and Rose (2001) pointed to the following guidelines for managing intranets: • Balance control and individual ownership—Continually evaluate the balance of employee empowerment and standardization as the intranet and organization infrastructure have been shown to evolve. • Cultivate intranet as a medium, not a system—Using collaborative and facilitative managerial style to garnish the individual user and task requirements with the system. Nielsen (2000) referred to this as evangelism outreach because every department needs to be brought on board in the development of standards and intranet usage paradigms they are expected to follow, or else risk the disregard of management mandates. • Enforce tactics for a self-sustaining intranet—Convincing knowledge workers of the potential value of the intranet to their productivity and job satisfaction. This is often accomplished with the introduction of a "killer application" for users or subgroups of users that will draw in users (e.g., employee phone directories, Web-enabled legacy systems).
High Returns on Investment Versus High Costs of Ineffective Design The return on investment of an intranet has been calculated, as has the cost of poor intranet design to an organization. An intranet is a relatively inexpensive means of connecting multiple information system platforms. Use of an organization intranet can provide a significant cost savings with respect to administrative simplification (e.g., pre- vs. postintranet operating costs) that a 100% return on investment should be realized within a period of weeks after successful intranet implementation (Carlson, 2000; Lai, 2001). However, these are often the immediate gains of electronic document conversion to reduce document duplication and distribution costs. Curry and Stancich (2000) advise against measuring intranet progress using only the return on investment value. They advise that several companies calculate their return on investment based on intranet-based publishing functions, as opposed to intranet-based applications. Nielsen (1999) stated that it is common for intranet improvements to have 3 : 1 payout ratio. For example, an organization that invested $3 million in its intranet's usability saved an estimated $10 million per year for its 7,000 users (Nielsen, 1999). The 3 : 1 payout ratio for usability improvements and the 100%
return on investment show how potentially sensitive an organization's success can be to the stability and usability of its internal network. Intranet designers have the keen ability to comprehensively scope user hardware, experience, skills sets, and workflow needs. Human resources can provide much of the tangible information, but additional information needs to be gathered in order to effectively support the organization's rules, norms, and structures, and more comprehensive user requirements. Field studies or ethnographic studies are fairly easily facilitated within the confines of your organization (Nielsen, 2000). Essentially, the goal is to gather information on existing communication infrastructures and workflow through observations of employees during actual work. Accordingly, after initial planning of intranet application development, the exploitation of low-cost usability testing methods should be employed throughout the development of intranet applications, to ensure employee productivity and mitigate changes and misuse with the postlaunch application. Given the severe headcount reductions in many companies in more recent years, one may wonder who has the time or resources to accomplish these activities. Any investigation into existing communication infrastructures is better than none, and will moderate the people hours and other resources required throughout the design and implementation process. That said, tremendous care should be taken to ensure end user productivity and satisfaction with the intranet and its associated tools. It is of great importance to understand and track the skills, experience, and expertise of the people who will be creating, sharing, using, and disseminating information through the use of intranet. Research has demonstrated that in specific contexts of use, such as hospitals, different personnel have accepted and used Internet technologies much more readily than others (Jacko, Sears, & Sorensen, 1998, 2001; Sorensen, Jacko, & Sears, 1998a, 1998b). This is especially true for time-critical work environments that may by highly sensitive to a possible decay in the value of information when issues of design, information presentation, and organization limit information accessibility. Usability heuristics identified to influence intranet design include efficiency, memorableness, and error reduction (Nielsen, 2000). These heuristics and several others have been proposed and explored as analytical tools for assessing the usability or effectiveness of an interface. These attributes directly impact a user's productivity and level of frustration when working with the intranet.
CONCLUSION Intranets, a form of IT, have the capacity to promote gains in productivity, efficiency, and communication infrastructures within organizations. However, these applications of technology will not work independently. "Corporate efficiency goes up dramatically with clearer communication, and the intranet can be the infrastructure for the communication if—and only if—it is designed to make it easy for people to find information when they need it" (Nielsen, 2000, pp. 276-277).
29. Intranets and Intraorganizational Communication
547
FIGURE 29.9. Components of the research and development agenda and their interrelationships (Jacko et al., 2002).
According to Zuboff (1988), IT may induce innovative business initiatives and create an organization where knowledge should be used to make knowledge accessible to anyone with the capacity to understand it. IT is the product of human actions and material assets, and the effective integration of the two can lead to success. Huber (1991) recognized that IT could increase the speed and quality of intraorganizational decisions in organizations supporting an infrastructure of decentralized decision making. Intranet technology, unlike other forms of IT, is more forgiving and customizable to an organization's needs throughout the layers of an organization. Intranet technology and its utilization by organizations have matured since the late 1990s. The early adopters of intranet technologies are experiencing issues grounded in the later stages of intranet evolution, distanced from the problems of document publishing and altering. In their white paper on the topic of intranets and organizational learning, Jacko and colleagues (2002) pooled the expertise of researchers in the areas of intranets and organizational learning. The panel of experts was surveyed on their expectations for the future role of intranets, and ensuing research and development efforts to address their future use.
Experts indicated that the role of intranets in driving a new form of organization is an evolving tool upon which an organizational knowledge infrastructure can be built. Intranets were also identified as driving an organization to new forms of collaborative work, such as virtually distributed teams. Justification of this is the potential for the teams to possess optimal subject-matter expertise and limited time spans for collaboration, which serve as motivators for fast, efficient knowledge exchange. Four major categories for the research and development of intranets and organizational learning were ranked as follows: (a) organizational rules, norms, and structure; (b) changes in the nature of work; (c) knowledge manipulation, and (d) technology issues. Figure 29.9 presents a conceptual framework that resulted from this effort, depicting the interrelationships of the four categories of targets, with knowledge at the center to demonstrate its centrality at the core of this topic. Lam's (2000) proposed typology of knowledge, integrating the cognitive and organizational dimensions of knowledge, is adapted for Fig. 29.9 and shown at its center. "Knowledge manipulation" encapsulates knowledge in this diagram, and is composed of knowledge creation, knowledge structuring, knowledge processing, knowledge distribution,
548
EMERY ET AL.
knowledge interpretation, and evaluation and feedback. It is left to the researchers and developers to investigate the nature and degree of the interrelationships between the targets within each of the four categories of research and development. It is clear from this conceptual depiction and the experts' opinions that next generation intranets have the potential to serve as powerful tools for organizational learning, and that the existing scientific literature serves as a solid foundation from which future research and development efforts can be launched. Intranets provide new communication tools that organizations can use to support their evolution into learning organizations. Intranet use evolves over time to support the changing needs of intraorganizational communication. At the foundation of any organization are the users of information, whom possess different needs, abilities, and responsibilities at different times and places. As in any human-centered system, understanding
user characteristics, and the role these characteristics play in the nature of the user's work flow, can inform the types of tasks and structures that should be supported with intranet applications (and the way in which these organizational aspects can be recorded). The comprehensive, direct impacts of intranets on organizational learning are not yet known. Much of the current research in this area tends to be more anecdotal. This is due, in part, to the difficulty of measuring organizational learning, and the wide array of uses for intranets within and between organizations. Intranets only address a small area within the overall knowledge space of organizational learning. In fact, current applications of intranets for organizational learning focus mainly on encoded knowledge. As organizations integrate broader levels of intranet functionality and more sophisticated technologies, intranet applications will become fully intertwined in the evolutionary process of organizational learning.
References Akoumianakis, D., & Stephanidis, C. (2001). Computational environments for organisational learning. In M. J. Smith & G. Salvendy (Eds.), Proceeding of the 9th International Conference on HumanComputer Interaction (pp. 301-305). Mahwah, NJ: Lawrence Erlbaum Associates. Argote, L. (1999). Organizational learning: Creating, retaining, and transferring knowledge. Boston: Kluwer Academic. Argyris, C. (1982). Reasoning, learning, and action: Individual and organizational. San Francisco: Jossey-Bass. Argyris, C., & Schon, D. A. (1978). Organizational learning: A theory of action perspective. Reading, MA: Addison-Wesley. Baines, A. (1996). Intranets. Work Study, 5, 5-7. Borella, M. S., Sears, A., & Jacko, J. A. (1997). The effect of Internet latency on user perception of information content. Proceedings of IEEE Global Telecommunications Conference: GlobeCom 97 19321936. Bullinger, H. J., Mueller, K., & Kempf, F. (2001). Using integrated platform solutions in work-oriented learning environments. In M. J. Smith & G. Salvendy (Eds.), Proceeding of the 9th International Conference on Human-Computer Interaction (pp. 282-286). Mahwah, NJ: Lawrence Erlbaum Associates. Cangelosi, V E., & Dill, W. R. (1965). Organizational learning: Observations toward a theory. Administrative Science Quarterly, 10, 175203. Carlson, P. (2000). Information technology and emergency of a worker centered organization. ACM Journal of Computer Documentation, 24(4), 204-212. Coleman, D. (1997). Collaborating on the Internet and intranets information systems-collaboration systems and technology. Proceedings of the Hawaii International Conference on System Sciences, 2, 350-358. Collins, H. (2001). Corporate portals: Revolutionizing information access to increase productivity and drive the bottom line. New York: AMACOM. Collins, H. (2003). Enterprise knowledge portals: Next generation portal solutions for dynamic information access, better decision making, and maximum results. Chicago: AMACOM. Collins, H. M. (1993). The structure of knowledge. Social Research, 60(1), 95-116.
Cross, R., & Baird, L. (2000). Technology is not enough: Improving performance by building organizational memory. Sloan Management Review, Spring, 69-78. Curry, A., & Stancich, L. (2000). The intranet—An intrinsic component of strategic information management? International Journal of Information Management, 20, 249-268. Cyert, R. M., & March, J. G. (1963). A behavioral theory of the firm. Englewood Cliffs, NJ: Prentice Hall. Daft, R. L., & Weick, K. E. (1984). Toward a model of organizations as interpretation systems. Academy of Management Review, 9, 284295. Damsgaard, J., & Scheepers, R. (1999). A stage model of Intranet technology implementation and management. In J. Pries-Heje, C. U. Ciborra, K. Kautz, J. Valor, E. Christiaanse, & D. Avison (Eds.), Proceedings of the Seventh European Conference on Information Systems. Dasan, V, & Ordorica, L. (1998). Hands-on intranets. Mountain View, CA: Sun Microsystems Press. Dix, A., Finlay, J., Abowd, G., & Beale, R. (1998). Human-computer interaction (2nd ed.). London, England: Prentice Hall. Drucker, P. F. (1988). The coming of the new organization. Harvard Business Review, 66(1), 45-53Duane, A., & Finnegan, R (2000). Managing intranet technology in an organizational context: Toward a "stages of growth" model for balancing empowerment and control. Proceeding of the 21st International Conference on Information Systems, 242-258. Duffy, V G., & Salvendy, G. (1999). Relating company performance to staff perceptions: The impact of concurrent engineering on time to market. International Journal of Production Research, 37(4), 821-834. Duffy, V G., & Salvendy, G. (2000). Concurrent engineering and virtual reality for human resource planning. Computers in Industry, 42, 109-125. Gill, Z. (2001). Webtank design—Intranet support for collaborative problem-solving, hi M. J. Smith & G. Salvendy (Eds.), Proceeding of the 9th International Conference on Human-Computer Interaction (pp. 292-296). Mahwah, NJ: Lawrence Erlbaum Associates. Gonzalez, J. S. (1998). The 21st century intranet. Upper Saddle River, NJ: Prentice Hall.
29. Intranets and Intraorganizational Communication •
Greer, T. (1998). Understanding intranets. Redmond, WA: Microsoft Press. Guengerich, S., Graham, D., Miller, M., McDonald, S. (1996). Building the Corporate Intranet. New York: John Wiley & Sons. Harvey, M., Palmer, J., & Speier, C. (1998). Implementing intraorganizational learning: A phase-model approach supported by intranet technology. European Management Journal, 16(3), 341354. Harvey, M. G., Palmer, J., & Speier, C. (1997). Intranets and organizational learning. Proceedings of the 1997 SIGCPR Conference, The Impact of the New Connectivity (pp. 100-116). New York: ACM. Hazel, H., & Jones, A. M. (2000). Show off the corporate library. International Journal of Information Management, 20, 121-130. Hedberg, B. (1976). How organizations learn and unlearn. In P. C. Nystrom & W. H. Starbuck (Eds.), Handbook of organizational design (pp. 3-27). New York: Oxford University Press. Hills, M. (1996). Intranet business strategies. New York: Wiley. Hills, M. (1997). Intranet as groupware. New York: Wiley. Hills, M. (1998). Managing the corporate intranet. New York: Wiley. Hope, B. (2001). Service quality in the virtual world: The case of extranets. In S. Dasgupta (Ed.), Managing Internet and intranet technologies in organizations: Challenges & opportunities (pp. 21-34). Hershey PA: Idea Group Publishing. Huber, G. R (1991). Organizational learning: The contributing processes and the literatures. Organizational Science, 2(1), 89-115. Jacko, J. A., & Duffy, V (1997). Interface requirements: An analysis of technologies designed for individual and group work. Proceedings of the 7th International Conference on Human-Computer Interaction, 347-350. Jacko, J. A., Salvendy, G., Sainfort, E, Emery, V K., Akoumianakis, D., Duffy, Y G, Ellison, J., Burley Gant, D., Gill, Z., Ji, Y. G, Jones, R M., Karsh, B., Karshmer, A. I., Lazar, J., Peacock, B., Resnick, M. L., Sears, A., Smith, M. J., Stephanidis, C., & Ziegler, J. (2002). Intranets and organizational learning: A research and development agenda. International Journal of Human-Computer Interaction, 14(1), 93130. Jacko, J. A., Sears, A., & Borella, M. S. (2000). The effect of network delay and media on user perceptions of web resources. Behaviour and Information Technology, 19(6), 427-439Jacko, J. A., Sears, A., & Sorensen, S. J. (1998). The effect of domain on usage patterns and perceptions on the Internet: Focusing on healthcare professionals. Proceedings of the Human Factors and Ergonomics Society 42nd annual meeting, 521–525. Jacko, J. A., Sears, A., & Sorensen S. J. (2001). A framework for usability: Healthcare professionals and the Internet. Ergonomics, 44(11), 989-1007. Ji, Y. G., & Salvendy, G. (2001). Development and validation of intranet organizational memory information system for improving organizational learning. In M. J. Smith & G. Salvendy (Eds.), Proceeding of the 9th International Conference on Human-Computer Interaction (pp. 297-300). Mahwah, NJ: Lawrence Erlbaum Associates. Jones, P. M. (2001). Collaborative knowledge management, social networks, and organizational learning. In M. J. Smith & G. Salvendy (Eds.), Proceeding of the 9th International Conference on HumanComputer Interaction (pp. 306-309). Mahwah, NJ: Lawrence Erlbaum Associates. Koehler, J. W, Dupper, T, Scaff, M. D., Reitberger, E, &Paxon, P. (1998). The human side of intranets: Content, style, & politics. Boca Raton, FL: St. Lucie Press. Lai, V S. (2001). Intraorganizational communication with intranets. Communications of the ACM, 44(7), 95-100. Lam, A. (2000). Tacit knowledge, organizational learning and societal
549
institutions: An integrated framework. Organization Studies, 2/(3), 487-513. Lazar, J. (2001J. User-centered web development. Sudbury, MA: Jones and Bartlett Publishers. Levitt, B., & March, J. G. (1988). Organizational learning. Annual Review of Sociology, 14, 319-340. Mahapatra, R. K., & Lai, V S. (1999). Evaluation of intranet-based enduser training. Proceedings of the 20th International Conference on Information Systems, 524-527. Malhotra, Y. (1996). Organizational learning and learning organizations: An overview [On-line]. Available: http://www.brint.com/ papers/orglrng.htm March, J. G., & Olson, J. P. (1975). The uncertainty of the past: Organizational learning under ambiguity. European Journal of Political Research, 3, 147-171. Martin, E T. (1999). Top secret intranet: How U.S. intelligence built inteUinet—The world's largest, most secure network. Upper Saddle River, NJ: Prentice Hall. Miner, A. S. (1990). Structural evolution through idiosyncratic jobs: The potential for unplanned learning. Organizational Science, 1, 1995210. Miner, A. S., & Mezias, S. J. (1996). Ugly duckling no more: Pasts and futures of organizational learning research. Organization Science, 7, 88-99. Mueller, K. (2001). Learning from work—How to use working tasks in virtual learning environments. Proceedings of the 20th ICDE World Conference on Open and Distance Education. Nambisan, S., & Wang, Y. (1999). Technical opinion: Roadblocks to Web technology adoption? Communications of the ACM, 42(1), 98101. Nielsen, J. (1997). The difference between intranet and Internet design. Jakob Nielsen's Alert Box [On-line]. Available: http://www.useit. com Nielsen, J. (1999). Intranet portals: The corporate information infrastructure. Jakob Nielsen's Alert Box [On-line]. Available: http://www. useit.com Nielsen, J. (2000). Intranet design. In designing web usability (pp. 263294). New York: New Riders Publishing. Nonaka, I., & Takeuchi, H. (1995). The knowledge creating company. New York: Oxford University Press. Rademacher, R. A. (1999). Applying Bloom's taxonomy of cognition to knowledge management systems. Proceedings of the 1999 ACM Conference on Computer Personnel Research, 276-278. Riggins, E, & Rhee, H. S. (1998). Toward a unified view of electronic commerce. Communications of the ACM, 41, 88-95. Rosen, A. (1997). Looking into Intranets & the Internet: Advice for managers. Chicago: AMACOM. Sainfort, E (1987). Innovation and organization: Toward an integration. Dissertation thesis, Ecole Centrale des Arts et Manufactures, Paris, France. Scheepers, R., & Damsgaard, J. (1997). Using Internet technology within the organization: A structurational analysis of intranets. Proceedings of the International ACM SIGGROUP Conference on Supporting Group Work: The Integration Challenge, 9-18. Scheepers, R., & Rose, J. (2001). Organizational intranets: Cultivating information technology for the people, by the people. In S. Dasgupta (Ed.), Managing Internet and intranet technologies in organizations: Challenges & opportunities (pp. 1-17). Hershey, PA: Idea Group Publishing. Sears, A., & Jacko, J. A. (2000). Understanding the relationship between network quality of service and the usability of distributed multimedia documents. Human-Computer Interaction, 15, 43-68. Sears, A., Jacko, J. A., & Borella, M. (1997a). The effect of Internet
550
EMERY ET AL.
delay on the perceived quality of information. Proceedings of the 7th International Conference on Human-Computer Interaction, 335-338. Sears, A., Jacko, J. A., & Borella, M. (1997b). Internet delay effects: How users perceive quality, organization, and ease of use of information. Proceedings of the ACM Conference on Human Factors in Computing Systems, 353-354. Sears, A., Jacko, J. A., & Dubach, E. M. (2000). International aspects of WWW usability and the role of high-end graphical enhancements. International Journal of Human-Computer Interaction, 12(2), 243263. Senge, P. M. (1990). The fifth discipline: The art and practice of the learning organization. New York: Doubleday. Sorensen, S.J., Jacko, J. A., & Sears, A. (1998a). A characterization of clinical pharmacists' use of the Internet as an online drug information tool. 1998 Spring Practice and Research Forum of the American College of Clinical Pharmacy, 846. Sorensen, S. J., Jacko, J. A., & Sears, A. (1998b). Hospital pharmacists'
use, perceptions, and opinions of the Internet. Pharmacotherapy, 18(2), 438. Suresh, S. (1998). Decision support using the Intranet. Decision Support Systems, 23, 19-28. Telleen, S. L. (1997). Intranet organization. New York: Wiley. Tkach, D. (1999). Knowledge portals [On-line]. Available: http://www4.ibm.com/software/data/km/advances/kportals.html Wachter, R., & Gupta, J. N. D. (1997). The establishment and management of corporate intranets. International Journal of Information Management, 17(6), 393-404. Weick, K. E. (1991). The nontraditional quality of organizational learning. Organization Science, 2(1), 41-73. Young K. (2000). Entre nous. The Banker, February, 84-85. Zhang, R., & Chen, J. (1997). Intranet architecture to support organizational learning. Proceedings of the annual meeting of the Decision Sciences Institute, 729-731. Zuboff, S. (1988). The age of the smart machine. New York: Basic Books.
30: WEB-BASED COLLABORATION Martijn van Tilburq Microsoft Corporation
Timothy Briggs Microsoft Corporation
In terms of place, the group of collaborators can either be situated together or the group can be distributed. It is important to note, however, that even groups that are primarily co-located can benefit from features supporting less proximal interaction (Olson & Teasley, 1996). In terms of time, a task can be synchronous or asynchronous. During synchronous collaboration, people work together at the same time, discussing more granular topics through means such as audio-, video-, graphics-, and text-based systems. During the more common asynchronous collaboration, contributions are captured and stored for a longer period of time, for instance, using e-mail, teamware databases, and conferencing groupware. Collaborative applications can also be categorized by the granularity of the tasks they support, as shown in Table 30.1. This granularity can range from being a really specific task (like making a decision together) to being a very large and complex set of tasks (like working on a team project). The scope of the information used or created for these tasks can also distinguish collaborative applications. For example, the information shared within e-mail applications and discussion boards is topic based, within conferencing software it is object based (like a document), within teamware it is project based, and within intranets and portals it is corporate based. Finally, differences in time, place, task granularity, and scope of information can translate to critical differences in a user's frequency of visit. People might check e-mail hourly, where they might only consult the corporate intranet once a month—an important consideration in the design of each one.
INTRODUCTION Collaborative systems have been around for decades. Most frequently, they have used a dedicated client-server architecture to support the common goals, communication, and work products that typify this area (e.g., sharing documents, managing process, discussing tasks). The successes and failures of these systems are well represented in the industry and academic computersupported collaborative work (CSCW) literature. Now, like so many other domains, collaborative systems are becoming more Web-based as Internet and intranet technologies are ideally suited to support the broad access that such systems require. However, this shift to Web-based collaborative applications introduces yet another layer of opportunity and challenge to the design of the foundations and individual installations. This chapter focuses on the intersection of designing collaborative systems, specifically teamware, and designing for the Web. The research, guidelines, and solutions in both domains are generally still applicable, but they take on unique qualities when viewed from the perspective of Web-based collaboration. Overview of Collaborative Systems. With such a broad definition, collaboration encompasses a lot of applications. A common way, as followed by Wheeler, Dennis, and Press (1999), to categorize different types of applications is to distinguish between place and time dimensions and then the different aspects within each, as shown in Fig. 30.1.
* Microsoft, Windows, Office, SharePoint, Internet Explorer, and Word are either registered trademarks or trademarks of Microsoft Corporation in the United States and/or other countries.
551
552
VAN TILBURG AND BRIGGS
FIGURE 30.1. Overview of collaboration systems. TABLE 30.1. Collaborative Applications by Granularity of Task Example application
E-mail, instant messaging
Example task
Discussions and decision making
Scope
Teamware
Enterprise portal systems
Content creation and review
Project and team coordination
Knowledge management
Topic
Document or object
Team content
Corporation content
Typical visit frequency
Hourly
Daily
Weekly
Monthly
Typical type of collaboration
Synchronous and asynchronous
Synchronous and asynchronous
Asynchronous
Asynchronous
Conferencing and collaboration tools
The differences described have produced a myriad of CSCW applications. Because most work consists of a combination of activities that fall in all quadrants of the matrix and that span from individual content to organizational content, supporting this range of work is nearly impossible for a single application. Moran and Anderson (1990) discussed three such paradigms as the shared workspace, coordinated communication, and informal interaction paradigms. A very generic tool that tries to do all would not solve the unique problems within a quadrant adequately (Dix, 1997). Therefore, computer-supported collaborative work usually takes place in a multitude of tools with a minimum of integration and user interface (UI) consistency. Web-Based Collaborative Applications. There are Web-based applications to support much of the collabora-
tive work in Table 30.1 (Dix, 1997). However, many of them are direct descendants of earlier non-Web products. The Web version is often a copy of the earlier functionality or merely a Web-based access point to the data. Examples of this are Webbased e-mail systems, conferencing tools, and Web frontends for groupware databases. Often, little attention is paid to the intersection of design for collaborative work and design for the Web. This is less true for the category of teamware applications. We define teamware as a collaborative application that intends to support the needs of a team. However, first, what is a team? We generally ascribe to the definition used by Guzzo and Dickson (1996), which uses "team" and "group" interchangeably, "recognizing that there may be degrees of difference, rather than fundamental divergences, in the meanings implied by these terms" (p. 309). Only when it becomes important for design decisions
30. Web-Based Collaboration
553
do we tend toward a tighter description of teams as a subset of a group with perhaps more shared work items and goals. Many teamware applications have been developed specifically for the Web because of the benefits to collaborative work that are afforded by Web technologies. First, members of collaborative groups (especially more distributed ones) are more able to access Web-based teamware applications because of the Web's platform independency. Second, the success of teamware applications is highly dependent on the ability to customize installations; the Web provides a relatively simple protocol for making changes (Morris, Neilson, Charlton, & Little, 2001). Third, the browser is a tool that is already available on many computers, so it does not directly require the learning and installation of new software (Bentley et al., 1997). Thus, the Web architecture can lower some of the hurdles for teamware noted by Grudin (1999): establishing critical mass, unobtrusive accessibility to features, and customization in the adoption process. For this chapter, we focus on Web-based teamware applications because of their specific relevance to the intersection of designing for CSCW and designing for the Web.
elsewhere (e.g., Orlikowski, 1992). Web-based collaborative applications add yet another level of possible failure points.
ISSUES IN WEB COLLABORATION
Building in Process. The necessary development work to make an organizationally selected solution useful at the team level frequently falls to the organization's technology support group. They are more likely to have the training and experience with the Web's quickly evolving standards and technologies—more so than at the team level. Even if this additional development cost is accurately budgeted for (which is rare in our experience), the centralized information technology (IT) groups rarely possess sufficient knowledge of a team's needs and process to accurately customize the solution in meaningful ways. Although this customization is critical in initial development, the success of systems usually relies on optimization and tuning that can only come from use. Indeed, many organizations choose Web-based systems on the promise of adaptability once rolled out. However, often development projects taken on by IT groups do not contain sufficient planning for subsequent maintenance. Neither time nor resources are budgeted for additional work. Therefore, the necessary tuning rarely takes place and the resulting inefficiencies adversely affect reaching the "critical mass" that contributes to success.
Successful design of any system requires an understanding of the requirements and issues the system is intended to support. The specific users and use of the system must be investigated, understood, cataloged, communicated, and evaluated against throughout the entire design process. (For more information, see Volk & Wang, chap. 17, this volume, and Zhu, Vu, & Proctor, chap. 18, this volume). This section describes a three-part structure that can act as a catalog of what must be investigated pertaining to a particular Web-based collaborative installation and as a general list of issues that seem to face all such collaborative systems. The issues of Web-based collaboration can be divided into three components: • Organization issues describe the constraints of the larger environment. • Team issues involve the goals and needs of the intended collection of individuals. • End user issues affect an individual's performance at the interaction level. Issues obviously overlap the components and thus the categorization should not be considered mutually exclusive. It supplements but cannot replace the necessary work to exhaustively describe a specific installation.
Organization Issues The general CSCW literature and our own investigations are studded with examples of good-intentioned and otherwise welldesigned collaborative systems eventually failing in practice due to organizational issues. These issues are well documented
Selecting a Web Architecture. For some organizations, Web-based systems hold the promise to address three major cost points: software licensing fees, maintenance, and training. The hope is that expensive desktop applications can be replaced by ubiquitous (cheap) browser frontends that can reduce not only number of seats purchased, but also the cost of upgrading and maintaining them. The browser frontend promises little additional training because internal clients are familiar with standard browser interaction. Unfortunately, as Dix (1997) pointed out, solutions purchased for these reasons and at the organizational level versus team level are usually too generic to be useful to individual teams—they lack sufficient power or customization to address the process and business needs of each group. To support the various teams' needs, they need significant customization, which in turn necessitates additional training costs. Total cost of ownership can far exceed initial purchase and installation.
Changing, Then Maintaining the Culture. Our own ethnographic work supports the major adoption findings in the CSCW literature—collaborative systems often fail because organizations do not adequately foresee and account for the required changes in the organizations' culture. New systems mean new work practices. For instance, collaborative systems supporting document storage are usually replacing the sharing of files through shared network folders. Competing with this familiar and successful method requires a Copernican shift in how users are expected to collaborate. As addressed previously, sometimes this shift never happens and the system never reaches the critical mass that will sustain it. Equally common though is an initial adoption followed by a slow erosion: Gradually the users stop trusting that the new system contains the most up-to-date information, they themselves start using the old
554
VAN TILBURG AND BRIGGS
system, and the entire effort spirals down through a cycle of disuse. At roll out, collaborative systems need strong technology mediators and champions at the organizational level, and then throughout the long adoption phase. This is even more true for Web-based systems, which are usually radically different than existing methods.
Team Issues We define the "team" or "group" to be the collection of individuals with shared work goals, products, and dependencies— organizationally determined, or "virtual." Much has been written about supporting geographically dispersed teams through technology, usually with the goal of more tightly coupling their work to a point approaching co-location (Olson & Teasley, 1996; see also Kiekel & Cooke, chap. 6, this volume). However, our interests have focused on the more common arrangement of teams in close geographic proximity. The challenge here is not re-creating co-location, rather it is finding ways to improve the sharing of information by augmenting the usual communication methods including face-to-face methods. Being Aware. The common takeaway from the team situation awareness literature (e.g., Dourish & Bellotti, 1992; Gutwin & Greenberg, 1999; Posner & Baecker, 1993) is that in order to be effective, teams need to be aware of the content, status, changes, and work they share. Presence, knowing when a team member in a space, has become an especially hot topic. As Dix (1997) noted, this awareness may be explicit in the form of direct communication and discussion, or implicitly derived "through the artifacts themselves." Either way, awareness poses significant problems for Web-based systems. First, because these systems are accessed through the browser, they are most often outside the team members' immediate work context. Thus, users must either set goals to visit the space (thereby reducing the frequency with which they can receive updates) or be notified through other communication channels (interrupting work and further fracturing the collaboration into yet another context). Second, reconstructing the history of changes is made difficult by the usual lack of metainformation, such as accurate depiction of who, when, how, where, and why an explicit change was made. Web-based systems, because they are not as integrated at the system level, are more likely to lack important information that can directly or indirectly contribute to a team's awareness. Communicating. Effective collaboration requires lots of communication. Particularly in co-located teams, this can occur through many different channels: in person, across cubicles, on phone, in e-mail, through work products, in system-based discussions, and so on. More important, note that all except the last two usually occur outside a collaborative system. The user is expected to live within or visit the system in order to communicate. This is unrealistic. Individual users may belong to multiple -workgroups and thus need to constantly switch between spaces to monitor communication. Thus, the solution is not adding more communication into teamspaces, but rather integrating information from multiple teamspaces into users'
primary personal communication tools (notably e-mail). This has the additional benefit of leveraging users' existing communication tool knowledge, but also better supporting the escalation of conversations through multiple channels: voice mail into e-mail, e-mail into messaging, and so forth. Improving Existing Methods. For obvious reasons, the most frequent communication channel among team members who are in close geographic proximity is face-to-face meetings. Where do Web-based collaborative applications fit in this realm? They should seek to support rather than replace meetings. Setting an agenda, distributing discussion materials, taking attendance, capturing work items—the actual meeting may be face-to-face, but the preparation should be supported online. That said, one measure of a system's success might be the reduced need for face-to-face, meetings. The system and its contents might enhance effectiveness to the point that fewer scheduled or ad hoc meetings are necessary. Expanding Audiences. Teams have shared work products. A goal of all teamware is to support the production of that work, ideally through three phases: production, publication, and archiving. Production, as in the authoring of a document, usually involves a single individual or at most one to two other team members. The needs of the audience for this phase are very different than for publication (the sharing of work with other team members) or for archiving (the storing of work for unknown possible utility to an unknown audience). One very important issue for any collaborative system, but especially Webbased ones, is adapting the functionality of the space to support the work as it changes across phases. The most common solution for many systems at the moment is creating new shared workspaces for each phase. The unintended result, however, is multiple workspaces all containing roughly the same document and/or collateral materials, introducing questions of which version is the most recent, which should be accessed by whom and for what purpose, as well as the increased server demands for storage.
End User Issues The organization issues can constrain the success of a collaborative system. The team issues can constrain the work of its members. Finally, being Web-based can constrain the actions of the user. The general usability of Web-based systems is discussed in other chapters of this book (see Zhu et al., chap. 18, this volume). Suffice it to say that Web-based collaborative applications are probably subject to most of the issues illuminated elsewhere. Therefore, we concentrate on the most salient and distinct issues befalling the end user of a Web-based collaborative application. Rich Expectations. In their infancy, Web-based collaborative applications were very rudimentary in their functionality. The next generation promises to add much more functionality, making them viable products in which to spend extensive durations of time (Wheeler et al., 1999). As this happens, users will come to expect a richer experience, and richer interaction.
30. Web-Based Collaboration We expect that this will translate to more problems such as those we have witnessed in our early evaluations of other teamware, to be discussed later: • Users searching the browser menus for system functionality found only in the content area of the window • Users trying to right-mouse click to bring up context menus, which in this case provide only commands for navigating the browser, not the system • Users expecting to be able to drag and drop within and between the browser windows and components We also expect that to support this richer interaction, new controls and elements will have to be invented, further adding to the learning curve. Replacing the Venerable File Share. One of the most common methods for sharing files in use today is shared network folders. The shared context and work of a team is frequently contained in folders they can all access to store and retrieve work. As this is the method that Web-based collaborative applications will most likely be replacing within an organization, we expect that the familiar containment metaphor of folders will generate powerful expectations at the interaction level. However, many of the collaborative systems today use a general database architecture for storing the work products, along with the metadata that describes them, such as subject, project, collaborators, and status. This explicit metadata affords better searching, viewing, and tracking. In file share systems, some of this is accomplished by users moving items between folders, for instance, moving a report into a project's FINAL folder to represent its new status. In a database system, this is more likely to be done by changing the value of the Status property to FINAL. This different action can represent a significant change (and perhaps even degradation) of the user experience. Poor Integration. Currently, most Web-based systems are poorly integrated into the users' work. They are separate systems that are merely "visited," with most of the effort to produce, review, and discuss work products to be done elsewhere. Users therefore are constantly switching contexts, introducing the usual problems related to transfer of learning and varying interaction expectations. However, there are other problems that are particularly important within Web-based collaboration: replication or loss of important information between contexts. Often, shared work has to be described in two places, once within the work product itself and once in the metadata that allows others to find it.
TEAMWARE
555
programs. One aspect of teamware applications is generic containers that need extensive customization and population of existing business data. Another aspect of teamware consists of one place that brings together specific tools used to collaborate around this content. Examples of this are discussion tools, instant messaging, desktop application sharing, and Web conferencing. In general, the different teamware applications offer similar collaboration functionality. However, there are fundamental distinctions based on technology and focus. In terms of technology there is the distinction between peer-to-peer technology and a centralized server-based technology. The first technology does not require a server for storage and offers great off-line access to the information. The server-based technology provides central storage of information and does not necessarily require any installation on the client, besides a generic Web browser. In terms of focus, there is a distinction between teamware applications specifically aimed at vertical markets, for example, professional services or product development and applications with a more generic purpose. Microsoft Windows® SharePoint™ Services (WSS) is a centralized server-based teamware application with a more generic purpose.
Case Study Microsoft WSS technology allows users to create ad hoc Web sites to store and create content specific to a team or project. A WSS site is composed of a homepage and various building blocks in the form of lists (tasks, discussions, events, announcements, surveys) and document libraries. A document library is a special type of list that stores uploaded files and their associated custom information. A WSS site can also be further subdivided into other WSS sub-Webs to provide a more granular level of focus. WSS can work as a stand-alone solution or as part of a SharePoint Portal Services installation, which is intended to handle an entire organization's portal and document management needs. At press time, WSS is in its second version. Design Goals. The goal in designing the WSS teamware application was to create a generic collaboration Web site that provides a browser-based user experience consistent with other productivity tools, while allowing for extensive customization to reflect specific business logic. The structure of a WSS site is intended to support this, as shown in Fig. 30.2. WSS seeks to address many of the issues introduced previously (especially integration into the users' productivity tools). As the product matures, we expect to further refine and enhance its interaction and feature set based on feedback and formal evaluations. Different user roles. The interaction with the Web site can be split into four distinct activities and corresponding user roles:
Definition of Teamware A teamware application is a persistent Web-based virtual workspace composed of multiple tools that coordinate the collaboration and information exchange activities of workgroups around business processes, projects, or cross-functional
• Readers—all users are considered readers as it is the most common activity and entails consuming all the information on the site. • Contributors—anyone who adds content to the site, such as documents or discussion items.
556
•
VAN TILBURG AND BRIGGS
FIGURE 30.2. The WSS site structure supports creation and customization of sections, lists, and items. • Designers—those who change the site to reflect specific business needs, including adding and customizing lists and document libraries. Note that the design goal here is to allow any user to be a designer and, therefore, design activities assume little HTML or systems knowledge. • Administrators—more advanced users who manage the server and especially access rights. In the design of WSS version 1, we mainly focused on the first three activities. Obviously, within any one user you might find all these types of activities. However, the way of interacting within each role is quite unique and the goal was to design a user interface that supports all these uses in the best way. For most users, the time spent doing any of the first three activities
(reading, contributing, and designing) is not equal. Reading is assumed to be done most frequently, contributing less so, and designing is done sporadically. Supporting user roles. The first layer of the Web site is optimized for reading—consuming information assumed to be the most frequent activity. The application is structured like a Web site, using a standard navigation model and links. The site has a homepage with a summary of the information that is stored in it. A main navigation bar at the top of the site allows for efficient access to documents and lists, as shown in Fig. 30.3. If the user clicks on Documents and Lists in the top navigation bar, he or she is brought to a page displaying all the containers that have been used to collect the information: documents,
30. Web-Based Collaboration
557
FIGURE 30.3. The WSS homepage can be customized to allow efficient access to the most important information. pictures, lists, discussion boards, surveys, and so on, as shown in Fig. 30.4. The contents of each container or collection is displayed in a dedicated page, with options and interface optimized to work with the type of information they contain, for instance, documents as shown in Fig. 30.5. It was decided to make the editing functionality (changing previously posted information) more hidden and require a deliberate action to initiate. Every time a user wants to edit something, they would first have to browse to the item and then click the "Edit" button. At that point they would be presented with
the necessary options to make changes, as shown in Fig. 30.6. Although this is a great optimization for the reading activity, unfortunately, the sacrifice can be less efficiency for the contributing activity. The third activity is designing. Although the time spent doing this is least, this activity is probably the most important one. As discussed under "Issues in Web Collaboration," a collaboration Web site can only be useful if it reflects the way work is done in an organization. WSS is an empty frame that needs to be customized to reflect its work context. This customization activity will most likely be done just after the site is deployed
558
•
VAN TILBURG AND BRIGGS
FIGURE 30.4. The Documents and Lists page has comprehensive access points for all content. and intermittently during the use of the site. Because of the sporadic nature of this activity, it was decided to create a more verbose user interface for this task that optimizes for learnability over efficiency. To achieve this, every part of the site that can be customized uses a structure of an overview page where the user can read all the current settings and options, as shown in Fig. 30.7. From this overview page, the user can go into more detailed control pages that are designed with more explanation and description, as shown in Fig. 30.8. In the context of role-based design, reading, contributing, and designing are thus seen as distinct activities, occurring in
isolation from each other. However, to support switching between these activities, entry points into the other activities are accessible from all levels of the site. For instance, the contributor role is supported by having generic "create" functionality on the top level and context-specific "create" functionality at each sublevel. The designer role is supported in a similar vertical way by having high-level customization tools available at the top level and more context-specific tools elsewhere. Consistency with productivity applications. Because a teamware application is rarely used in isolation from other tasks, special attention needs to be paid to the applications that are
30. Web-Based Collaboration
•
559
FIGURE 30.5. List pages such as Shared Documents are optimized for browsing and reading tasks. producing and viewing the content around which the collaboration is taking place. Examples of these other applications are word processors, spreadsheets, and e-mail software.
User Interface. Most of the content-producing applications use the traditional graphical user interface (GUI) model of direct manipulation, operands, and operators. Teamware, in contrast, is typically Web based and might be used in conjunction with other Web sites. This means that some of the user interface aspects unique to the Web (hyperlinks, browsing within
one window, back button) should also be honored. WSS was designed has a hybrid of both models: The main structure is designed as a Web site, but some familiar GUI elements are also used. The challenge is in denning and clearly communicating which model is used in which circumstances. For instance, the decision was made not to use a selection model or right-click functionality within the WSS environment. In GUI applications, one expects to be able to select an item (a file icon, a word, an object) and then to operate on that selection by choosing commands in menus or right-clicking to see a contextual menu. For the most part, this behavior is unexpected
560
•
VAN TILBURG AND BRIGGS
FIGURE 30.6. Editing previously posted information is less common than reading; thus, it has limited access, but controls that designers can customize.
within a Web browser. To maintain a clear model that WSS is Web based, other solutions had to be sought for allowing these frequent and well-rehearsed interactions. For example, the context menu usually invoked by right-clicking became a drop-down menu accessed by an arrow that only appears on hover, as shown in Fig. 30.9Distinct from that is the use of toolbars and list headers within WSS—features not usually associated with Web browsing. Although most of the operations that users perform in WSS could be handled by standard hypertext links (and many indeed
still are, such as switching views), toolbars are used to handle the most frequent operations in WSS, managing lists. Choosing from icons in a toolbar and clicking column names in list headers are familiar interactions in GUI interfaces such as Microsoft Windows Explorer and Microsoft Office. These interactions were carried forward to WSS, as shown in Fig. 30.10. These examples are intended to show the blending of typical GUI and Web-based interaction to recognize the need for user interface consistency between WSS and productivity applications.
30. Web-Based Collaboration
561
FIGURE 30.7. The settings page is intended to support the infrequent task of designing, and thus is more verbose and comprehensive.
Task Integration. Consistency can also refer to the integration of tools throughout a task. A typical collaboration task rarely starts and finishes within one application. Often the content is created in one application, then shared and discussed in another, and possibly archived in yet another. Especially because teamware is often not optimized for content creation, it is important that there are the right "hooks" between them. For example, it is important that the user can save a document directly to a teamware application from within a productivity application. A common scenario is a user working in Microsoft
Word and needing to post the active document to a WSS site. Rather than forcing the user to save the document to his or hard drive and then upload it to the site, they should be able to invoke the Save As dialog from within Word, browse to the site, see the files list from the site in the familiar WSS presentation, and make their decisions to save from within that dialog, as shown in Fig. 30.11. Other examples of integrating collaborative and productivity tools via "hooks" between them are inline discussions and document workspaces. A document workspace is a special type
562
VAN TILBURG AND BRIGGS
FIGURE 30.8. The WSS controls page supports infrequent customization and personalization tasks using a verbose UI.
of WSS site that allows users to manage information around a document, such as members, files, tasks, and status. This information can be displayed in Word via the task pane that appears to the right of the active document in Word, as shown in Fig. 30.12. The tasks supported by a document workspace can be accomplished either in the Shared Workspace task pane or on the WSS site. Although they appear in different environments, they have similar experiences and allow the user to switch between contexts.
Major usability findings. Throughout development of WSS, usability evaluations were conducted to understand interaction design problems (for further information, see Zhu et al., chap. 18, this volume). In addition to the typical, detailed usability problems found, a few implementation-agnostic issues appeared, which hold certain import for the design of any Webbased collaborative application. Web UI Model Issues. Our findings support the delicacy of the right balance between traditional GUI and Web UI.
30. Web-Based Collaboration
Although some familiar aspects of traditional productivity applications were used in WSS, users continued to run into limitations typically associated with Web applications. For example, we observed users trying to right-click on several different sections of the Web site. Typically, right-clicking in a Web browser invokes the standard Web page context menu (Fig. 30.13). This frustrated users who expected to see a context menu containing commands specific to the target on which they had clicked—for instance, right-clicking a file icon in a WSS site to open or delete the file. The standard Web page context menu does not contain such commands. We also observed users continuously looking in the browser menus for specific WSS functionality, for example, to customize the appearance of a WSS site. Because WSS is somewhere between an application and a Web site, it was unclear to them which environment contained their desired functionality. Based on these findings, changes were made to the design of WSS. The goal of these changes remained to optimize for the browsing experience and to maintain the overall feel of a Web site, while allowing control similar to a GUI application. One example is the mouse-over control discussed earlier, which can approximate a right-click context menu. Another example is the addition of a data sheet control to support the editing of multiple items in a list, as shown in Fig. 30.14.
FIGURE 30.9. GUI context menus were replaced by a dropdown menu accessed by an arrow that only appears on hover.
563
This control turns the entire list into an editable grid, similar to a spreadsheet. After making the changes, users confirm and return back to the read layout of the list. In a way, the division between read and edit mode has been expanded from item level to list level. Still there is a continuous trade-off in how much of the familiar GUI to implement, while maintaining the benefits of a Web site. Especially concerning issues around list manipulation and customization, there is a balance of how much of this should happen in-place within the page or list (possibly sacrificing the Web feel) or in a separate place, such as a customization Web page (possibly sacrificing the advantage of using familiar mechanisms).
Increasing Integration Issues.
Currently wss is
primarily a place where team information is stored and accessed. This singular focus keeps consistency issues to a minimum—our usability evaluations indicate that users have a relatively clean model at this level of where particular functionality can be found and how to interact with it. In general, users now work with content in productivity applications and manage it with WSS. From field work, we understand the need for more direct collaboration. However, as WSS matures and begins to consolidate more and disparate collaboration functionality there is some concern that the resulting hodge-podge of tools and functionality will lead to an increase in usability issues. Future design and evaluation efforts will remain concentrated on finding the right balance of supporting environmentwide integration. An extreme example of this issue is the infusion of realtime collaboration tools. Capabilities such as screen sharing and video conferencing may be supplied by others, but accessed through more generic teamware site, perhaps WSS. These tools may not follow the same mapping of functionality or model of interaction, thereby creating in the user confusion as to where the teamware/conferencing integration and experience begin and end. The challenge, as with any emerging framework, is to develop a concise application model for where and how such components can plug in and out of the foundation. Customization and Personalization. Our usability evaluations indicate some confusion around who controls specific parts of a site's experience and where to go to make such changes. As noted earlier, customization of teamware is critical to a site's utility and therefore survival—it needs to grow and adapt as the needs of a project or team change from initialization through to completion. Typically, such customization for all members is done by an administrator or project leader who is fulfilling the designer role. Equally important though for the same reasons is the personalization that individual members can undertake to make a site more useful and usable for themselves.
FIGURE 30.10. Choosing from icons in a GUI-like toolbar were transferred to WSS.
564
•
VAN TILBURG AND BRIGGS
FIGURE 30.11. Integration between content-producing applications and content management applications is important, such as saving to a WSS from a word processing environment.
Designers customize the site for all members; individual contributors or readers personalize the site for themselves. For instance, a collection of different views on a list might contain views available to all users (customized), as well as views an individual user has created and are thus only available to that user (personalized). In the current design of WSS, both activities are supported by the same user interface. We have noticed that these issues often lead to confusion about where to go and who has permissions to make certain changes, and the effect of such changes. For example, we observed users changing the display properties of a list to only highlight their contributions; however, their actions inappropriately affected the display of that list for all users. Separating and clarifying the customization and personalization of a site remains a high design priority in WSS, as it should for any teamware application. Tangential, but related to the previous issues, is the need for different representations of the team information based on the phase of the project's life cycle. This is the expanding audience problem discussed earlier. Initially, a WSS site might be used by only a few team members for preparation, then its contents need to be shared with a larger audience for feedback, then it is shared with a larger audience for information, and finally when the project is completed, the information needs to be archived.
All these situations warrant a different representation of the information and thus a way to customize it as such. Customization for this purpose remains a design priority.
GUIDELINES FOR WEB COLLABORATION Building from the general issues in Web collaboration and based on the experience of designing and evaluating WSS technology, some general guidelines emerge for building and customizing Web-based collaborative applications and specifically teamware. It should be noted that, on the whole, Web-based collaborative applications should follow good Web design principles wherever possible (see Mayhew, chap. 19, this volume). However, because Web-based collaborative applications are primarily used in parallel with the more traditional GUI-based content-producing applications (as well as different Web-based components), special attention and solutions may be required. 1. Organization Guidelines 1.1. Total cost of ownership Planning for deployment of a Web-based collaborative application should include careful consideration for the
30. Web-Based Collaboration
565
FIGURE 30.13. A typical context menu opened by rightclicking in a Web browser.
FIGURE 30.12. Dedicated tools such as the document workspace can also support the authoring/management integration. continuing effort to update and maintain the system. Where possible, systems should allow team members to do most of the necessary customization and personalization. However, to ensure the adoption of the system, organizations should have clear plans for allocating the time and resources needed to address circumstances where significant development is required. The most common examples are the porting or connection of legacy information to the new system and the integration of important tools into the system. Our field work indicates that failure to do so is one of the leading causes of abandoned systems. 1.2. Language Processes and objects will have certain established names within an organization. Familiarity with new systems can be created by using those words within the Web-based collaborative application. 1.3. Structure Hierarchical structure of the functionality within the Web-based collaborative application should follow organizational practices. The roll out of a new Web-based
collaborative application may not be the time to reengineer business processes, organizational charts, and service models. The cost of learning system-independent changes may overwhelm the learning curve of a new Web-based collaborative application. 1.4. Customization As discussed throughout, a significant key to the adoption and success of a Web-based collaborative application is customizing it to meet the needs of a specific installation, and continuing to do so throughout its life cycle. This cannot be overstated. 1.5. Championing the solution Because of the initial inertia for a group to adopt a new system, garnering buy-in can be difficult. Championing the solution can take many forms, but essentially entails helping prospective users see and understand the value of the new system over existing methods. 2. Team Guidelines 2.1. Web-based collaborative applications as a hub Web-based collaborative applications can be seen as a hub for all team collaboration and communication, especially in the teamware scenarios. The following is a list of different functionality that should be considered for any specific application. The necessity or selection of any one piece of functionality over another is a critical and often difficult choice for the administrator of a specific solution. To this degree, it is often wise to create a test site (or "sandbox") to experiment with different combinations. If a Web-based collaborative application
566
VAN TILBURG AND BRIGGS
FIGURE 30.14. The WSS datasheet view supports spreadsheet-like operations. is the hub, the spokes are storage, communication, information management, and awareness. 2.1.1. Information storage Information storage refers to the functionality around capturing and storing of information around which users might typically collaborate. The customization challenge here is choosing a useful and distinct set of functionality so it is clear to users where to store different types of information. 2.1.1.1. Document management (and pictures) Functionality to store documents related to the project. This store could have additional functionality, such as checkin/checkout, versioning, folders, and filter-based views. A critical consideration for this spoke is the design and capture of metadata around each document (e.g., project, document type, purpose, status). These are the attributes by which subsequent users will find stored information. Prior to deployment, significant thought and experimentation should go into denning what metadata is important to a team—it can be costly and painful to change metadata schemes once a storage space contains a large amount of material. 2.1.1.2. Business processes Functionality to create and use templates that users fill in for applicationspecific business workflow, such as
2.1.1.3.
2.1.1.4.
2.1.1.5.
2.1.1.6.
2.1.1.7.
2.1.1.8.
Issue Tracking or Approval. The fields and their values of the form can then be displayed as columns and cells in a list or table. Surveys Functionality that allows polling of the team members and analysis of the acquired information. Announcements Functionality to store and share general information with the team members. Contacts Functionality to store and manage information about people. Integration with an organization's personnel system (organization chart) and primary communication service (e-mail, instant messaging) is highly valued. The latter will become even more important as such things as real-time communication become more popular. Discussion forums Functionality that allows users to ask and answer questions, and view those in a threaded fashion. Link libraries Functionality to store collections of related uniform resource locators (URLs). Calendar Functionality to maintain a shared project or events calendar. As with Contacts, integration with an organization's
30. Web-Based Collaboration de facto personal or project scheduling system is highly desirable. 2.1.1.9. Task management Functionality to keep track of and assign action items. A key consideration in the choice of this functionality is its relation to the other timebased types of information available— calendaring, project scheduling, and even events/announcements. 2.1.2. Communication Effective collaboration requires specific functionality for communication. The choice and final design of the following components should reflect the organization's culture around communication. (Sproull & Kiesler, 1991). It should also be noted that this communication spoke of the collaborative hub as the most likely to occur outside the Web-based collaborative application, and therefore poses special issues such as integration and capture of relayed information. 2.1.2.1. E-mailing Functionality for accessing, composing, or even archiving e-mail messages. 2.1.2.2. Web conferencing Functionality for application sharing and "white-boarding." This may or not also support live-stream videoconferencing. 2.1.2.3. Instant messaging Functionality for instant, synchronous collaboration around team topics. This could also manifest itself in the form of chat rooms. 2.1.2.4. Meeting support tools Functionality that allows for better team meetings. This could be support before (agenda), during (Web cam), and after (archiving) the meeting. 2.1.3. Information management Information retrieval is an important aspect in collaboration. The goal and value of any collaborative environment is being able to efficiently access and update information relevant to the rest of a team's members. Breakdowns in this area are usually, however, not the result of flaws in the retrieval or notification systems—they are inadequacies of the input and update systems that prevent contributors from maintaining detailed and accurate information. There can be an often fatal cycle of disuse: The harder it is to efficiently update the information, the less likely users are to do so, which can lead to a lack of trust that the system contains the most up-to-date information, finally leading to team members resorting to other methods of collaboration.
•
567
2.1.3.1. Search and filtering Functionality to search and filter all the team content that is stored in the teamware application (see also Fang, Chen, & Chen, chap. 11, this volume). 2.1.3.2. Notification Functionality to monitor changes in the content and be notified through communication applications such as e-mail or instant messaging. 2.1.4. Awareness Teams need to see what is happening and what has happened so they can understand the reasons and react to them. 2.1.4.1. Status and history Functionality that allows users to see status of the project and status of individual parts of the project. 2.1.4.2. Presence Functionality that allows users to see whether team members are online and what they are currently are working on. 2.2. Access to collaboration information outside the system As mentioned pertaining to communication components, sometimes important interactions occur outside of the Web-based collaborative application. One solution is to increase the degree to which the Web-based collaborative application is acting as a hub, moving interaction in external applications into the system so the system can be aware of it. A more realistic solution is bringing the information in the system out to the applications, especially communication applications. Rather than building e-mail or instant messaging into the system, those communication applications should be able to recognize, display, and act on shared collaboration information. 2.3. Control over content 2.3.1. Contextual create structure Functionality to create new content or to upload information. Each level of the site should have a clear means to create corresponding content. At the higher levels of the site, this means the ability to create new containers or add functionality. At the lower levels, this means the ability to add information to those containers. 2.3.2. Contextual customization structure Functionality to customize parts of the site. This should be available on all hierarchical levels of the Web application, based on the context of that specific level. For example, at the higher levels of the site, users should be able to customize/personalize pages; at the item level, users should be able to edit the properties of any item. 2.4. Change focus of space The teamware application should provide customization functionality that allows users to change the focus
568
VAN TILBURG AND BRIGGS
of the space to support the work as it changes across phases (initial plan, sharing information, archiving). 3. End User Guidelines 3.1. Consistency with supporting applications Web-based collaborative applications are only part of a larger set of productivity and line of business applications. All these applications will be used interspersed and in parallel. There should be basic user interface consistency among these applications. This does not mean that the tools have to follow traditional GUI guidelines literally, but there should be similarity in aspects such as the placement and appearance of familiar items [e.g., commands and objects (documents)], as well as consistency of language. 3.2. Consistency with Web applications Typically the Web-based collaborative application lives within the browser and, consequently, traditional Web user interface conventions should be followed. The main aspects are appropriate use of hyperlinks; the use of a model where pages load in the same window, rather than in several windows; and finally support for the back button throughout the Web application. 3-3. Sufficient integration with supporting applications Web-based collaborative applications are often the storage and collaboration hub for a certain project. The content, however, is often created and discussed in dedicated applications. Those dedicated applications should have sufficient hooks into the teamware application in order to integrate cross-application tasks. For example, it should be possible to save directly to the Web-based collaborative application, or discussions around content of a document should be displayed within the document. Other aspects to think about are integration with a spreadsheet application for data analysis, and integration with messaging applications to display related discussions in context of the team information. 3.4. Use of "rich" UI mechanisms hi Web-based collaborative application There are some mechanisms from the traditional GUI model that users begin to expect from productivity applications, even if they are typically not used for Web applications. The main mechanisms to think about are context-sensitive commands (e.g., right-click), selection of list items, and direct manipulation (e.g., drag and drop). This is directly supported by our usability findings that indicate users' expectations for GUI-like interaction within WSS. For instance, in early prototypes, 80% of users attempted to find functionality for editing an item in the context-sensitive menu opened by a right-click. 3.5. Appropriate design for the task While adhering to the previous guidelines concerning consistency, the feature design can be unique for the task at hand. Frequency of use is an important metric in creating a successful design for the particular task.
3.5.1. Design for most frequent task Most likely, consuming and sharing information will be the most frequent task for Web-based collaborative applications, especially teamware. The UI of the teamware application should be optimized for this. One way of doing this is to purposely downplay other functionality, such as customization, in the main UI. Another aspect to think about is allowing for a method that permits users to skip repetitive navigation links. 3-5.2. Design for less frequent tasks Some tasks, such as customizing a Web site, are assumed to be performed only sporadically. Therefore, each attempt to complete one of these tasks will approximate first-time usage every time. The design should give a clear overview of all the offered options and tasks. Because efficiency is not of primary concern, it is appropriate to use a more elaborate and verbose UI. Once a user chooses a task, a new page specifically designed for this task should be provided. Verbose UI should guide the user through the steps and the title should clearly state what the task is about. There should be an obvious confirmation and cancel mechanism in place, which leads back to where the user chose the task in the first place, most likely the overview page of tasks. 3.6. Obvious separation between content and tools Attention to clear and consistent separation of application UI and contributed content is an important aspect for Web-based collaborative applications. Traditional GUI applications typically have a clear separation between the two, for example, the application's toolbars versus the active document. Because Web applications use the same UI mechanisms for both the UI and the content, this separation can easily become unclear. For example, hyperlinks are often used for buttons and for navigation to list items. Both are valid uses for hyperlinks—yet the behavior, after the user clicks it, can differ. By creating a unique design for different elements in the application, the user can develop a set of expectations for those different behaviors. The importance of the consistent and accurate use of these new elements is critical.
SUMMARY Web-based collaborative applications are replacing dedicated non-Web solutions because of the many relative benefits the Web can provide, such as ubiquitous access, familiar Web interface conventions, and perceived ease of deployment and maintenance. The implementation of these systems poses interesting design challenges for the solution providers who design the enabling platforms and the organizations who must customize their specific deployments. The most important challenge is
30. Web-Based Collaboration
incorporating years of knowledge about collaborative systems into the new solutions that Web technology can provide. Innovations in this technology will continue to emerge that offer new and promising solutions for addressing collaborative issues at the organizational, team, and end user levels. Our case study of Microsoft WSS technology highlights the importance of finding the right balance between traditional GUI
569
and emerging Web UI design principles. As productivity applications and Web-based collaborative environments become more tightly integrated, we will need to design more powerful, richer means of supporting the end-to-end collaborative experience. The design challenge for WSS and all Web-based collaborative applications is to understand and develop guidelines that address this important intersection of interaction styles.
References Bentley, R., Appelt, W., Busbach, U., Hinrichs, E., Kerr, D., Trevor, J., & Woetzel, G. (1997). Basic support for cooperative work on the World Wide Web. International Journal of Human-Computer Studies, 46, 827-846. Dix, A. (1997). Challenges for cooperative work on the Web: An analytical approach groupware and the World Wide Web. Computer Supported Cooperative Work, 6, 135-156. Dourish, P., & Bellotti, V (1992). Awareness and coordination in shared workspaces. Proceedings of the 1992 ACM Conference on Computer-Supported Cooperative Work, 107-114. Grudin, J. (1999). CSCW and groupware: Their history and trajectory. In Y. Matsushita (Ed.), Designing communication and collaboration support systems (pp. 1-15). London: Taylor & Francis. Gutwin, C., & Greenberg, S. (1999). The effects of workspace awareness support on the usability of real-time distributed groupware. In ACM Transactions on Computer-Human Interaction, 6(3), 243281. Guzzo, R., & Dickson, M. W. (1996). Teams in organizations: Recent research on performance and effectiveness. Annual Review of Psychology, 47, 307-338. Moran, T, & Anderson, R. (1990). The workaday world as a paradigm
for CSCW design. Proceedings of the Conference on ComputerSupported Cooperative Work, 381-393. Morris, S., Neilson, I., Charlton, C., & Little, J. (2001). Interactivity and collaboration on the WWW—Is the 'WWW shell' sufficient? Interacting with Computers, 13, 717-730. Olson, J. S., & Teasley, S. (1996). Groupware in the wild: Lessons learned from a year of virtual collocation. Proceedings of the ACM 1996 Conference on Computer-Supported Cooperative Work, 419427. Orlikowski, W. J. (1992), Learning from notes: Organizational issues in groupware implementation. In J. Turner, & R. Kraut (eds.), Proceedings of CSCW'92 Conference (pp. 362–369). New York: The Association for Computing Machinery. Posner, I., & Baecker, R. (1993). How people write together. In R. Baecker (ed.), Groupware and computer-supported cooperative work (pp. 239—250). San Mateo, CA: Kaufmann. Sproull, L., & Kiesler, S. (1991). Connections: New ways of working in the networked organization. Cambridge, MA: MIT Press. Wheeler, B., Dennis, A., & Press, L. (1999). Groupware comes to the Internet: Charting a new world. The DATA BASE for Advances in Information Systems, 30(3-4), 8-21.
This page intentionally left blank
Part
X WEB PSYCHOLOGY AND INFORMATION SECURITY
571
This page intentionally left blank
31 KEY PSYCHOLOGICAL FACTORS THAT AFFECT UPTAKE AND USE OF WEB-BASED SERVICES Ron Henderson ClientWise Pty Ltd.
Desley Hennessy Megan Divett University of Wollongong
This chapter discusses the key psychological factors associated with the uptake and use of Web-based services. Our central thesis is that to be successful, a Web-based service must attract and then retain an economic user base. Our concern is that in the past the central focus of many developers has been on the technological aspects of the service rather than the client user and that such a focus has led to low use and high churn rates. This chapter presents and discusses core psychological aspects related to volitional behavior and how these relate to the decision to use a Web-based service. Empirical research is then examined. The chapter concludes by discussing how to apply this knowledge to the Web service environment. In particular, it is argued that service failure risk may be reduced by systematically applying this knowledge to tune the service to meet the needs of the client groups.
for Web-based services to deliver real returns to sponsors and developers alike. In addition, there is much ambiguity as to the immediate and longer-term direction and evolution of online services. Even obtaining consensus regarding the definition of Web services is difficult (Landgrave, 2003). Definitions of Web services vary from online services, such as purchasing of goods and services, to online programs, such as software available for download on the Internet. Intelligent Enterprise magazine (2002) claimed that Web services are the next wave in business process automation, with Web services being defined as a "business function run from a service provider's Web server using a request and response process between two applications managed via the service application programming interface" (p. 26). In this case, the functionality could be a single piece of data, a document, or an entire file. Howerton (2002) described a Web service as nothing more than a piece of functionality that has been deployed on a Web server and contains an externally visible interface, making it eligible for remote command. Kalita (2002) provided a more succinct definition by simply pointing out that Web services can include anything in this world that can be accessed online, anywhere and at anytime. In this chapter, we propose a broad definition of services. We define Web services as those services available to any consumer over the Internet. These include retail purchases, such as online grocery shopping, purchase of books, electronic equipment, or pharmaceuticals; financial services, such as checking the balance
Web-Based services. Web-based services are now offered in almost every conceivable theatre of life. Web services range from business, educational, and health care concerns to leisure activities for individuals or groups. The uptake of Web-based services has proceeded at a phenomenal rate. Simply conducting a Web-based search on the term "Web services" highlights the pervasiveness of the Web, where in early 2003 approximately 8 million references were returned. Despite the popularity of Web-based services, it must be acknowledged that development of the Web is still in its infancy. There is much uncertainty and discussion about the potential
573
574
HENDERSON, HENNESSY, DIVETT
of bank accounts and paying bills; requests for information, such as phone numbers, flight information, and government- or community-provided information; and search engines, such as Google and Yahoo! The central thesis to this chapter is that a Web-based service must ultimately provide some form of return on investment to the sponsor and developer, and that the only way to do this is by capturing and maintaining an economical client base. To this end, this chapter first provides an overview of current issues in Web service development, concluding with a focus on the need to attract and keep customers. The chapter then reviews and conceptually integrates theory relevant to the evaluation of the development and uptake of Web services. This integration is followed by a selective review of empirical research findings related to Web service utilization. Finally, the chapter concludes with examples of current client-focused applications with suggestions for client-centered design. Web-Based service uptake and use issues. Kocharekar (2001) reported that currently organizations emphasize the "information" rather than the "technology" component of information technology (IT), and this reflects the evolution of their underlying business models toward a greater customer focus (Kocharekar, 2001). This change in focus highlights the need to rethink the IT customer support model, concentrating on the user rather than the product. In this vein, Kocharekar (2001) asserted that IT's new mission is "the right information to the right people at the right place at the right time" (4Rs). Unfortunately, there appears to be a vast gap between what is purported to be the new IT mantra and the actuality of the business world. Specifically, we believe that the focus has been, and largely still is, on the technology rather than the end user/client. That is, the focus seems to be on displaying the technical power of Web-enabled services rather than solving customer problems. This approach has been displayed on a number of occasions to us, where developers have demonstrated prototypes of business Web services that, if implemented, would increase the user task workload without adding any real value to the business process. Needless to say, these days, such ventures seldom get off the ground. Just because a service is technically feasible does not necessarily make it a good idea—what seems like a good idea may in fact be the next White Elephant. When considering the provision of a Web service, it is paramount that organizations determine what is needed to attract and then subsequently keep customers/users. There must be something more valuable to the customer than the novelty value of interacting with an organization via the Internet, particularly, in today's Web-saturated world. At the end of the day, without a customer group the service benefits no one. The key concept to grasp is that, in many cases, the Web service is simply an extra service channel. To compound issues further, the channel is a self-help service channel that relies on a relatively lean form of communication. The question that needs to be repeatedly asked is "what would make a person access this channel rather than that of competing service channels (shop front, telephone, etc.)?" If the proposers and developers of the
system cannot clearly articulate the answer to this question, and see clear advantages to the Web service, then the service may simply become redundant. There is little doubt in our mind that Web services can be developed into an important adjunct to traditional service delivery and, in some instances, may become the prominent mode of service delivery. The sheer number of users on the Internet and its growth rate are staggering. Moston (1997) reported that there were more than 32 million registered Internet users in North America alone in 1997. Fellenstein and Wood (1999) reported that, by 1999, there were more than 80 million users in North America representing a 340% growth during the 1997 to 1999 period. By March 2000, Nielsen//NetRatings reported that this figure had increased to an Internet universe of 129.7 million in the United States and 221 million by March 2003. Moreover, Internet use has not been confined to North America alone, with Nielsen//NetRatings reporting an Internet universe of 6 million in Australia in 2000 (and 12.7 million in 2003), 693,000 in Ireland, 1.2 million in New Zealand, 1.7 million in Singapore, and 17 million in the United Kingdom in 2000 (expanding to 34.4 million in 2003). These figures provide a global estimate of 399-5 million for January 2003. Moston (1997) reported that between 15% and 19% of online users make purchases via electronic means, a figure that appeared to be remaining stable. Given the rapid growth in the numbers of Internet users, the financial impact becomes obvious. For example, these figures equate to an approximation of between 23-4 and 29.7 million purchasing users in these six countries in the year 2000 alone. Furthermore, Moston reported that up to 40% of online users use information gained via electronic means to support purchasing decisions, highlighting the importance of the information dissemination role of Web-based services. In addition to these commercial transactions, there is also the potentially larger use of purely information-only sites traditionally belonging to government and community groups, and the impact of business-to-business interactions. However, again, we must emphasis that the continued growth and development of Web-based services will largely depend on the ability of services to meet client needs. As well as competing with other physical providers, the online service provider often needs to supply a better product/service than other online providers to remain competitive. Despite the apparent popularity of Web services, there is still much disagreement about their current and potential impact. Although the Gartner Group group predicted that Web services would have a mainstream impact by 2004, others have predicted that it will take another 10 years before the potential of Web services is fully realized (Kalita, 2002). At the same time, it has been suggested by still others that the concept of a Web service will die out in the near future (Kalita, 2002). Online Web services with mass consumer access are definitely "here and now," contrary to some predictions. Available services extend beyond buying groceries and books, and making bill payments. A digital version of The Australian newspaper was launched on August 8, 2001. As Mark Day, The Australian's Media columnist put it: "Suddenly, the Holy Grail. This is what
31. Uptake and Use of Web-Based Services we've dreamed of for decades—an electronic form of distribution of newspapers" (Weaver, 2002). There are also theoretically many reasons to contemplate providing a Web-based service as part of the service channel mix. Such reasons include attempting to gain a competitive advantage by offering a Web-based service when competitors do not, the ability to overcome geographic or organizational size barriers, the ability to provide services to remote clients, the ability to provide services during nontraditional work hours, cost savings due to enhanced supply chain management, lower costs due to the supplementation of face-to-face interaction by providing electronic media, attempting to gain access into a new rapidly expanding market, or simply the excitement of a new business venture. Whatever the reasons for contemplating the provision of online services, the venture must be profitable in the medium to longer term by achieving organizationally relevant, strategic objectives. Although new Web services appear on a daily basis, it is unclear what, if any, theory and research drive the planning and development of these services. Presently, and for the foreseeable future, the use of such services will be largely volitional in nature. That is, people are free to decide whether to access the service. For this reason, psychological theories of volitional behavior are examined in the following section. This leads into a discussion of empirical research relevant to the application of these theories to Web service design. Practical implications and guidelines for designing Web services are then detailed. Finally, the chapter concludes with suggestions for consideration when developing and managing client-centered Web services.
THEORIES OF VOLITIONAL BEHAVIOR APPLIED TO THE UPTAKE AND USE OF WEB SERVICES Discussed briefly are the three generic theories of human behavior that we believe should act as the conceptual basis for the development of client-oriented, Web-based services. These theories include the Theory of Reasoned Action (TRA; Fishbein & Ajzen, 1975), a generic theory of volitional behavior; the BentlerSpeckart (Bentler & Speckart, 1979, 1981) modification to the TRA; and the Theory of Planned Behavior (TPB; Ajzen, 1991), an augmentation of the TRA. Along with these generic theories of behavior, two specific theories are also discussed. The Technology Acceptance Model (TAM) is a technology uptakespecific exemplar of the modified TRA (Davis, 1986), and the Expectation-Disconfirmation Model (Oliver, 1977), which is a consumer-oriented exemplar of the TRA. It should be noted that the versions we are advancing are conceptual simplifications of these theories, as a chapter of this length cannot do justice to each theory on its own, much less to all these models. Consequently, we have presented each in, what we believe, is its core nature as relevant within the current context. Readers are encouraged to examine more detailed reading on each topic and a suggested list is provided as an Appendix to this chapter.
575
FIGURE 31.1. Core attitude-intention-behavior link.
Core Attitude-Behavior Model Common to all the mentioned models is the concept that the precursor to behavior is an individual's intention to undertake the behavior, and that one of the precursors to the formation of the intention is an individual's attitude toward conducting the behavior. Simply put, the behavior is the result of intentions to conduct the behavior, which was a result of attitude toward performing the behavior. In the current context, the model would say that an individual used a particular Web site because they intended to access the site and that the intention was a result of the attitude formed toward the site. This basic sequence is illustrated in Fig. 31.1. The model at this stage is, of course, a simplification. Sometimes behavior is performed without the formation of an intention, such as in the case of habitual behavior. Other an individual may have an intention, but then not perform the behavior. For example, an individual may intend to log on to a site to obtain certain information, but cannot access the system. To improve the predictive power of the model, we need to present a more complex model of human volitional behavior. This is where the models start to differ. However, we consider the TRA as the base model because it is the core theory within all the models under discussion. The Theory of Reasoned Action. The TRA (Fishbein & Ajzen, 1975) expands upon our core model of volitional human behavior by including one further construct to account for the individual's perception of social influence, sometimes known as subjective norms. Specifically, as mentioned previously, the TRA views an individual's intention to perform a volitional behavior as the immediate determinant of subsequent overt action. That is, we form intentions to undertake a behavior. For example, the intention to shop for music CDs leads to the act of shopping for a compact disc. According to the TRA, an individual's behavioral intention is a function of two basic factors, one personal in nature (attitude) and the other reflecting social influence (normative impressions). Because most behaviors of social relevance are under volitional control, behavioral intention is viewed as both the immediate determinant and the single best predictor of behavior. The model also assumes that variables external to the model influence intentions only to the extent that they affect either attitudes or subjective norms. This extension to the core model is presented diagrammatically in Fig. 31.2. Within the TRA model, attitude is defined as a person's positive or negative evaluation about performing a behavior (Fishbein & Ajzen, 1975). Subjective norms refer to a person's beliefs that specific individuals or groups approve or disapprove of performing the behavior. Each factor will have a different weight
576
HENDERSON, HENNESSY, DIVETT
FIGURE 31.2. Simplified model incorporating the Theory of Reasoned Action elements (Fisbein & Ajzen, 1975). for each individual. It should also be noted that Attitudes and Subjective Norms are related, hence an arrow between Attitude and Subjective Norm in Fig. 31.2. The theoretical model based on the TRA has been -widely used to predict behavioral intentions and behavior across a variety of subject areas. At the same time, the theory has stimulated theoretical research aimed at understanding the model's limitations, testing key assumptions, and developing refinements and extensions to the theory/model. These assumptions and limitations must also be taken into consideration. Generally, if the model is applied as its authors intended it, the model is useful and provides a good conceptual starting point in any discourse of IT uptake. However, the authors articulate three conditions for the applicability of the model, where violation of these boundary conditions reduces the magnitude of the relationship between intention and behavior. These conditions are (a) the degree to which the measure of intention corresponds in its level of specificity to the behavioral criterion (b) the stability of intentions between the time the intention is assessed and the time at which the behavior is performed, and (c) the degree to which carrying out the intention is under the person's volitional control. For example, you cannot ask people what they think about online services in general and expect the reply to relate strongly to a particular service. If you are interested in a particular service, then you should ask people about that specific service. In addition, it is not useful to ask people if they would use a service now and then expect that opinion to relate to their actual use in a year's time because circumstances may change over time. The relationship between intention and actual behavior decreases over time. Furthermore, if performing the behavior is not fully under the individual's volitional control the intention-behavior relationship may be weaker. Unfortunately, in the current authors' opinions, these boundary conditions are often violated, resulting in research results that are less than impressive or erroneous in nature. The third boundary condition also severely limits the application of the model, as even very mundane activities that can usually be undertaken at will are sometimes influenced by factors beyond an individual's control. This weakness in the model has been accommodated in a revision, the TPB, discussed later in this section. In the current context, the TRA model would imply that an individual's intention to engage in using an online service would depend on how favorably they personally view using the service (attitude), and what the individual thinks the impressions of significant others are about them using the service (subjective norms). Thus, an executive may form an intention to use an online grocery service based on how convenient the service is to
her and what she thinks her work colleagues think of her using the service. Conversely, a homemaker may form his intention based on the monetary savings and what he thinks his parents or partner think of him using the online service. A youth may, in turn, base his intention to use an online music service based on price and what he thinks his school colleagues think of him using the service. It is important to note that attitudes and subjective norms may be different for each service and customer group. In addition, the relative importance of these two factors will be different for each service and customer group. When using the conceptual model based on the TRA, the online service provider needs clear knowledge of what the important attitudes toward the service are, as well as the identity of the significant referents.
The Bentler-Speckart Model Using the TRA (Fishbein & Ajzen, 1975) as a starting point, Bentler and Speckart (1979) developed a model of attitudebehavior relations that had two major modifications. First, they argued that attitude could affect behavior directly and that it did not always affect behavior through intentions. Second, they proposed that past behavior might directly affect both attitude and intentions. A topic previously researched by Regan and Fazio (1977), who found that the more an attitude was based on direct behavioral experience, the more likely it would be that there was greater attitude-behavior consistency, reporting that direct behavioral experience produced an attitude that is more clearly, confidently, and stably maintained than an attitude formed by more indirect means. Specifically, Bentler and Speckart (1979) argued that behavioral intention is cognitive in nature, and that consequently there must be some mechanism by which emotion or attitude can influence behavior without the regulation of cognition. It was further proposed that behavior might similarly influence subsequent behavior directly, in addition to indirectly through intention. In other words, prior behavior impacts attitude, current intentions, and future behavior. In the context of Web services, prior utilization of the Internet for shopping would influence a consumer's attitude and intention to repeat the Internet shopping experience (see Volk & Kraft, chap. 32, this volume). Unlike attitude, Bentler and Speckart (1979) proposed that subjective norms did not have a direct effect on behavior. Finally, intention was included in the model, and was found to impact the relationship between attitudes and behavior measured at different points in time. Relating this to the uptake and use of Web services, a consumer's initial attitude toward purchasing goods via the Internet will influence and be influenced by subjective norms, current behavior (current mode of shopping), and current intentions (using the Internet to shop). This initial attitude will influence the consumer's future attitudes toward using the Internet to shop, as well as whether they continue to use the Internet to shop in the future. It was argued by Bentler and Speckart that the dominant influence of attitudes over behavior might be partly due to the high correlation between intentions and attitudes. Both studies by Bentler and Speckart (1979, 1981) found that the effects of
31. Uptake and Use of Web-Based Services
577
FIGURE 31.3. Simplified model incorporating elements of the Bentler-Speckart modification of the TRA.
intention are context specific. Supporting the first boundary condition for the TRA outlined by Fishbein and Ajzen (1975). In addition, future behavior was found to be significantly affected by attitude and prior behavior. Rentier and Speckart concluded that previous behavior, attitudes, and intentions are necessary elements of attitude-behavior relational models. As mentioned previously, research has demonstrated a direct relationship between prior behavior and attitude. These studies have led to such models as the Attitude-Accessibility model articulated by Fazio (1986, 1989). Readers interested in this relationship should also refer to the research conducted by Triandis (1977) examining the role of habitual behavior, a notion that has received some empirical support (e.g., Landis, Triandis, & Adamopoulos, 1978; Pare & Elam, 1995). Our model now becomes slightly more comprehensive (see Fig. 31.3). The model now stipulates that behavior will be influenced by intention, attitude, and prior behavior. In turn, intention is influenced by attitude, subjective norms, and prior behavior, and attitude is influenced by past experience and behavior. As mentioned previously, neither the TRA nor BentlerSpeckart models apply when behavior is not entirely under volitional control, that is, when there are no choices. In the current context, the social and cognitive elements (intentions, attitudes, and subjective norms) will have no relationship to behavior when the individual simply has to use the service, or conversely, does not have the means to use the service. Obviously, volitional control is a continuum running from total control to no control, and the model's applicability will depend on the amount of volitional control the individual has with regard to accessing the service. This is an obvious limitation to the models that can be addressed using the later reformation of the TRA, the Theory of Planned Behavior. The Theory of Planned Behavior, in an attempt to extend the TRA to encompass behaviors that are not entirely under volitional control, the Theory of Planned Behavior (TPB) was developed (Ajzen, 1991). The TPB specifies the two original TRA determinants of intention—attitude and subjective norm— as well as a third determinant of intention, namely, perceived behavioral control. Perceived behavioral control refers to an individual's belief about how easy or difficult it will be to perform
FIGURE 31.4. Simplified model incorporating the Theory of Planned Behavior elements. a particular behavior, and can conceptually be broken into two broad beliefs: internal personal beliefs and external beliefs. An example of an internal belief in the current context is the belief that an individual has the requisite skills to use a computer. An external belief may be that the person believes that there is no access to a Web-enabled computer available. In the TPB, perceived behavioral control operates to affect behavior directly as well as indirectly, through intention. Our conceptual model can now be presented (Fig. 31.4). Specifically, this model articulates that behavior will be influenced by intention, attitude, perceived behavioral control, and prior behavior. In turn, intention is influenced by attitude, subjective norms, perceived behavioral control, and prior behavior. The psychological theories here discussed are generic in nature and are intended to be applied to the prediction of human behavior in a range of settings. It will come as no surprise to learn that these theories have been modified to examine human behavior in specific applications. Two models are of particular relevance to the current context. The first is the Technology Acceptance Model (TAM), which has been specifically developed to examine the uptake and use of IT systems. The second is the Expectation-Disconfirmation Theory (Oliver, 1977), which essentially integrates the generic attitude-behavior model within a consumer purchase process, while expanding the model to explore the impact of expectation-disconfirmation.
The Technology Acceptance Model Although the TRA and the TPB are useful generic models of behavior, they are not IT specific. Davis (1986) developed the TAM to specifically model user acceptance of end user information systems. Davis recognized the merit of using a proven behavioral intention model to predict computer usage and acceptance, and consequently, based the TAM on the principles expressed in the TRA (the TAM was developed prior to the articulation of the TPB).
578
•
HENDERSON, HENNESSY, DIVETT
The TAM was intended to be used prior to system implementation to obtain feedback on system design features, as well as after implementation to diagnose problems in user acceptance. Therefore, the model aims to explain user acceptance while promoting understanding of how to improve user acceptance through system design. According to the TAM, design features of a computer system are external stimuli about which a user forms certain cognitive responses. These cognitive responses are causally linked to a user's attitude (affective response) toward use of the computer system. Their intention to use a computer system is jointly determined by the user's attitude toward using the system and the perceived usefulness of the system, with the behavioral response being the actual use of the computer system. The TAM, however, focuses on the internal psychological variables that influence IT acceptance. Based on the TRA, it is postulated that all other external variables influence behavioral intention and ultimately behavior, indirectly, via their effect on attitude. This assertion is particularly helpful from an information systems perspective, as variables such as system design characteristics, nature of development or implementation processes, specific user characteristics, and organizational and political structure all fall into the category of "external variables." This position makes the TAM much more "useable" by collapsing the number of variables to a manageable size, and consequently allows the model to be more readily empirically tested. The TAM originally proposed that two major beliefs, perceived usefulness and perceived ease of use, were fundamental determinants of system usage. Perceived usefulness was originally defined as the degree to which a person believes that using a particular system would enhance his or her job performance (Davis, 1993). Perceived ease of use was defined as the degree to which a person believes that using a particular system would be free of effort (Davis, 1993). Although the perceived ease of use definition is directly transferable to the online situation, some minor modification is required to the perceived usefulness aspect of the model before it can be applied to an online context. In the online context, perceived usefulness can be seen as the degree to which a person believes that using a particular system would help complete the task at hand, whether it be shopping for a CD, paying speeding fines, or gathering information on local community events. When attempting to implement Web services, organizations need to consider both perceived usefulness of the system to the consumer and the perceived ease of using the Web-based service. If the consumer perceives using the Internet is a useful way to purchase good, this view will influence whether the consumer purchases goods and services over the Internet. Similarly, whether the system is easy to use will have an effect on customer usage. In an attempt to accommodate the observation that some activities are interesting to undertake for their own sake, and hence are fun to perform, Davis (1986) later included the notion of perceived fun of using a system or technology. Perceived ease of use, perceived usefulness, and perceived fun are postulated to represent the salient belief set for system usage. This set of beliefs can be applied to different computer systems and user populations.
The TAM has served as the basis of much research examining the uptake and use of technology systems, and as such, has had a number of additions and reformations of varying quality and conceptual clarity. Some additions openly overlap with the generic theories articulated previously. Others contradict the explicit boundary conditions of the parent TRA. The articulation of service aspects for consideration such as perceived fun, computer playfulness, control, and previous behavior were the result of these modifications. There have been a wide range of other individual personal variables, such as computer anxiety and self-efficacy, that have also been considered within the TAM framework, although detailed coverage of their relative influence and utility is beyond the scope of this chapter.
The Expectation-Disconfirmation Theory. The Expectation-Disconfirmation theory of consumer behavior (Oliver, 1977) is a cognitive model of consumer behavior that has its roots in the TRA and Helson's (1964) Adaptation Level Theory. The model examines how the appraisal (thought) process about the service interaction (how well it met expectations) impacts the perceptions (quality) and emotions (satisfaction) of the consumer. In terms of the TRA, quality and satisfaction are viewed as the core attitudes held by the consumer, which then directly impact the intentions to repurchase or use the service again. In simplified form, the model asserts that when an individual is asked to evaluate a service or product that they have experienced, they will go through a set cognitive process. Specifically, the model holds that the prediction (expectations) customers make in advance about the service or product act as a standard against which customers measure performance. Expectations are believed to create a frame of reference when the comparative judgment is made. The model argues that confirmation of expectations result when the actual level of performance is as expected. If performance is better than expected, positive disconfirmation results and, in turn, leads to enhanced perceptions of quality and satisfaction with the experience. Conversely, if performance is worse than expected, the outcome is negative disconfirmation and leads to poor perceptions of quality and satisfaction with the experience. Researchers have suggested that if the performance of the evaluated service or product is within a reasonable range, no disconfirmation will occur. They refer to this range as the "zone of indifference," and it has been found that experiences outside this zone attract attention and the resulting disconfirmation is magnified. For this reason, the service supplier requires knowledge of the service recipients' expectations, and the ability to control or manipulate the expectations of this group (predominantly through communication). Although expectations as predictions is a dominant paradigm, the satisfaction literature suggests alternative standards consumers may employ when forming opinions about anticipated performance. Such alternative standards may include ideal expectations, desired expectations, normative expectations, experience-based expectations, minimum tolerable expectations, comparative expectations, and equity-based expectations. When considering the notion of expectations, previous research has also indicated there may be two generic classes of dimensions associated with consumer stimuli: focal and contextual.
31. Uptake and Use of Web-Based Services
579
FIGURE 31.5. Expanded model of theoretical elements predictive of Web uptake and use. Focal stimuli are those that the consumer directly responds to, for example, the service or product. Conversely, contextual stimuli are those that are the background for the perception of focal stimuli, such as interface and the delivery process. Consequently, services should be seen as an interaction among the purchaser, service personnel, and the environment in which services are produced and used. Moreover, focal and contextual stimuli may influence perceived quality and satisfaction in different ways. For instance, consumers may experience a positive disconfirmation of perceived expectations on one dimension, yet simultaneously experience a negative disconfirmation of expectations on another dimension.
Summary and Enhancements of the Models of Human Behavior So how do these two specific theories relate to the generic theories of behavior? Essentially, the TAM (and its variants) and Expectation-Disconfirmation models explore the attitude and precursor-to-attitudes side of the model. As a result, the core model outlined in Fig. 31.5 becomes more detailed in the sorts of variables that are believed to impact the uptake and use of a Web service. For example, rather than thinking of attitudes in general, we can put some names to the key attitudes (or determinants of attitudes) in the form of perceived usefulness perceived, ease of use, perceived fun, satisfaction, and quality. As a result, the core model is likely to be a sufficient theoretical starting basis for predicting Web service uptake in most contexts. Another question that arises is "how good are these models at predicting use of human behavior in general, and Web service use in particular?" Unfortunately, some of the most wellarticulated and internally consistent theories have been found wanting when placed under empirical scrutiny. As a result, the topic of the next section is an examination of the empirical evidence pertaining to the relationships espoused in the expanded model. After an empirical evaluation, the road is then open to developing explicit guidelines and suggestions to the development of a client-based Web service. Specifically, we present a checklist that may be useful when conceptualizing and evaluating Web services.
EMPIRICAL STUDIES OF THEORETICAL LINKS UNDER CONSIDERATION This section provides a brief overview of the research that has explored the relationships discussed in the previous section. The aim is to establish empirical support, or otherwise, for these theoretical relations in an attempt to identify the applicability of these models to the real world. Specifically, each relationship detailed in Fig. 31-5 is discussed in light of the empirical research that has been conducted. Other variables that have been taken into account are briefly mentioned, where relevant. The relevance of these findings for online service uptake is also detailed. It should also be pointed out that the results are reported in simplified form and no account of multivariate relations is discussed. Multivariate relations are generally not considered, as we desire the reader to focus on each hypothesized link as presented in Fig. 31-5-
Relationship Between Attitudinal Elements, Intention, and Behavior Our model of Web service use (Fig. 31.5) argues that the immediate predecessor to behavior is the formation of an intention to perform that behavior. This is derived from, among other things, an attitude toward the service. First, we need to consider the notion that an immediate precursor to enactment of behavior is the formation of an intention. General research has supported the notion that intentions predict behavior. For example, Fishbein and Ajzen (1975) reported an average intention-behavior correlation of between .85 and .66. Ajzen and Fishbein (1980) reported similar correlations of between .82 and .73. In a meta-analysis of 150 studies, van den Putte (1993) found an average intention-behavior correlation of .62 and, in a further meta-analysis of 32 studies, Hausenblas, Carron, and Mack (1997) reported an average intention-behavior correlation of .47. Within the Web domain, Roberts Roberts (1999) explored the intention-behavior link within a sample of 143 users of an
580
HENDERSON, HENNESSY, DIVETT
TABLE 31.1. Summary of Empirical Findings for the Relationship Between Intention and Behavior Intention Intention to use an online supermarket
Intention to use an online supermarket
Intention to use an online supermarket Intention to use bulletin board
Behavior
Correlation
Log-on activity
.34
Purchase value Deliveries per month Current behavior
.36 .37
Log-on activity Purchase value Deliveries per month Future behavior Log-on activity Purchase value Deliveries per month Subsequent use
.53 .60 .63
Self-reported use
electronic supermarket. In this study, users were requested to report their intentions to use the supermarket. Behavior was measured in terms of log-on activity, deliveries, and purchases value per month. Roberts reported intention-behavior correlations of .34 (log-on activity), .37 (deliveries), and .36 (purchase value) within the month the survey was conducted. Roberts reported concern with these low relationships, but could not identify whether the low correlations were due to measurement errors, or the fact that rather than relying on self-reported behavior, as in many other studies, electronically recorded measures of behavior (log-on/deliveries/purchase value) were used. In an attempt to clarify the concern articulated by Roberts (1999), Anschau (1999) conducted follow-up research within the same setting, but using a refined measure of intentions. Using a separate sample of 113 electronic supermarket users and the same behavioral use measures (log-on/deliveries/purchase value), Anschau reported intention correlations of .53 (log-on), .63 (deliveries), and .60 (purchase value) for the month the survey was conducted, and .49 (log-on), .58 (deliveries), and .57 (purchase value) within the following month. This research went some way in highlighting that measurement might have been the reason for the low correlations observed by Roberts. Similarly, Mathieson, Peacock, and Chin (2001) reported and intention-behavior relationship of .47 when exploring technology use. Henderson (1999) examined the ability of intentions to predict subsequent uptake of an electronic supermarket. In this research, a Weighted Application Blank was derived from intention items and then used to predict subsequent shopping use over a 6-month period. The weighted blank used statistical analysis to identify those items that best predicted whether a person would shop or not. It was found that the derived function correctly predicted shopping behavior (shop/nonshop) with a 77% accuracy rate. As well as the general empirical support associated with the TRA and TPB (e.g., Fishbein & Azjen, 1975), the relationship
Researchers Roberts (1998)
Anshau (1999)
.49 .57 .58 77% accuracy .47
Henderson (1999) Mathieson, Peacock, & Chin (2001)
between intention and behavior has received consistent empirical support within the online environment. Indeed, intention to use the Web service leads to actual use of the service. The function was also found to adequately predict frequent, intermittent, and nonshoppers. Table 31.1 presents a summary of these findings. The TAM articulates a number of core attitudinal elements, specifically, perceived usefulness, ease of use, and fun/playfulness. Each element has been explored within the online environment and received empirical support. For example, in applying the TAM to the use of e-mail and a text editor, Davis (1993) found that perceived usefulness had a strong effect on attitude to use these technologies. Ease of use had a smaller effect on attitude, yet a strong effect on usefulness. Attitude toward the technologies had an effect on usage, but perceived usefulness had a strong direct effect on usage over and above attitude. In a survey of 62 North American companies, perceived fun was found to be significantly correlated with both computer satisfaction and usage (Igbaria, Schiffman, & Wieckowski, 1994). The results supported the predictions that perceived fun would have a relationship with both elements of user acceptance. Computer satisfaction was correlated with usage, again as predicted. Henderson and Divett (2003) explored the relationship between usefulness and ease of use and purchases (value/ deliveries/log-on) of a group of supermarket users, and found that although both factors held a significant bivariate correlation with each behavioral indicator that ease of use did not contribute unique predictive power when usefulness was considered within the equation. Hu, Chau, Liu Sheng, & Yan Tarn (1999), who when exploring intention to use telemedicine technology (technology that utilizes live interactive video for medical purposes) found that, although both usefulness and ease of use related to attitude and attitude related to intention, ease of use was nonsignificant in the presence of usefulness. In a study by Shim, Eastlick, Lotz, and Warrington (2001), intention to use the Internet for purchases, as well as intention
31. Uptake and Use of Web-Based Services to use the Internet for search, were directly affected by attitude toward Internet shopping (.35, .28, respectively). Anschau (1999) conducted an empirical test of the TAM model with a sample of supermarket shoppers. In that study, the TAM was largely supported. As predicted, intentions related to all behavioral indicators (log-on/deliveries/purchase value), attitudes related to intentions and perceived ease of use, and usefulness and fun related to attitudes. However, when the full model was tested, perceived ease of use had its impact on attitudes indirectly through perceived usefulness and perceived fun. A model based on the TAM was tested by Chen, Gillenson, and Sherrell (2002) using data from a survey of 253 online consumers. These researchers attempted to explore the precursors to attitude in an online environment, including relative advantage, compatibility, complexity, trialability, and observability. It was concluded that TAM is a valid model to explain and predict user behavior in the business-to-consumer context. Specifically, these researchers demonstrated that attitude determines intention, which in turn predicts consumer acceptance and use of virtual stores. The primary determinants of consumer attitude consisted of compatibility, perceived usefulness, and perceived ease of use. Compatibility and perceived ease of use were found to affect perceived usefulness of virtual stores. The implication is that enhancing a virtual store's compatibility with consumers' values, needs, and lifestyle, as well as the perceived usefulness and perceived ease of use of the store, will result in a more positive consumer attitude. It should be noted that some empirical studies draw a distinction between perceived ease of use, perceived usefulness, and attitude, while others treat the constructs as equivalent. Attitude is seen to represent ease of use and usefulness, as well as many other factors such as fun/enjoyment, motivation, trust, and risk. This should be remembered when considering research in this section. For example, Henderson and Divett (2003) explored the relationship between system perceived usefulness and perceived ease of use and behavior (logon/deliveries/purchase value) with an electronic supermarket context. Other researchers, however, such as Hu et al. (1999) recorded usefulness, ease of use, attitude, and intention, and Mathieson, Peacock, and Chin recorded usefulness, ease of use, attitude, and behavior, as did Anschau (1999). In part, these differences may represent the practicalities of conducting applied research, particularly when some are commercially sponsored, whereas others represent the particular theoretical and research background of the researchers. In a comparison of hedonic (pleasurable) and utilitarian (useful) online shopping, the TAM, with the inclusion of enjoyment, was tested by Childers, Carr, Peck, and Carson (2001). All three variables—usefulness, ease of use, and enjoyment—were found to have a direct effect on attitude for both types of shopping. However, the strength of the predictors differed depending on the setting. In the utilitarian context (e.g., grocery shopping), usefulness was a stronger predictor of intention, while ease of use and enjoyment were found to be equally predictive. In the hedonic environment (e.g., browsing a book store), enjoyment was the strongest predictor, and ease of use and usefulness predicted shopping attitudes equally.
581
Koufaris (2002) also examined the value of hedonic and utilitarian aspects within an online environment. Using a questionnaire completed by Web users, Koufaris showed that, for new customers, perceived usefulness (an information systems variable) and enjoyment (a psychological variable) were important in predicting intention to return with perceived usefulness explaining 49% of the variance of intention to return. Although recommender systems and express checkouts increase customer convenience, shopping enjoyment was shown to be beneficial in retaining customers, indicating a need for both hedonic and utilitarian value in online stores. Motivation to perform a behavior can be divided into two main classes: extrinsic and intrinsic motivation. The drive to perform a behavior in order to achieve specific goals/rewards is known as extrinsic motivation, while intrinsic motivation relates to the satisfaction gained from performing the behavior. Within the TAM, extrinsic motivation is captured by perceived usefulness. Are the technologies useful to perform a specific function? The TAM does not explicitly include intrinsic motivation; however, Venkatesh (2000) proposed that the role of intrinsic motivation relates to the perceived ease of use construct. If the technologies are easy to use, the consumers will be satisfied with the process. Enjoyment and perceived fun are also closely related to the construct of intrinsic motivation. In an investigation of a sample of 64 consumers of an electronic supermarket, Henderson, Rickwood, and Roberts (1998) found significant relationships between perceived fun and intention to use the service, as well as between perceived ease of use and perceived usefulness. Perceived usefulness and perceived ease of use were directly related to behavioral intention, as well as with each other. The addition of intrinsic and extrinsic motivation to the TAM was also investigated by Venkatesh, Speier, and Morris (2002). Intrinsic motivation was found to be a catalyst for both perceived usefulness and perceived ease of use. Because perceived usefulness and perceived ease of use are both important predictors of usage intention, this implies that intrinsic motivation plays an indirect role in understanding short-term acceptance or rejection of new technology. An interesting finding was that perceived usefulness was not influenced by training interventions, and continued to exert a strong positive effect on intention to use technology. If the system is useful in completing a task, people will intend to use it. A finding that appears to not be influenced by training. Chen et al. (2002) recommended enhancing the playfulness of using a virtual store as a high priority, as playfulness enhanced motivation to use a system. In addition, Igbaria, Schiffman, and Wieckowski (1994) reported that perceived fun had a stronger effect on satisfaction than on perceived usefulness. Likewise, Venkatesh (2000) demonstrated a correlation between perceived usefulness and system usage that was higher than that between perceived fun and system usage. Perceived fun was found to be the only significant predictor of computer satisfaction, whereas perceived usefulness was the only significant predictor of system usage. In a related study, perceived ease of use and perceived usefulness explained approximately 35% of the variance in intention to use new technology, lending strong support to the TAM (Venkatesh & Morris, 2000). Perceived ease
582
HENDERSON, HENNESSY, DIVETT
of use was found to have a direct effect on intention and an indirect effect (via perceived usefulness) on intention to use technology. Trust and risk were incorporated into the TAM by Pavlou (2001) in an effort to predict consumer adoption of e-commerce. Pavlou argued that trust is possibly the most important component of consumer-marketer transactions. Trust increases the amount of control consumers believe they have over the situation through having confidence in the Web retailer's behavioral actions. Incorporating risk and trust into the TAM explained 64% of the variance in intention to transact (Pavlou, 2001). It was concluded that trust affects intention to transact through attitude and perceived behavioral control, whereas risk influences intentions through control alone. Similarly, Jarvenpaa, Tractinsky, and Vitale (1999) concluded that trust may operate to qualify purchase decisions for Internet consumers. Jarvenpaa et al. (1999) argued that lack of trust in an Internet store may mean that customers are unlikely to make purchases from that store. The level of trust consumers have in an Internet store influences their attitude toward the store, and hence the risk they perceive in purchasing from that store. However, trust was not significantly related to intention to transact, suggesting that trust acts indirectly through perceived risk and attitude. Perceived risk reduction was shown to be a direct antecedent of intention to transact online, suggesting that this is a key factor in consumer adoption of e-commerce. Reduced risk associated with buying from an Internet store would likely increase the likelihood a consumer purchases from that store (Jarvenpaa et al., 1999). In an attempt to understand the predictors of Web use, Liaw (2002) adapted the TAM to create a conceptual model termed the "technology use model" (p. 141). This model proposed that perceived usefulness, perceived enjoyment, and behavioral selfefficacy (a perceived behavioral control element) predict intention to use the Internet. Successful interactions (past behavior) with technology were viewed as influencing self-efficacy. The model proposed that perceived enjoyment has a direct effect on perceived usefulness (Liaw, 2002). All three variables (perceived usefulness, perceived enjoyment, and behavioral selfefficacy) significantly predicted behavioral intention to use the Web. Usefulness was found to be the most important predictor of intention, whereas Web enjoyment was found to be the least important. Experience, both with computers in general and the Internet in particular, was found to affect Web self-efficacy, a result consistent with the TPB (Ajzen, 1991), which states that self-efficacy expectations are one determinant of user behavior (Liaw, 2002). Store reputation has been shown to significantly predict intention to shop online (Pavlou, 2001). Similarly, Jarvenpaa et al. (1999) found that the reputation of a store had a greater effect on trust than the size of the store. These findings validate the expectation that Web retailer reputation plays an important role in consumer behavior (Jarvenpaa et al., 1999; Pavlou, 2001). A conceptual model based on constructs from the TRA was tested for predictive power with regard to future intentions to purchase online (Fitzgerald, Kiel, & Drennan, 2001). For both adopters and nonadopters of online purchasing, attitude was significant in predicting future use, with 49% of variance for
adopters and 32% for nonadopters being explained. It was concluded that attitude should be included in the model to predict future use intentions for both adopters and nonadopters of online purchasing. In a study by Shim et al. (2001), transaction services, convenience, sensory experience, and merchandise emerged as four factors that together explained 61.7% of the variance in attitude. Attitude toward Internet shopping was represented by transaction service, which explained 38.9% of the variance by itself. Intention to use the Internet for purchases was indirectly affected by previous Internet purchase experience, perceived behavioral control, and attitude toward Internet shopping. Intention to use shop-bots was investigated by Gentry and Calantone (2002), who applied the TRA, the TPB, and the TAM and used structural equation modeling to compare the results. Shop-bots are software that informs buyers which Internet retailer will give them the best price for a particular product. When common method bias was taken into account, the hypothesis that all three models would be appropriate in a network environment was supported. The TAM explained 91.6% of the variance in behavioral intentions, with the TRA explaining 75.8% and the TPB, 85.8%. It was concluded that within a procurement context, the TAM was superior because it explained more variance in behavioral intention and demonstrated superior model fit. It was believed that the TAM benefits by not considering subjective norms, which the authors believed could be unreliable. The Expectation-Disconfirmation theory of consumer behavior outlines the two key outcome factors of perceived quality and satisfaction. McKinney, Yoon, and Zahedi (2002) outlined the development and application of such measures, reporting that the metrics possessed a high degree of validity and reliability. They concluded by calling for research examining the relationship between these constructs and subsequent purchase behavior. Henderson and Divett conducted such research exploring the relationship between expectation-disconfirmation satisfaction and use (log-on/deliveries/purchase value). The results indicated that satisfaction was significantly correlated with each use variable and that satisfaction could be adequately modeled using an expectation framework. Although there is some disparity about whether usefulness, ease of use, and fun are precursors to attitude, or attitudinal elements themselves, there appears to be substantial empirical support for the influence of these constructs upon intention. Usefulness, ease of use, and enjoyment (fun) directly effect intention to use the system. Furthermore, the effect of perceived ease of use upon intention appears to occur indirectly through perceived usefulness (see Table 31.2 for a summary). Service elements such as the quality of presentation and information and service satisfaction also need to be considered, as should the aspects of risk/trust when purchase processes occur as these may have a direct impact on system use. Subjective Norms, Social Pressure, and Intention. Although the TAM represents an important theoretical contribution to understanding IT usage and IT acceptance behaviors, it fails to account for social influences. Research conducted by Malhotra and Galletta (1999) found that social
31. Uptake and Use of Web-Based Services
•
583
TABLE 31.2. Summary of Empirical Findings for the Relationship Between Attitude and Intention Attitude Precursors
Attitude Attitude toward Internet shopping
Attitude Usefulness Ease of use Fun
Internet shopping intention Internet search intention Intention
Attitude
Intention Intention to use online service Intention
Satisfaction
Service use
Usefulness
Attitude
Intention/behavior
Ease of use Usefulness
Attitudes
Ease of use
Enjoyment Usefulness and ease of use
.32-49
.49 .55
.73 .10 .30–.47
Intention to use the Web Intention to use new technology
.18-20 .30–.46 .20–36
.81 .69 .35
Trust
Intention to transact Risk/control
.64 —.56— .77
Risk/control
Willingness to buy
—.16— .29
influences played an important role in determining the acceptance and usage behavior of adopters of new information technologies (Malhotra & Galletta, 1999). Specifically, when social influences generated a feeling of compliance, the users' attitude toward use of the new information system "was negatively influenced. Expectations of future use based on compliance demonstrated a weak relationship with actual system use. Conversely, when social influences generated a feeling of internalization and identification on the part of the user, the effect was a positive influence on user attitude toward the system. The subsequent relationship between attitude and expectations of future use was strong. Internalization of the induced behavior by the adopters of new information system was found to play a stronger role than perceived usefulness in predicting consumer behavior. Although social influences were related to acceptance and usage of IT, no statistically significant relationship was demonstrated between social influences and behavioral intentions
Researchers Shim, Eastlick, Lotz, & Warrington (2001)
.28
.08 .34–.46
Supermarket use
Ease of Use Usefulness
.35
.43 .45
Attitudes
Ease of Use Enjoyment Usefulness
Correlation
.65 . 13 .41
Satisfaction
Usefulness Usefulness Ease of use Usefulness
Intention/Behavior
Fitzgerald, Kiel, & Drennan (2001) Davis (1993) Igbaria, Schiffman, & Wieckowski (1994) Koufaris (2002) Henderson, Rickwood, & Roberts (1998) Hu, Chau, Liu Sheng, & Van Tarn (1999) Henderson & Divett (in press) Mathieson, Peacock, & Chin (2001) Childers, Carr, Peck, & Carson (2001)
Henderson, & Divett (in press) Liaw (2002)
Venkatesh & Morris (2000) Pavlou (2001) Jarvenpaa, Tractinsky, & Vitale (1999)
(Malhotra & Galletta, 1999). It appears that social influence has direct effects on the users' attitude and indirect effects on behavioral intention via attitude. Contradictory evidence regarding the relationship between Subjective Norms and Intentions has also been reported by other researchers. For example, Davis (1986) excluded subjective norms from the original TAM specification because of its "uncertain theoretical and psychometric status" (p. 986). This choice of exclusion seems to have been supported by the findings reported by Gentry and Calantone (2002), who believed that the TAM benefits by not considering subjective norms, which can be unreliable. In contrast Liao, Shao, Wang, and Chen (1999) found support for the efficacy of Subjective Norms, as did Henderson et al. (1998). Conversely, Roberts (1999) found no support for the inclusion of the subjective norms construct. These findings tend to suggest that our understanding of the Subjective Norms construct has not really advanced since the
584
HENDERSON, HENNESSY, DIVETT
TABLE 31.3. Summary of Empirical Findings for the Relationship Between Subjective Norms, Attitude, and Intention Subjective Norms
Attitude
Psychological attachment Subjective norms
Attitude toward using the system Attitude
Subjective norms
Attitude
Explained Variance/Correlation
Intention
.30 Intention to use online purchasing
.58
Intention to use virtual banking
.54
comments made by Davis (1986), that is, "of uncertain theoretical and psychometric status" at least within the online environment (Table 31.3). Further research should be conducted that explores the relationship between subjective norms and behavior within this context. It appears to these authors that some moderating factors are present. One likely area of investigation is the nature of the activity, and in particular, its public/private nature. One could suggest that subjective norms may be more relevant within public activities, and not as important with regard to private activities.
Relationship Between Perceived Behavioral Control, Intentions, and Behavior Purchasing via the Internet does not occur simply because a consumer decides to shop. Factors such as skill, time, and Internet connections—"control" factors need to be further investigated. Control is divided into perceptions of computer self-efficacy (internal control) and perceptions of facilitating conditions (external control). Control is a construct that reflects constraints to behavior. Where the achievement of behavioral goals depends on resources such as access to computers and computer skills, the concept of perceived behavioral control is essential. Perceived behavioral control can be seen as the consumer's perception of the ease or difficulty associated with performing a behavior. The TPB proposes that perceived behavioral control, in conjunction with attitude and subjective norm, is a direct predictor of behavioral intention (Fig. 31-5). When studying intention to purchase books online, Koufaris (2002) added intrinsic enjoyment, perceived control, and concentration (used in flow research) to the TAM. Unplanned purchases were also included in the model, which attempted to predict intention to return. Venkatesh (2000) extended the TAM to include the determinants of system-specific perceived ease of use. These determinants were classified as anchors ("general beliefs about computers and computer usage," p. 345) and adjustments ("beliefs that are shaped based on direct experience with the target system," p. 345). Prior to direct experience with the target system, users will "anchor" their perception of the ease of use of the technology based on their general beliefs regarding computers and prior experience with computers/software in general. As their experience with the system increases, users adjust their
Researchers Malhotra & Galletta (1999) Henderson, Rickwood, & Roberts (1998) Liao, Shao, Wang, & Chen (1999)
perception of the ease of use of the system in accordance with their interaction with the system. The revised model proposed that control, intrinsic motivation, and emotion were the anchors that determine early perceptions about the ease of use of a new system. Anchors were found to explain 40% of the variance in perceived ease of use (Venkatesh, 2000). Adjustments made for higher levels of experience were found to play a key role in determining perceived ease of use. In this instance, the variance explained increased to almost 60% (Venkatesh, 2000). Venkatesh and Morris (2000) concluded that perceived control is an important precursor to the perceived ease of use for the system. Specifically, control (internal and external conceptualized as computer self-efficacy and facilitating conditions, respectively), intrinsic motivation (computer playfulness), and emotion (computer anxiety) serve as the anchors that users employ when forming an impression about how easy a new system is to use. With increasing direct experience with the system, objective usability, perceptions of external control (related to the specific system environment), and perceived enjoyment from system use were the contributing adjustments. Computer selfefficacy and facilitating conditions were stronger determinants than were these adjustments. These adjustments could play a more important role in influencing perceived ease of use of the new system when the user's experience is inconsistent with their initial anchors. Perceived resources can be viewed as a specialized subset of perceived behavioral control. This construct refers to the extent to which an individual perceives that she has the necessary resources to use a particular information system (Mathieson et al., 2001). Despite believing that a system is useful and easy to use, an individual may perceive that the resources needed to use it are not available. A perceived resources instrument was developed and tested by Mathieson et al. (2001). Their analyses suggested that this instrument captured a construct different from usefulness and ease of use. The results showed that perceived resources were related to usage intention and perceived ease of use, with a weak link to perceived usefulness. A version of the TAM that included social influences and behavioral control attempted to model computer usage for experienced and inexperienced users (Taylor & Todd, 1995). The model accounted for 21% of the variance in behavior and 43% of the variance in behavioral intention for experienced system users. For inexperienced user, 17% of the variance in behavior and 60% of the variance in intention was accounted for.
31. Uptake and Use of Web-Based Services
•
585
TABLE 31.4. Summary of Empirical Findings for the Relationship Between Perceived Control, Intention, and Behavior Perceived Control
Intention
Perceived control
Intention
Perceived resources Self-efficacy
Intention
Self-efficacy
Intention
Behavior
Explained Variance/Correlation
Researchers
.43-60
Taylor & Todd (1995)
Behavior Behavior Usage
.17–.21 .13-29
Mathieson, Peacock, & Chin (2001) Compeau, Higgins, & Huff (1999) Liaw (2002)
.18 .50
TABLE 31.5. Summary of Empirical Findings for the Relationship Between Past Behavior, Intention, and Behavior Prior Behavior Prior behavior
Intention
Behavior
Intention to use an online supermarket
.39–.41
Use of online supermarket Internet purchase experience
Explained Variance/Correlation
Internet shopping intention Internet search intention
Researchers Roberts (1998)
.58-62 .53
Shim, Eastlick, Lotz, & Warrington (2001)
.47
A longitudinal study by Compeau, Higgins, and Huff (1999) showed that self-efficacy was a strongly significant predictor of anxiety and affect, as well as use, 1 year later. Self-efficacy explained 18% of the variance in usage when both direct and indirect effects were included. This research finding has been supported within the Web environment by Liaw (2002), who reported that Web self-efficacy positively related to use of the Web in the future. Table 31.4 provides a summary of the perceived control/intention/behavior research.
for information via the Internet was previous Internet purchase experience. These researchers demonstrated relationships of between .47 and .53 between previous Internet purchase experience and intention to use the Internet to search for information or to shop. In summary, empirical research conducted within the online environment appears to consistently report a direct relationship between prior behavior and intention, as well as prior behavior and future behavior (Table 31.5).
Influence of Past Behavior on Intention and Behavior
Role of Individual Difference Variables
Again, as with the intention-behavior link, the influence of past behavior upon intention and actual behavior has also received general support. Meta-analytic research of 150 studies conducted by van den Putte (1993) reported that the knowledge of past behavior significantly improved the prediction of both intentions and future behavior. Specifically, the results indicated that knowledge of past behavior improved the prediction of intentions by 11% and behavior by 34%. Similarly, Roberts (1999) reported correlations of prior behavior and intentions between .39 and .41, and of prior behavior and behavior between .58 and .62. However, in Robert's study, prior behavior was the most powerful predictor of future behavior. This view is supported by Shim et al. (2001), who reported that the most influential predictor of intention to search
It should be noted that individual difference variables are usually beyond the control of the Web service designer. Similarly, the developer cannot manipulate self-efficacy, but can manipulate such things as menu complexity and recovery, and hence make the system more fault tolerant. As a result, traditional HCI design principals need to be considered during the interface design process and the developer should at least be cognizant of individual difference factors that may impact service use. Computer playfulness and computer anxiety are two such individual difference variables. It may be prudent to keep in mind the results of the following studies. Computer playfulness is seen to be system independent. The more "playful" an individual is, the more likely it is that the system will be used "just for the sake of it." This notion is
586
HENDERSON, HENNESSY, DIVETT
noteworthy, as playful individuals on average may "underestimate" the difficulty of using a new system as they enjoy the process and do not perceive it as being effortful. There is likely to be a positive relationship between general computer playfulness and system-specific perceived ease of use. In general, more playful individuals are expected to rate any new system as being easier to use compared with those who are less playful (Venkatesh, 2000). However, the role of computer playfulness is expected to diminish over time. As users become familiar with the system, computer playfulness should give way to systemspecific perceived enjoyment. Computer anxiety relates to an individual's general perceptions about computer use. These perceptions usually involve
apprehension, or in some cases, fear, when confronted with the possibility of using a computer. Computer anxiety has been shown to have a significant impact on attitude, intention, behavior, and performance (Venkatesh, 2000).
Summary The research highlighted suggests that the theoretical model as outlined in Fig. 31-5 provides a useful conceptual beginning point. A brief summary is presented in Table 31.6. The research also highlights that the model can be expanded to gain a better understanding of web uptake and use (Fig. 31.6).
TABLE 31.6. Brief Research Summary of Factors Related to Web-Based Service Uptake and Use Element Perceived usefulness Perceived ease of use
Research Support
Perceived behavioral control
High High, but may act through other elements High, but may be context related Moderate, more research required Moderate, more research required Moderate, more research required Weak, more research required High
Past behavior
High
Perceived fun Service satisfaction Information quality Trust/risk Subjective norms
Degree of Developer Control High High High High High High Low, may have some indirect control through media Low, some elements not subject to developer control Low
FIGURE 31.6. Detailed model of elements impacting Web service uptake and use.
31. Uptake and Use of Web-Based Services
APPLICATION OF THEORY TO UPTAKE AND USE OF WEB SERVICES The research evidence, although scant in parts, is clear. Each element as proposed in Fig. 31.5 and detailed in Fig. 31.6 holds a statistically significant and positive relationship with Web-based service use. This means that these factors all hold a positive relationship with service use that is better than chance. That is, an increase in users' Positive past behavior experiences (with both similar and target systems) Attitudes (perceived usefulness, ease of use, etc.) Subjective norms (with regard to friends, relatives, and significant others) Perceived behavioral control (in terms of access boundaries, self-efficacy, etc.) toward the Web-based service will be accompanied by an increase in the probability of their using the service (and other similar services). It makes some sense therefore to consider each element when developing a Web-based service, as such consideration will reduce the risk of service failure. It should also be emphasized that the developer has more control over some elements than others. For example, the developer has no direct control over the past experiences of clients, they also have no direct control over the subjective norms held by the client group, nor do they have direct control over some perceived behavioral control elements. The developer does have direct control, however, over the usability of the site, they do have direct control over the usefulness of the site, and they do have control over the service deliver elements and the quality of material published on the site. The developer also has control over the messages that are sent from the site and, therefore, may indirectly have some control over the subjective normative elements. For example, Roberts, Henderson, and Rickwood (1997) found that some home-based people with young children believed that it was their job to pick the groceries for the family, and hence believed some normative groups (partner/friends/parent) may disapprove of their use of the service. This scenario was quite easily countered by emphasizing that the groceries in their delivery were being hand picked by professional grocers, ensuring they would receive only the best produce, and that this freed time to spend with children. Having established that there is empirical support for the conceptual model advocated in Fig. 31.5 and detailed in Fig. 31.6, the topic of how to practically apply this knowledge becomes pertinent. This component of the chapter examines determining factors, and discuss the practical implications. We will then conclude with a conceptual checklist that developers may use to consider the potential of their site to better meet client needs in such a way that client will first choose to access the site, and then choose to return and use the Web-based service.
•
587
The Client: A Short Interlude The first step when applying volitional models of behavior to a service is to consider the client base of the Web service. Specifically, we need some conceptualization of the core target client groups. We are continually surprised as to the lack of conceptualization of the client and client needs undertaken by many Web-based service developers. When we ask, "who are your clients?", we still receive blank stares on many occasions. It amazes us that a Web service can be designed and implemented without some conceptualization of the intended user, perhaps akin to manufacturing clothes without knowledge of body size. This reinforces to us that it is the technology and not the client that has been the focus. Similarly, we have yet to encounter a Web service that has an entirely homogenous client type. The main problem with not adequately considering a client type is that one size does not fit all: Tuning the service for one client type may inadvertently compromise the service for another client type. This second aspect of tuning a service was vividly highlighted to us when we examined the optimization of service delivery within an electronic supermarket. Our focus group research indicated that at least four quite different client types existed. First, we found that one distinct group was high-income executives who valued the convenience of the service in the form of time saving. Premium prices seemed not to be a concern for this user type, but convenient delivery times were important. Second was a group of people who were homebound (elderly/parents with young children/physically disabled). This group valued the home delivery aspect of the service, but did not desire to pay a price premium. Third, there were inner-city flat dwellers that valued the home delivery aspect of the service due to parking problems associated with unloading the groceries. Finally, there were individuals who clearly enjoyed the use of the electronic media as a shopping media for the novelty of it. The concern was that optimization required the identification of each type and then offering different service delivery models to meet the needs of each type, or at least optimizing service delivery to the valued client types. For example, the executives were happy to pay a premium for tight delivery windows, whereas the home-based clients were not. To met home-based client needs, a lower cost service could be offered that had broader delivery windows. There are a host of ways to articulate the main client types that vary in complexity and expense. We have found using a simple Who/Why matrix (Table 31.7) to elicit client types to be a cost-effective starting point. When using this approach start by preparing a matrix with approximately six rows and five columns. Next, ask yourself, "who comes to the site?" We have found that developers or administrators can usually articulate between three and six client types, and that this proves a satisfactory starting point. Place the "who" in each row as applicable. Next, for each client type ask "why do they come to the site?" We find that it is useful to consider four possible reasons: to obtain information, to obtain product/service, to provide information, and to provide product service. For each client type, place the relevant reason in the relevant cells. Although this is not exhaustive, having completed this exercise, you will have
588
HENDERSON, HENNESSY, DIVETT
TABLE 31.7. WHO/WHY Client Breakdown Why (Why do the clients access the service?)
Who (Who accesses the service?)
Obtain Information
Obtain Product/Service
Provide Information
Provide Product/Service
Client type 1 Client type 2 Client type 3 Client type 4
a general feel for the main client types and why they access a service. The important thing to note about the Who/Why system is that it is generally the "why" that is the more informative than the "who." In this context, we have explored the predictive power a number of psychographic attributes (who-type features) and motivations (why-type features) to predict Web service uptake and use, and have consistently found that motivations are better predictors of Web service uptake and use than psychographic attributes. We have also found this exercise to be quite helpful in shifting the focus of developers from a technological focus to a client needs focus. It must be stressed, however, that this exercise is not comprehensive, but from experience we have found that this technique elicits client types that are satisfactory for initial investigations. We have also found that, in most cases, between four to six main client types can be derived. Our suggestion that if you derive more than six, you may want to reconsider the client groupings, and if you have less that four you may have too few categories.
Uptake: A Change in Behavior Having identified our main client types, the next step is to consider the service in relation to the identified reasons for using the service. Here we find it useful to consider core change concepts. Remember we are considering how to change an individual's behavior (to use our service) and then consolidate their behavior to continue to use the service. There are two general schools of thought that may be applied. The first is the classic approach based on individual attitudes. This approach assumes that if we alter an individual's attitude they will alter their behavior to be consistent with their attitude. Thus, we aim to present information that results in favorable attitudes that in turn lead to the formation of intentions to use and subsequent use. The second model is the converse. That is, changing a behavior will result in a change in attitudes to reflect the behavior. Thus, offer use and attitudes will alter to justify the use. Beginning with the first model, the developer should consider the core aspects of change as conceptualized by Beer (1980). Specifically the developer should consider three steps that a user will consider before considering whether to use a service: (a) is there a tangible problem with the current system/
way of doing things?, (b) does the service offer a solution to the problem?, and (c) what are the costs (time, effort, money) of changing from the current way of doing things and adopting the new approach. Each step should be considered as a hurdle to be overcome in order for the Web service to be a success. Specifically, if there is nothing wrong with the current approach, then why change? If the new approach is no better than the old approach, then why change? If the effort/cost of adopting the new approach is not worth the effort, then why change? These notions have lead Beer, Eisenstat, and Spector (1990) to highlight that if the cost of changing behavior is high (greater than the perceived benefit), then change will not occur. The notion of reducing entry barriers to be as low as possible therefore becomes pertinent. For example, we more recently examined an online office produce purchasing system. Although the system was easy to use, we found the payment registration so arduous that we simply gave up and continued using our existing telephone ordering system. In this case, we believed that the telephone system created enough frustration to warrant exploring the online system and we believed that the online system would remove our frustrations, but the cost of making the change made the exercise simply not worth it. So how can we ascertain if there are significant frustrations with current systems? We suggest either running online or offline research to elicit salient beliefs. One way to do this is to ask people what are the three best and worst aspects of the current approach and the three best and worst aspects with the online approach. If no system is present at this stage, we have found storyboarding will suffice. You can then examine the topics that are raised. Table 31-8 presents the outcome of this exercise when exploring a proposed online home appliance system, and Table 31-9 presents the outcome when exploring an online cosmetics and pharmaceutical outlet. For simplicity's sake, we have combined the data to form one client type, but the exercise may be applied to different client types to further optimize the system. The advantages of the physical appliance retail outlet were the product range, prices, and staff expertise. Conversely, the salient disadvantages were pressure from staff, poor advice, parking problems, and store-preferred brands. When considering the online retail outlet, the product comparison and time to browse factors were advantages, and the disadvantages were not being able to physically touch/view and size the products.
31. Uptake and Use of Web-Based Services TABLE 31.8. Advantages and Disadvantages Associated With Brick-and-Mortar and Proposed Online Home Appliance Outlet Bricks-and-Mortar Product range Prices Expertise Product comparison Time saving (one location) Pressure from staff Poor advice from staff Parking problems Store preferred brand Mark up
Online Advantages Product comparison Time to browse Time saving (travel) Convenience Overcome poor service Disadvantages Physical touch/view picture Inability to size object Inability to compare with other stores Uncertainty of prices Inability to view manufacturers' quality information
Return of large products Busy weekends Poor competitor price comparison TABLE 31.9. Advantages and Disadvantages Associated With Brick-and-Mortar and Proposed Online Cosmetics and Pharmaceutical Outlet Physical
Online Advantages
Expert advice Range Personal assistance Pampering Specials Fresh products
Price Inconvenience Pushy staff Queues Time Range Opening hours Parking Disorganized Access Nonexperts
Convenience Time No pressure Speed Parking Fun Delivery Easy access to departments Price Self-education Lack of queues Disadvantages Lack of advice Unable to see/touch Ambience Reading time Wrong product orders Credit card hassles Delivery lag No sampling System log on
From this knowledge, the retailer can plan an online system to complement the traditional physical store by building a system that uses the advantage of range variety and prices of the physical store, but enhances the product comparison feature, while removing the problems associated with poorly trained and pushy staff. This then allows an individual to gather infor-
589
mation at their leisure. The main disadvantages of the electronic system are the lack of physical touch and sizing of objects. This problem may be alleviated by a good return policy. An examination of the beliefs with regard to the cosmetics and pharmaceutical dispensing agency highlights a less optimistic view of the online service features than was the case of the home appliance service. Specifically, the main problem is associated with the cosmetics component, where users viewed the online service as deficient for a number of reasons. First, many cosmetics are of an olfactory nature, which cannot be conveyed well via electronic media; second, many viewed the process of shopping for cosmetics as pampering, and therefore valued the richness of the interaction with the staff and the premises of the cosmetics department. Conversely, the online component of the health product section did not suffer from these deficiencies because many health products are not of a sensory nature. Thus, it was found that the main advantages of shopping at a physical pharmacy are the expert advice and product range. In contrast, the disadvantages are the price, inconvenience, and pushy staff. The online store does in some way complement the physical shopping system in that it alleviates some of the inconvenience factors and the pressure to purchase. Concern must be expressed, however, because the price disadvantage is not alleviated, and more important, if the online store cannot supply at least the same level of expert advice offered by the physical pharmacy, it may be seen as lacking. This shortcoming may be compounded by the fact that the purchaser cannot touch and smell products in the online store. We suggest that the developer consider the problems facing each client group. What are the problems they are experiencing, and what is the inconvenience associated with each problem? Next, how well does the service reduce or eliminate the problem? Remember, if the service does not eliminate or significantly reduce the pain associated with problems, the service uptake will be low. Finally, the developer needs to consider the pain associated with the change of behavior. Does the system have a complicated or long registration procedure? How similar is the look and feel of the interface to existing systems used by the user group? How about the navigation of the system? Speed of operation? And so on. It should be noted that some organizations (e.g., banking institutions, government, sole providers) could increase the pain associated with the current forms of interaction (e.g., by closing outlets, stipulating the form of interaction, reducing service levels), thereby making a new system appear more attractive. However, as highlighted by Malhotra and Calletta (1999), such manipulations may result in resentment with associated negative outcomes. Generally speaking, however, developers do not have control over the pain associated with current systems. Developers potentially may exert some influence by increasing the perceived pain via advertising, but such efforts may be expensive, and the effectiveness may be very context specific. The developer does, however, have the ability to reduce the cost associated with the change and increase the benefit. Both of these may be addressed by examining each of the elements highlighted in Figs. 31.4 and 31.5 and relating them to the need identified for each client type.
590
HENDERSON, HENNESSY, DIVETT
The second approach to change is the view that if behavior is altered, then attitudes will alter to be consistent with the behavior. When applying this model, the strategy would be to provide a service to the user (presumably through some form of incentive). The hope would then be that once the user was accessing the service they would continue to use. This approach may be useful when there is short repeat use cycle time. Specifically, by providing incentive to access the service, the provider is essentially lowering the entry barriers to the system. The logic is that once the individual has become familiar with using the system the developer has essentially raised the exit barrier as to use another system the user would need to learn the other system, which has an inherent cost. We believe that this strategy may have some merit: however, it does still rest upon the assumption that the system meets the client needs in terms of the elements discussed in Fig. 31.5. That is, if the system is not useful, easy to use, etc., then the client will not develop a stable use pattern, and has an increased potential to stop using the service. For this reason, it may be useful to discuss the elements associated with repeat use now.
Repeat Use: Retaining Clients Our thesis has been that in order to be successful any Web-based service must retain core clients. Figures 31.4 and 31.5 outlined some of the aspects that have been identified as being predictive of use. We must now consider what can be done in the operational environment to increase the probability of repeat use. Past Behavior. As highlighted previously in this chapter, past behavior is a powerful predictor of future behavior. Unfortunately, how to use knowledge of past behavior in the operational environment is problematic. Also, the developer does not have direct control over past client behavior, and so does not have the ability to directly influence this aspect of use. There are, however, at least two ways in which past behavior may be used in the operational environment. One is to identify potential users of a service, another is in modeling continued use of a service and detecting potential client defections. When considering potential users, the analysts could look for either other online services that perform similar functions, or brick-and-mortar services that conduct similar functions. In both cases, the analyst should undertake the task of asking themselves (or client) "do these current services fulfill the complete needs of the client groups?" "Are there significant frustrations with the current service?" and "Can my, or the proposed, service overcome frustrations?" If the answer to these questions is "No," then the merit of the Web service may be dubious. The second way past behavior may be used within the Web services domain is through examining behavioral shifts in service use and, in particular, reductions in service use. Humans tend to be creatures of habit. Once a behavioral pattern of use has been established, relatively stable patterns of use should continue. This is particularly so with short repeat frequency cycle service, such as grocery shopping. As a result, changes
in patterns of use may indicate potential churn. Through the use of such things as short-run control charts, the Web service provider may identify individuals that are decreasing their use of the service. In such cases, we suggest deploying low-cost service recovery interventions, for example, an automated e-mail to the potentially defecting client designed to increase client voice and perceived service responsiveness. Such a system has been reported and tested within the brick-and-mortar environment by Divett (2002). Divett demonstrated that client repeat patronage could be enhanced through of a low-cost information booklet designed to enhance perceived approachability and responsiveness. In the electronic world, such interventions may be triggered by electronically recorded behavioral shifts in use. The advantage of this approach is that as the information system is electronic and therefore has no cost to the organization. In addition, these systems represent no immediate risk to the organization, in that if the e-mail is sent to a client that is not thinking of defecting, the intervention will only serve to further increase repeat patronage.
Attitudinal Elements. The key attitudinal elements outlined in Fig. 31.5 consisted of three site elements (perceived usefulness, perceived ease of use, perceived fun), two service elements (perceived quality and service satisfaction) and trust/risk/reputation. For each need identified for each client type, the developer should consider how the service would address each element of importance. What is useful for this client type? How is it useful in fulfilling their service needs? What has been done to ensure the system is easy to use? Is the service of a high quality? What have we done to ensure satisfaction with the service? How has the service been developed to convey trust for economic transactions? What is the client economic risk when interacting online? What is the service reputation? Once a user has accessed the service, we believe that it is imperative that the organization also takes steps to encourage repeat patronage. In the past, we have witnessed organizations conducting large user acquisition drives, only to then see a high churn rate. For this reason, we advocate a continuous sampling of user attitudes predictive of subsequent use. The logic is that through a steady inflow of client service performance information the Web developer can monitor service performance, identify service failure trends and then act upon such trends in a responsive way. We suggest that data should be gathered from active Web users and data should be presented in real-time to service managers. Data should be presented in such a format that allows for a brief examination, yet still provide enough content to allow for detailed analysis when service failure is identified. We also advocate a mix of quantitative ratings for benchmarking purposes and qualitative feedback for diagnostic purposes to provide a balanced view of service performance. We have developed one such system that may be used as an illustrative example. The system is the ClientWise Web services monitoring system. Data are gathered from Web users through a random pop-up requesting service feedback and through the explicit feedback link on the site. The pop-up requests site feedback and displays a brief privacy statement. The site user is given the options of answering
31. Uptake and Use of Web-Based Services
•
591
FIGURE 31.7. Sample OCCA™ screen view displaying client feedback. the inventory, not answering the inventory, or completing the inventory later. Using the system, we find that the response rate varies depending on the type of site, but generally a 1 in 40 pop-up to response rate is obtained. By this knowledge, and the knowledge of the psychometric properties of the instruments, we can set the pop-up rate to obtain the minimum amount of responses required to make robust decisions. Through this approach we also effectively ensure we do not sample more clients than needed to make decisions. It makes no sense to ask more clients than is absolutely necessary to obtain a robust service estimate. Upon agreeing to complete the inventory, the user is presented with a short service performance inventory. The inventory requests the user to identify their client type and to rate the Web service on six short item sets (site usefulness and usability, service quality and satisfaction, preference for interacting with the organization online, and their intention to use the service the next time they interact with the organization). The user is then invited to provide any further comments they may have about the site and service. Once the inventory is submitted, depending on the results given, the user will either be asked to elaborate on areas that received low scores or thanked for their assistance. When low scores are given, one further page is generated requesting the user to elaborate on the areas they believed were below standard. It should be noted that these further items are all open ended in nature. We have tested both rating style and openended formats and believe that the open-ended feedback provides better diagnostic information that can lead to immediate remedial action. As a result, the inventory is a benchmarking and alerting and service performance improvement tool. Scores are gathered on elements that have been shown to be both theoretically and empirically related to service use. Where scores are deemed to be below the acceptable service performance threshold, the system moves to gather more qualitative diagnostic data in the form of targeted open-ended comments. Gathering of such client-based service performance data, however, is only the first step in the service monitoring pro-
cess. Once data has been submitted, they are available for display through the OCCA™ display tool. The OCCA™ system is a real-time display mechanism that presents the data to the Web manager in a format that has both graphic and numeric components. The graphic component provides the manager with a simple visual impression of the site's service performance. Numeric data, in contrast, are provided to give a more accurate estimate where further examination is desired. The open-ended feedback is also displayed to let managers identify problems experienced by users (Fig. 31.7). Of course, there are many other features to the system, such as service failure alerting, trend analysis, and action planning, that are beyond the scope of this chapter. The point to note, however, is that with such systems the developer may actively monitor theoretically and empirically verified client service impressions, leading to close service monitoring and action where required to enhance repeat patronage.
Subjective Normative Elements. Generally speaking, the developer does not have direct influence on the subjective norms held by the user community. Remember that subjective norms are the views that the user holds with regard to what they feel significant others would think of them using the service. The empirical support regarding the importance of this construct is mixed. Obviously, if the site is not of a public nature, that is, something that is not openly discussed, the importance may be low. However, in other cases, such as grocery shopping it may be more important. Again, the importance will also depend on the social reference group. We suggest to again conceptually explore the most common referent groups a user type would reference, and to ask the question, "how could these people influence the community?" and "can we use this knowledge to influence our target group?" The developer should always be considering possible negative and positive influences may be used when positioning the service. Remember, the aim is to ensure the service provides an attractive alternative to other methods of interaction by reducing the pain associated with conducting tasks.
592
HENDERSON, HENNESSY, DIVETT
Perceived Behavioral Control. The impact of perceived behavior control is twofold. The first, and most obvious, is the accessibility notion. Here, if people cannot access, or perceive they cannot access, the system use will then be low. As a result, individuals or groups that have limited access to systems will be problematic. Conversely, sites that have been designed to employ certain restricted technologies may also have problems. Other aspects of perceived behavioral control worth considering are self-efficacy, computer anxiety, and so on. The concern here is that these factors are not under the control of the site developer because they are essentially individual personal characteristics. We suggest, however, that the designer remember that not all users will be of a certain disposition, and build in traditional good HCI practice when considering the interface.
CONCLUSION As highlighted at the beginning of this chapter, developers must continually ask the question, "why would someone use my service?" When asking this question, the developer should first consider basic behavior change elements. Specifically, for change to occur the client must experience a significant frustration with regard to the current system. The replacement system must then offer significant advantages over the existing system. Moreover, the effort required in the change of behavior must be less than the cost of making a change. To ensure the developer adequately considers change concepts, we advise the developer to begin by exploring the services client types and, in particular, identify why a particular
TABLE 31.10. Developer Checklist Consideration Area Experience with similar systems Experience with bricks-and-mortar systems Usefulness Ease of use Fun Satisfaction
Information quality
Trust/risk Reputation
Parents
Friends
Partner
Self-efficacy Access barriers
Elements to Consider (for Each Client Type) Past behavior experience (no direct developer control) What are the shortcomings of current systems? How does the Web service overcome current shortcomings? How different is the Web service from other services? What are the shortcomings of these systems? How does the Web service overcome these shortcomings? Attitudinal elements (direct developer control) How useful is the system at facilitating task completion? Can the complete task be conducted online? What has been done to make the system easy to use? Is the system fun to interact with? What is the client satisfaction level? What systems have been put in place to measure and track service delivery? What is the quality of the material contained on the site? What systems have been put in place to measure and track service delivery? What is the user financial/information risk associated with interacting with the site? What is the reputation of the service/organization? How will the reputation impact the willingness to interact online? Subjective norms (indirect developer control) Are parents a reference group to this client type? What message will they convey to potential users? What message can you convey to potential users? Are friends a reference group to this client type? What message will they convey to potential users? What message can you convey to potential users? Are partners a reference group to this client type? What message will they convey to potential users? What message can you convey to potential users? Behavioral control (Indirect Developer Control) What has been done to improve perceived self-efficacy? Is the system fault tolerant? Is the system easy to access for the clients? Are specialist downloads going to restrict use? Registration/interaction procedures?
31. Uptake and Use of Web-Based Services
client type would use the service. The Who/Why system is one way of conducting this exercise. Once reasons for accessing the service have been identified, the developer should consider how the service would meet the needs of each client type. Remember, the aim is to provide a service that reduces or eliminates significant problems for the client. We suggest that the developer should examine the service in relation to the factors highlighted in Figs. 31.5 and 31.6 and presented in Table 31.10. It should be noted, however, that this checklist should be further enhanced to meet the individual needs of each client type and service offering. When using the checklist, the developer should consider how the service addresses each problem identified for each client type. The aim is always to provide a service that solves people's problems.
593
Once operational, we suggest that service performance monitoring systems are implemented and continually used to ensure the probability of repeat patronage is enhanced. Monitoring should consist of measuring the psychological factors discussed within this chapter that have been theoretically and empirically shown to be predictive of future use.
ACKNOWLEDGMENTS The authors want to acknowledge the assistance of Professor Frank Deane and Peter Caputi from the Department of Psychology, University of Wollongong, and Meredith Monroe of ClientWise Pty Ltd. for their assistance in the preparation of this manuscript.
References Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179–211. Ajzen, I. & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Upper Saddle River, NJ: Prentice Hall. Anschau, K. (1999). A meta-analysis and empirical evaluation of the Technology Acceptance model. Unpublished doctorial dissertation, University of Canberra, Australia. Beer, M. (1980). Organizational change and development: A systems view. Glenview, IL: Scott, Forestman. Beer, M., Eisenstat, R., &Spector B. (1990, November-December). Why change programs don't produce change. Harvard Business Review. 158-166. Bentler, P. M., & Speckart, G. (1979). Models of attitude-behavior relations. Psychological Review, 86(5), 452-464. Bentler, P. M., & Speckart, G. (1981). Attitudes "cause" behaviors: A structural equation analysis. Journal of Personality and Social Psychology, 40(2), 226-238. Chen, L.-D., Gillenson, M. L, & Sherrell, D. L. (2002). Enticing online consumers: An extended technology acceptance perspective. Information and Management, 39(8), 705-719. Childers, T. L., Carr, C. L., Peck, J., & Carson, S. (2001). Hedonic and utilitarian motivations for online retail shopping behavior. Journal of Retailing, 77, 511-535. Compeau, D., Higgins, C. A., & Huff, S. (1999). Social cognitive theory and individual reactions to computing technology: A longitudinal study. MIS Quarterly, 23(2), 145-158. Davis, F. D. (1986). A technology acceptance model for empirically testing new end-user information systems: Theory and results. Cambridge, MA: MIT Sloan School of Management. Davis, F. D. (1993). User acceptance of information technology: System characteristics, user perceptions and behavioral impacts. International Journal of Man-Machine Studies, 38(3), 475-487. Divert, M. (2002). Consumer loyalty re-assessed: Measurement, determinants and intervention. Unpublished doctorial dissertation, University of Wollongong, Australia. Fazio, R. H. (1986). How do attitudes guide behavior? In R. M. Sorrentino & E. T. Higgins (Eds.), The handbook of motivation and cognition: Foundations of social behavior (pp. 204-243). New York: Academic Press. Fazio, R. H. (1989). On the power and functionality of attitudes: The role of attitude accessibility. In A. R. Pratkanis, S. J. Breckler, & A. G.
Greenwald (Eds.), Attitude structure and function (pp. 153-179). Hillsdale, NJ: Lawrence Erlbaum Associates. Fellenstein, C., & Wood, R. (1999). Exploring e-commerce: Global e-business, and e-societies. Upper Saddle River, NJ: Prentice Hall. Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Reading, MA: AddisonWesley. Fitzgerald, L., Kiel, G., & Drennan, J. (2001). Applying a consumer acceptance of technology model to examine adoption of online purchasing. Proceedings of the Australian and New Zealand Marketing Academy Conference. Gentry, L., & Calantone, R. (2002). A comparison of three models to explain shop-bot use on the Web. Psychology & Marketing, 19(11), 945-956. Hausenblas, H. A., Carron, A. V, & Mack, D. E. (1997). Application of the theories of reasoned action and planned behavior to exercise behavior: A meta-analysis. Journal of Sport and Exercise Psychology, 19(1), 36-51. Kelson, H. (1964). Adaption-level theory: An experimental and systematic approach to behavior. New York: Harper & Row. Henderson, R. (1999). Choosing electronic supermarket shoppers. Behaviour & Information Technology, 18(3), 225-228. Henderson, R., & Divett, M. (in press). Customer satisfaction and purchase behavior within an electronic supermarket service. Henderson, R., & Divett, M. (2003). Perceived usefulness, ease of use and electronic supermarket use. International Journal of HumanComputer Studies, 59, 383-395. Henderson, R., Rickwood, R., & Roberts, P. (1998). The beta test of an electronic supermarket. Interacting with Computers, 10, 385399. Howerton, B. (2002, January 1). Keeping promises. Intelligent Enterprise, 46-47. Hu, P, Chau, P., Liu Sheng, O., & Yan Tarn, K. (1999). Examining the technology acceptance model using physician acceptance of telemedicine technology. Journal of Management Information Systems, 76(2), 91-112. Igbaria, M., Schiffman, S. J., & Wieckowski, T. J. (1994). The respective roles of perceived usefulness and perceived fun in the acceptance of microcomputer technology. Behaviour & Information Technology, 13, 349-361.
594
HENDERSON, HENNESSY, DIVETT
Jarvenpaa, S. L, Tractinsky, N., & Vitale, M. (1999). Consumer trust in an Internet store. Information Technology and Management [Special issue on e-commerce], 1(1), 45-71. Kalita, M. (2002, December 16). Will Web services slip or leap? Pinnacle Systems, Inc. [On-line]. Available: SearchWebServices.com Kocharekar, R. (2001, November 12). Strategic support. Intelligent Enterprise, 14-15. Koufaris, M. (2002). Applying the Technology Acceptance Model and Flow Theory to online consumer behavior. Information Systems Research, 13(2), 205-223. Landgrave, T. (2003). Web services: The confusion continues. TechRepublic. [On-line] Available: http://techupdate.zdnet.com/ techupdate/stories/main/0,14179,2908615,00.html Landis, D., Triandis, H. C, & Adamopoulos J. (1978). Habit and behavioral intentions as predictors of social behavior. Journal of Social Psychology, 106, 227-237. Liao, S., Shao, Y, Wang, H., & Chen, A. (1999). The adoption of virtual banking: An empirical study. International Journal of Information Management, 19, 63-74. liaw, S. S. (2002). Understanding user perceptions of World-wide Web environments. Journal of Computer Assisted Learning, 18, 137148. Malhotra, Y, & Galletta, D. F. (1999). Extending the technology acceptance model to accountfor social influence: Theoretical bases and empirical validation. Paper presented at the 32nd Hawaii International Conference on System Sciences Maui, Hawaii. Mathieson, K., Peacock, E., & Chin, W W (2001). Extending the technology acceptance model: The influence of perceived user resources. The DATA BASE for Advances in Information Systems, 32(3), 86112. McKinney, V, Yoon, K., & Zahedi, F. (2002). The measurement of web-customer satisfaction: An expectation and disconfirmation approach. Information Systems Research, /3(3), 296-315. Moston, S. (1997). Shop until you drop. Internet Business, 10, 30-32. Oliver, R. (1977). Effect of expectation and disconfirmation on postexposure product evaluations: An alternative interpretation. Journal of Applied Psychology, 62, 480-486. Pare, G., & Elam, J. L. (1995). Discretionary use of personal computers by knowledge workers: Testing of a social psychology theoretical model. Behavior and Information Technology, 14, 215-228. Pavlou, P. A. (2001). Consumer intentions to adopt electronic commerce—Incorporating trust and risk in the technology acceptance model. Paper presented at the DIGIT Workshop, New Orleans, Louisiana. Regan, D. T., & Fazio, R. H. (1977). On the consistency between attitudes and behavior: Look to the method of attitude formation. Journal of Experimental Social Psychology, 13, 38-45. Roberts, P. (1999). Predicting and understanding electronic supermarket shopping behavior: A comparison of the theory of reasoned action and the theory of planned behavior. Unpublished dissertation, University of Canberra, Australian Capital Territory, Australia. Roberts, P., Henderson, R., & Rickwood, D. (1997). Assessment of modal beliefs about using an electronic supermarket. In S. Howard, J. Hammond, & G. Lindgaard. Human-Computer interaction INTERACT'97, 577-578. Shim, S., Eastlick, M. A., Lotz, S. L., & Warrington, E (2001). An online pre-purchase intentions model: The role of intention to search. Journal of Retailing, 7(3), 397-416. Taylor, S., & Todd, P. A. (1995). Assessing IT usage: The role of prior experience. MIS Quarterly, /9(4), 561-570.
Triandis, H. (1977). Interpersonal behavior. Monterey, CA: Brooks/ Cole. van den Putte, B. (1993). On the theory of reasoned action. Unpublished doctoral dissertation, University of Amsterdam, Amsterdam, The Netherlands. Venkatesh, V (2000). Determinants of perceived ease of use: Integrating control, intrinsic motivation, and emotion into the technology acceptance model. Information Systems Research, 11(4), 342365. Venkatesh, V, & Morris, M. G. (2000). Why don't men ever stop to ask for directions? Gender, social influence, and their role in technology acceptance and usage behavior. MIS Quarterly, 24(1), 115-140. Venkatesh, V, Speier, C., & Morris, M. G. (2002). User acceptance enablers in individual decision making abut technology: Toward an integrated model. Decision Sciences, 33(2), 297-316. Weaver, B. (2002, March). The holy grail goes digital. Online Currents, 17, 13-14.
SUGGESTED THEORETICAL READINGS Ajzen, I. (1985). From intentions to actions: A theory of planned behavior. In J. Kuhl & J. Beckman (Eds.), Action-control: From cognition to behavior, (pp. 11-39). Heidelberg, Germany: Springer. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179-2 11h. Ajzen, L, & Fishbein, M. (1977). Attitude-behavior relations: A theoretical analysis and review of empirical research. Psychological Bulletin, 84, 888-918. Armitage, C. J., & Conner, M. (2001). Efficacy of the theory of planned behavior: A meta-analytic review. British Journal of Social Psychology, 40, 471-499. Davis, F. D. (1986). A technology acceptance model for empirically testing new end-user information systems: Theory and results. Cambridge, MA: MIT Sloan School of Management. Davis, F. D., Bagozzi, R. P., et al. (1989). User acceptance of computer technology: A comparison of two theoretical models. Management Science, 35(8), 982-1003. Eagly, A., &Chaiken, S. (1993). The psychology of attitudes. Fort Worth, TX: Harcourt Brace Jovanovich. Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Reading, MA: AddisonWesley. Foxall, G. (1997). The explanation of consumer behavior: From social cognition to environmental control. In C. L. Cooper & I.T. Robertson (Eds.), International Review of Industrial and Organizational Psychology (pp. 229-287). New York: Wiley. Madden, T. J., Ellen, R S., et al. (1992). A comparison of the theory of planned behavior and the theory of reasoned action. Personality and Social Psychology Bulletin, 18, 3-9. Rogers, E. M. (1995). Diffusion of innovations. New York: Free Press. Sheppard, B. H., Hartwick, J., et al. (1988). The theory of reasoned action: A meta-analysis of past research with recommendations for modifications for future research. Journal of Consumer Research, /5(3): 325-343. van den Putte, B., Hoogstraten, J., & Meertens, R. (1996). A comparison of behavioural alternative models in the context of the theory of reasoned action. British Journal of Social Psychology, 35, 257266.
•32CONSUMERS: THE THEORETICAL FOUNDATIONS OF ONLINE BEHAVIOR Fred Volk Georgetown University & SAS Institute
Frederic B. Kraft Grand Valley State University
Consumer behavior is a complex set of related mental and physical activities that people pursue to conduct the exchange and consumption aspects of their lives (Markin, 1974; Peter & Olson, 2002). E-commerce solution providers must understand how consumers employ the Internet in their participation of these exchange and consumption activities. These activities can be conceptualized as responses to various internal (physical and mental) and external (social and physical—including marketing) stimuli. When consumers recognize unmet needs as a result of processing these stimuli, they engage in a sequential (and sometimes reiterative) series of goal-directed mental and physical activities aimed at satisfying these needs. These activities include the perception of a deficit or problem, a determination of available solutions for the problem, an evaluation of the alternative solutions, a choice of solutions, and postchoice and postconsumption activities. These activities have become known as the consumer decision process, the study of which has become an organizing model for many textbooks in consumer behavior. The nature and extensiveness of each stage of this process depends on the skills, experience, attitudes, knowledge, confidence, and behavioral tendencies of the consumer; the goal object (target of acquisition); context of the customer experience; user motivations; and the interaction between two or more of these. We must consider these behaviors within the context of the target product of the behavior and how that target limits or enhances the users' application of the Internet to support their consumer experience. As providers of e-commerce customer experiences, it is generally our goal to guide potential customers to a positive
595
postconsumptive experience in a timely manner (both from the customers' and the providers' perspectives). Some users come to their Internet session with the consumer decision process already engaged, whereas others have hedonic goals that do not involve the framework of the consumer decision process. Each user group and subgroup presents designers of e-commerce solutions with an opportunity to design advertising, product information, and Web-based product representations in a manner that assists users in clarifying their needs, reducing their perceived risk of meeting those needs, and obtaining the products and services that will enable postconsumption satisfaction. USER CHARACTERISTICS We begin by examining some of the ways that consumer researchers have categorized e-commerce users for analysis purposes. These research efforts assist us in identifying the variables that influence the different e-commerce usage behaviors and must be understood for successful e-commerce site designs. These variables include demographics, behavioral patterns, and psychological factors.
Demographics A traditional method of consumer analysis has been the description of the demographic and behavioral characteristics of
596
VOLK AND KRAFT
relevant consumer market segments. Numerous such studies of Internet shoppers have taken this approach with varying results. Variables such as gender, age, income, and education have been shown to be related to Internet consumer behavior. For example, gender is a powerful predictor of purchase behavior across product categories: Men are more likely to purchase electronics and computer products, and women are more likely to purchase food, beverages, and clothing (Bhatnagar, Misra, & Rao, 2000; Kwak, Fox, & Zinkhan, 2002). In addition, there is convincing evidence that adults and teenagers of different genders engage the Internet in different ways (Weiss, 2001). At the same time, Li, Kuo, and Russell (1999) reported a regression analysis of gender, age, income, and education level explaining only 4% of the variance associated with purchase frequency. This suggests that demographics play a role in the types of online activities and products consumers prefer, but have less to contribute when considering users' frequency of interaction. In his analysis of Internet sales across eight product categories, Reibstein (2001) identified seven segments of online consumers based on life stages. For the third quarter of 1999, teens-plus (consumers under 25 years old with no children living at home) accounted for 6.6% of the revenue associated with online sales; young adult consumers (users between the age of 25 and 34 with no children living in the home) accounted for 19.2% of revenue; average income adults (users between 35 and 54 years old, with no children living at home and an annual income under $50,000) accounted for 6.4% of revenue; high-income adults (consumers between 35 and 54 years old, no children living at home and an annual income of greater than $50,000) accounted for 18.5% of online revenue; average income adults with children accounted for 8.4% of sales; high-income adults with kids accounted for 26.6% of sales; and mature adults, those consumers over 54 years old with no children, accounted for 8.3% of online sales revenue (Reibstein, 2001). Members of each of these seven life stage segments reported by Reibstein (2001) were asked to rank the importance of five point of sale and five fulfillment merchant attributes. The five most important attributes across all segments were price, product representation (the degree that the product is presented consistent with reality), product selection, shipping and handling, and on-time delivery, respectively. At the same time, there were some important differences between segments. For example, mature adults ranked product representation as most important, whereas all other segments ranked price as most important, and they ranked product selection eighth, whereas all other segments ranked it no lower than fifth (Reibstein, 2001). This decrease in importance of price for mature users is supported by Lightner's (2003) findings in her study of e-commerce site preferences. Product selection also differentiated the young adult and teens-plus segments from the other segments, ranking it second, whereas no other segment ranked selection more important than fourth (Reibstein, 2001). The average income segments and the mature segment ranked customer support third, fourth, and third, respectively, whereas the other four segments ranked customer service as no more important than seventh.
Behavioral Patterns There are a number of ways that e-commerce users can be categorized by behavior patterns. Rodgers and Sheldon (2002) developed the Web Motivation Inventory (WMI), which characterizes Internet users as primarily researchers, communicators, shoppers, and surfers. Li et al. (1999) describe shoppers in terms of their purchase frequency (frequent, occasional, and never). Others, including consumers themselves, choose to categorize users in terms of their personality (Baumgartner, 2002; Kwak et al. 2002); socioeconomic status, age, and gender (Dholakia & Chiang, 2003; Weiss, 2001); and purpose (i.e., experiential vs. utilitarian; Hirschman & Holbrook, 1982; Holbrook & Hirschman, 1982; Weiss, 2001). Each categorization implies an approach for improving or optimizing the target groups' experience, with the Internet as a commerce medium. In an analysis of consumer traffic patterns, Moe (2003) identified five clusters of behavior that were indicated by three categories of dependent measures: session (duration at site, number of pages, and average time spent per page), category (percent of search result; information related, and homepages and number of pages across categories, products, brands), and variety (percent of pages across categories, products, brands). The first cluster of users, "knowledge builders," spent extensive time on each page and had a limited number of views of product pages. Moe (2003) interpreted these results as an indication of the Internet site users' desire to acquire knowledge rather than purchase a product. This segment is important because it represents users who do not intend to purchase in that given session, yet their behavior indicates the likelihood of future purchase. Because most Internet consumer research has focused on actual buyers, little is known about these types of users and their potential for generating revenue. There are those who have proposed that brick-and-mortar businesses take advantage of the Internet as a medium to support an overall strategy for revenue growth both online and in the physical store. More specifically, Steinfeld and Whitten (1999) argued that small local merchants could adopt a business strategy that leverages an online presence to more effectively serve their geographically co-located customers. As an example of managing the postpurchase satisfaction, a customer that buys a book from the Barnes and Noble Web site can return the book for an in-store credit at any local Barnes and Noble location. The second cluster of users, "hedonic browsers," spent the bulk of their sessions perusing category and product pages with a high percentage of unique page views (Moe, 2003)- These first two clusters, hedonic browsers and knowledge builders, represent a unique opportunity for Web designers. Providing knowledge builders with the types of tools that facilitate information acquisition, an information architecture that enables their desire to gain knowledge, and the ability to contrast our offerings objectively with competitors supports them in a way that is consistent with an overall customercentric strategy. Also, the hedonic cluster represents an opportunity for us to provide these users with the type of online experience that is attractive to the surfer of our e-commerce sites. Realtor.com provides an excellent example by offering users an opportunity
32. E-COMMERCE AND CONSUMER BEHAVIOR
to sort the home listings associated with a query with the Virtual Tour listings displayed first. This certainly supports these surfers, while not slowing down the knowledge builders. Although these types of implementations for these first two clusters may not translate into sales for that particular Internet session, they are likely to create a positive brand image for the site and its partners. The third cluster of users, "directed buyers," spent most of their session at product-level pages with a high number of repeat views at the product page level. The fourth cluster of users, "deliberate searchers," viewed the most pages and spent the most of their session in relatively few product categories across the range of products within that category with few repeat views at the product level (Moe, 2003). A fifth cluster of users, "shallow," accessed only two shopping pages and then left the site. The sales conversion rates for the "shallow" and "knowledge building" users was .01%, 1.4% among "hedonic browsers," 20.0% for "directed buyers," and 6.4% for "deliberate searchers." The behavior of each of these clusters suggests that consumers have a specific cognitive orientation or mind-set with which they approach a specific Internet session (Dholakia & Bagozzi, 2001). Understanding the demographics and behavioral characteristics of users is one piece of the puzzle and plays a considerable role in the success of an e-commerce effort. Although this section is not meant to be a complete discussion of all the potential user characteristics that could affect the success of e-commerce solutions, a reasonable set of user experience and business guidelines can be extracted to facilitate a positive user experience across a number of solution scenarios (Table 32.1). It is important to note that the previous paragraphs related demographic and behavioral groupings to cognitive patterns among the members of these behavioral groups. Thus, although we have noted the value of demographic and behavioral analysis, we must also consider the psychological characteristics of e-commerce users.
597
Psychological Variables Numerous psychological variables are associated with goaloriented and problem-solving behaviors by consumers. Some of the most important have received considerable attention in the consumer behavior literature. Particularly relevant to our discussion are motivation, perceived risk, trust, and involvement. Mind-Set. Consumer mind-sets include goal-oriented (deliberative and implemental) and experiential (exploratory and hedonic) mind-sets (Dholakia & Bagozzi, 2001). A deliberative mind-set refers to a cognitive orientation where the consumer is intent on collecting and processing information, and is common in the problem identification, information search, and decision-making stages of consumer behavior. However, an implemental mind-set refers to an action focused, cognitive orientation that occurs after the decision has been made, and serves to facilitate smooth action execution for goal attainment. It is possible to conceive experimental mind-sets where the consumer's focus is on the experience itself, without a specific end. For example, an exploratory mind-set refers to the consumer's cognitive orientation to encounter new experiences and to satisfy his or her curiosity. A second type of experiential mindset is the hedonic mind-set where the individual deemphasizes cognitions, and focuses instead on the sensory elements of the experience (Dholakia & Bagozzi, 2001). Motivation. Human behavior is goal oriented. The difficulty is in understanding those goals and designing an Internet site that provides an appropriate user experience that facilitates consumers' goal accomplishment in a manner that is mutually beneficial for consumers and provider. The specificity of consumer goals varies from instance to instance. For example, a consumer may use the Internet on one occasion to obtain a very specific product (target), such as the latest Harry Potter
TABLE 32.1. User Characterstics Design Recommendations 1. E-commerce sites that are primarily transactional in nature should target households with a high income and design for efficiency of purchase process (Kim, Cho, & Rao, 2000; Moe, 2003; Weiss, 2001). 2. A significant portion of online sales are generated by high-income households that have children living in the home. For average income households, those with children living in the home account for nearly 25% more sales than those with no children. Offer a set of products and services that are likely to be desired by families with children and a user experience that is supportive of that lifestyle (Reibstein, 2001; Weiss, 2001). 3. Lower-income households use the Internet more experientially than higher-income households. Provide an experiential online environment for those users who have a lower household income (Weiss, 2001). 4. Adult men and women and teenage boys and girls use the Internet differently. Design a user experience that is consistent with those differences (Dholakia & Chiang, 2003; Weiss, 2001). 5. Provide products and services that are consistent with early adopters/opinion leaders needs (Kwak et al., 2002). 6. When targeting older populations, spend your solution resources on increasing the accuracy of product representation as opposed to product selection (Reibstein, 2001). 7. Design for efficient knowledge acquisition to support knowledge builders. This is especially important for those sites that are a part of a multiple-channel business strategy (Moe, 2003; Steinfeld & Whitten, 1999). 8. Provide hedonic browsers an opportunity for pleasurable e-commerce activities. For example, providing a virtual tour on a real estate Web site may not translate into a referral in that session, but will help develop a positive experience association with the site brand for future business (Moe, 2003).
598
VOLK AND KRAFT
TABLE 32.2. Motivation Design Recommendations 1 . Associate registration benefits with increased efficiency in information acquisition, purchase, fulfillment, and customer support for utilitarian users (Rodgers & Sheldon, 2002). 2. Associate registration benefits with access to features that enrich the communicative and multisensory interactive nature of the consumer experience for experiential users (Holbrook & Hirschman, 1982; Kamali & Loker, 2002; Rodgers & Sheldon, 2002). 3. Provide experiential users with the product representations that increase telepresence (Rosa & Malter, 2003). 4. Provide utilitarian users with clean efficient user interfaces, minimizing the time and effort it takes them to attain their target (Lohse & Spiller, 1999; Moe, 2003; Rodgers & Sheldon, 2002). 5. Provide users an opportunity to choose the type of experience that they prefer relative to next-generation interactive experiences (Moe, 2003). 6. Do not assume that the experiential user will prefer an experiential purchase process. Regardless of purchase type, allow users to move from the experiential to the utilitarian per the task. For all users, facilitate efficiency from the decision point of purchase to actual purchase and fulfillment. 7. Provide opportunities for spontaneous purchases to the experiential users that may be outside the topical scope, but consistent with the apparent desired multisensory experience of current browsing activity (Baumgartner, 2002). For example, a user that is participating in an interactive product customization activity may be predisposed to enjoy other activities that require similar interactivity. 8. For deliberate purchases, offer consumers the types of products and services that are consistent with that product, for example, offering accessories to a cell phone or offering a second book title in the same area as the target book.
book. On a second occasion, that consumer will be searching for a gift, but has no idea what that gift might be, and on a third occasion the same consumer will have a goal of surfing the net and checking his or her favorite online haunts. The motivations represented by each of these occasions represent an opportunity for providers to advertise, sell, and provide support in different ways. In fact, as an example of such contrasting motivations, Reibstein (2001) found that when consumers were purchasing a product for someone else, product representation and on-time delivery were most important and price was fifth in importance. However, when consumers were shopping for themselves, price was the most important. The WMI is a 12-item scale that assesses Internet users' behavioral orientation on four dimensions (Rodgers & Sheldon, 2002). Shopping motivation is measured by three purchase items; research is measured by a "do research" item, a "get information I need" item, and a "find out things I need to know" item; surf items include three experientially oriented items; and communication motivation is measured by "e-mail other people," "connect with my friends," and "communicate with others" items (Rodgers & Sheldon, 2002, p. 88). Each of these scale dimensions had an acceptable alpha coefficient and reasonable test-retest reliability (Rodgers & Sheldon, 2002). In confirmatory studies of the WMI, Rodgers and Sheldon (2002) found that each of the dimensions was predictive of evaluation, intention to click, or feelings of persuasion from corresponding banner advertisements in student and adult populations. That is, banner advertisements promoting communicating behavior such as e-mail were associated with the communicate dimension of the WMI. Although the WMI is useful in categorizing the motivation for a given session, it effectively illustrates the limited characterization of consumer behavior to purchase behavior. An expansion of the characterization of online consumer behavior that includes product research and alternative evaluations would result in a set of motivations that may in fact coincide with both of Rodgers and Sheldon's research and communication motivation categories.
Utilitarian Versus Experiential Motives. From an e-commerce perspective, consumers either view the Internet as an experiential opportunity or as a utilitarian tool. The distribution of these two types of users is unknown, but there is evidence the online population spends the bulk of their time in experiential activities. For example, more Web surfers seek spiritual advice on the Web than participate in online banking (Weiss, 2001). Informed businesses recognize that the Web presents a unique opportunity and attempt to engage users in an experiential way as well as a utilitarian way. Experience-oriented consumers are interested in interacting with products because it is fun (Table 32.2). Their preferred online experiences are characterized by multisensory input and feelings of fantasy and fun (Holbrook & Hirschman, 1982). If user registration is associated with highly interactive (i.e., multisensory) experiential features as opposed to merely utilitarian benefits, experiential users will be more likely to view those features favorably and register on a given Web site (Kamali & Loker, 2002). It certainly could be argued that the types of consumer experiences that appeal to the experiential consumer are different than those that appeal to the utilitarian consumer. The experiential nature of the Internet increasingly permits the creation of experiences that enable telepresence and bricolage (Rosa & Malter, 2003). Telepresence refers to the degree that users perceive their presence in a virtual environment. Bricolage is the process of creating something out of what is available. Providing these features has benefits for both consumers and businesses. For consumers, features that improve telepresence facilitate elaboration and recall, information acquisition, and provide an opportunity to evaluate a wider range of products more thoroughly (Li, Daugherty, & Biocca, 2002; Rosa & Malter, 2003). For businesses, experiential features enable competitive differentiation, greater user customization, and make their Web site more satisfying for consumers who prefer experiential online sessions (Kamali & Loker, 2002; Li et al., 2002). All these combine to help develop a positive brand image
32. E-COMMERCE AND CONSUMER BEHAVIOR
599
TABLE 32.3. Personal Involvement Design Recommendations 1. For high involvement products, take care to provide users with enough information to make an informed decision (Kwak etal., 2002). 2. Hedonic and high involvement products are more likely to have strong affect associated with them, so it may be useful to provide a user access to experiential descriptions and interactive interfaces that reinforce positive imagery (Mowen & Minor, 1998; Sherif & Sherif, 1967). 3. Frequently repeated purchases should be facilitated by primarily utilitarian or efficiency-oriented interfaces or low involvement user experiences (Lastovicka, 1979). 4. For high involvement products, it is important for the site to support an extensive search for information. Users may be building their knowledge in a given product area for future purchase, and site designers should optimize their experience for attaining product knowledge that is consistent with the strengths of their products (Bloch, Sherrell, & Ridgeway, 1986; Moe, 2003). 5. For high involvement hedonic products, it is important to provide a well-designed experiential customer experience that is consistent with the target products (Hirschman, 1980; Peter & Olson, 2002; Tauber, 1972). 6. Saving time is a principle motivation for researching and buying products online. High involvement products typically require an extensive information search. Providers that are selling high involvement products need to allocate more resources to the design of their site to ensure users can acquire information efficiently (Beatty & Smith, 1987; Stigler, 1961). 7. For low involvement products where users are not motivated to attend to product information, designers should employ hedonic messages or content to encourage the elaboration of the low involvement item (Petty, Cacioppo, & Schumann, 1983; Mowen & Minor, 1998). 8. The use of graphics for high involvement must be informative in either a multisensory aspect for hedonic purchases and/or in terms of providing more information to the user than could be efficiently or accurately delivered via text (Netguide, 1996).
for online businesses that account for the broader range of online users. Designing for the utilitarian consumer has been the focus of Web site designers for some time. In data collected in 1996, Lohse and Spiller (1999) used regression techniques to determine that 61% of the variance associated with monthly sales was accounted for by navigation mechanisms that reduced time to purchase. Although these results are dated, they suggest the importance of efficiency to utilitarian consumers, and it is reasonable to use efficiency as a principle in designing other aspects of consumer behavior, such as information search and alternatives evaluation. The fact that customers prefer efficiency during the purchase process is obvious when we consider that even the most experientially oriented consumer is engaged in the utilitarian task of purchase and monetary exchange. However, the focus of online purchasing as the ultimate measure of Web site success has resulted in the understandable emphasis on streamlining the purchase process and a lack of exploration of the importance of experiential motives of consumers who use the Internet. Web site designers should recognize that experiential interactions are a major goal for many consumers in using the Internet. Personal Involvement. Personal involvement is one of the most important concepts in the analysis of the consumer decision process and the various factors that determine its nature. Involvement, as originally suggested by Krugman (1965), refers to a person's perception of connections between his or her own life and a stimulus such as a product, idea, or activity. This perception of the personal relevance of the use, acquisition, or disposal of a product in a particular situation causes people to experience what Celsi and Olson (1988) referred to as "felt involvement." In one more recent study, product involvement as measured by the degree of informa-
tion required for purchase, was reported as the most powerful predictor of purchasing products online (Kwak et al., 2002; Table 32.3). Consumer involvement is highly important because it is a motivational state that influences the level of cognitive and behavioral effort consumers expend, as well as the level of affect they experience during the purchasing process (Antil, 1984). The level of consumers' involvement influences the amount of attention they pay to communications and how much they think about the product information communicated (Celsi & Olson, 1988). Involvement also determines the extent and type of information processing in which consumers will engage (Petty, Cacioppo, & Schumann, 1983), the amount of belief and attitude change produced by communications (Sherif & Sherif, 1967), the relationship of attitude development and change to purchase behavior (Ray, 1973), and the likelihood of cognitive dissonance following a purchase (Insko & Schopler, 1972). Also, it has been found that affect toward objects of high involvement may be experienced much more strongly than affect for those of low involvement (Sherif & Sherif, 1967), although the reverse is also true in that objects that create high emotions, such as hedonic purchases, may also be more highly involving (Mowen & Minor, 1998). Consumers experience involvement with products and services for two broad reasons. First, their knowledge of a strong linkage between a product and their own important personal goals produces an intrinsic or enduring involvement (Bloch, 1982). The linkage exists because consumers view the consequences of acquiring or using the product or service as a means for reaching these goals. Although this linkage and the resulting involvement level reside in the minds of consumers, clearly some types of products (e.g., frequently purchased commodity-like goods) are more likely to be ones that people have low involvement, whereas other goods (e.g., high-priced,
600
VOLK AND KRAFT
infrequently purchased and brand differentiated items) are ones with that people are likely to be highly involved (Lastovicka, 1979). A second reason why consumers may feel involvement is the purchasing situation. Purchase situations are temporarily perceived linkages between the purchase and important outcomes (Peter & Olson, 2002). The immediate physical or social environment of a purchase may activate a temporary awareness of important goals that a purchase may achieve. These may include purchases involving temporary emergencies, or purchases that confirm one's shopping acumen, that impress friends, or that take advantage of a sale (Bloch, 1982). Once the purchase is made and the temporary means-end linkage has decreased in salience, the level of involvement decreases. An important way that consumers' involvement level is of concern to commercial Web site designers is its influence on the extent that consumers will spend effort to search among sites and to probe succeeding layers of information within a particular site. In addition, involvement level determines the most appropriate design of information presented in a Web site. Two types of consumer search processes may occur, "ongoing" search or "prepurchase" search (Bloch, Sherrell, & Ridgeway, 1986). Ongoing search, sometimes referred to as browsing or window shopping, may be motivated for two reasons. First, consumers who have high intrinsic involvement with a product category search because of their continuing interest in and desire to build knowledge that may be accessed during future purchase situations. For example, an automobile enthusiast may examine photos of Corvette sports cars on an auto dealer Web site even though no purchase is contemplated in the foreseeable future. Such search is a product-specific phenomenon for members of particular market segments (Bloch et al., 1986). These consumers require that the information from sources be adequate in breadth and depth. Ongoing search may also be driven by an experiential or hedonic motivation. This form of search is motivated by the enjoyment and stimulation of the activities of the search process (Tauber, 1972) or by a desire to seek novelty (Hirschman, 1980). This hedonic form of search requires that the "aatmospherics" of the search source provide rewards, whether they be from attractive physical store designs that cause consumers to linger or from "asticky" Web sites that provide enjoyable interaction experiences (Peter & Olson, 2002). The second type of search is the "prepurchase" search, that is motivated by the consumer's need to acquire information to make a purchase decision. The higher a consumer's involvement with a product or product category, the greater the motivation to acquire information to reduce the risk of a poor purchase (Beatty & Smith, 1987). The value of searching, however, is also related to the amount and quality of knowledge consumers already have. The economist Stigler (1961) pointed out that from an economic perspective, consumers should search only while they perceive that the value of search outweighs costs. An important appeal of shopping via the Internet is that it decreases the costs of shopping in terms of time and travel. Therefore, Web page designers must make shopping as easy and straightforward as possible and work toward decreasing the perceived risks in purchasing via this medium.
Another way that the involvement construct is important in consumer analysis is its influence on perceptual processes. Celsi and Olson (1988) found that involvement strongly influences people's motivation to attend to and comprehend information. Not only are highly involved individuals more motivated to process information, but they also have a greater ability to comprehend information because they more easily activate knowledge from long-term memory. Related to the impact of involvement on perception, involvement also determines the role of communication in influencing consumer attitudes and decision making. Krugman (1965) was one of the earliest to note that when consumer involvement is low, attitudes do not change prior to purchase, as in the traditional Hierarchy of Effects (cf. Lavidge & Steiner, 1961). Rather, attitudes toward low involvement products change following simple belief change and purchase. Research on the low involvement hierarchy led to the development of the "Aelaboration likelihood model" by Petty et al. (1983), according to which there are two routes to persuasion. The first is the "central route," which corresponds to the traditional hierarchy of effects model of attitude change. This central route occurs when consumers are high in involvement and when they are motivated to process and elaborate on the product information they receive. The provision of relevant product information may lead to attitude change followed by a change in purchase behavior. The second route to persuasion is the "peripheral route" that occurs when consumers are low in involvement and not motivated to process information. In this case, consumers may be influenced in spite of their low product involvement because they still attend to peripheral, nonproduct information for its entertainment or hedonic value. Such messages encourage higher involvement with the advertising itself, and effective message might include humor, pleasant music, attractive models, or even soap opera-like stories. Positive feelings generated toward the advertisement itself may later become associated with the brand name sponsored in the ads (Mowen & Minor, 1998). A final point is that the research on low versus high involvement persuasion is very relevant to the type of information and graphics presented by a Web site. For high involvement purchasing, informative data and perhaps pictorial information is mandatory, whereas splashy graphics may be detrimental (Netguide, 1996). However, for low involvement items that are not actively sought, Web sites must capture attention and prolong interaction with attention-getting stimuli in order to influence buyers through the low involvement route. Involvement is fundamental to understanding (a) the motivation of consumers to search for information, (b) consumers' willingness to pay attention to information, (c) consumers' motivation and ability to comprehend information, and (d) the type of information most likely to influence them. Web site developers should be guided by the involvement construct as they determine the stimulus and information characteristics of the Web pages with which consumers will interact (Table 32.3). Purchase Type as Defined by Experiential and Cognitive Dimensions. Purchases can be categorized by the extensiveness of decision making, the level of consumer
32. E-COMMERCE AND CONSUMER BEHAVIOR
involvement, and the extent of consumers' emotional processes in the purchase. When the extent of these is known, several implications for Web site design may be derived. Baumgartner (2002) identified eight categories of purchases using multidimensional scaling and cluster analysis technique that were defined by three dimensions. He classified each purchase as either thinking or feeling, spontaneous or deliberate, and having high or low involvement. In this context, it is useful to define deliberate purchases as those consumer situations where a user approaches the consumer decision process and its associated shopping tasks with a defined purpose or goal (e.g., purchase office supplies). Purchases that require extended decision-making processes are also described as "goal-oriented" consumer purchases. Symbolic purchases include those that enhance a consumer's self-image or social status such as designer clothes or a particular make and model of automobile (Belk, Wallendorf, & Sherry, 1989). A purchase that has "... multi-sensory, fantasy and emotive..." evaluative components can be described as hedonic (Hirschman & Holbrook, 1982, p. 92; Holbrook & Hirschman, 1982,). This classic definition is consistent with Baumgartner's (2002) notion of symbolic purchases. Spontaneous purchases are unplanned purchases and include promotional, casual, exploratory, and impulsive purchases. These purchases occur in the context of a user's engagement in online activities and are often made without consideration beyond the moment. Promotional purchases refer to those unplanned purchases that occur as a result of a vendor's special offer, casual purchases are those that consumers perform without much deliberation, exploratory purchases are made when a user is curious or is seeking some variety, and impulse purchases are those that are executed without any thought beyond the immediate stimulus situation. People who have had the experience of purchasing books on Amazon.com have no doubt received a special offer that consists of the target book with another book on a similar topic or by the same author. That special offer has, in the past, been associated with a slight decrease in price from purchasing the two books separately. It is also useful to describe spontaneous purchases, like deliberate purchases, in terms of these three bipolar dimensions. If we take a closer look at the deliberate purchase categories and consider how these might play out behaviorally, it is evident that these clusters are, as clusters by their nature tend to be, generalizations, and we must take great care in how we apply design recommendations based on this type of research. It is probably more useful to think of each of these purchase types as dimensions of any given purchase. A consumer may employ an extended decision-making process for hedonic, symbolic, or even repetitive purchases when participating in the consumer decision process (Belk et al., 1989). For example, the purchase of a Ford F150 could be rated highly on all four purchase-type dimensions. It might require extended decision making, be symbolic relative to image, repetitive because we always buy Fords, and hedonic in terms of pleasure seeking. That said, this research is extraordinarily useful in conceptualizing products in terms of the interaction between users and their likelihood to feel a certain way about particular products. By rating our products on these dimensions with our various
601
target user populations, we are able to consider the design of advertising content and nature of desired user experience with our product pages. Perceived Risk and Trust. Perceived risk refers to the degree that a consumer estimates the potential negative outcomes of a given behavior. Is the risk of purchasing a CD from Amazon.com greater than the risk associated with purchasing the same CD on eBay? Perceived risk is associated with personality, store reputation, quality of information, return policies, cost, and personal experience with a medium (Bhatnagar etal., 2000; Lee, Kim, & Moon; Pavlou, 2001; Schaninger, 1976). These risks can be described in terms of finances, performance, time, image, and nuisance (Ha, 2002): • Financial or monetary risk is the monetary costs associated with doing business with an online vendor and is comprised of the security associated with the transfer of electronic financial information and implications of dealing with a potentially fraudulent vendor (Wang, Lee, & Wang, 1998). • Performance risk is the degree that product performance meets the consumer's expectations and is especially of concern for consumers of products that are single purchase, durable, high-cost items (Bhatnagar et al., 2000). • Time risk is the degree that a particular behavior or relationship will increase or decrease asymptotic performance of goal accomplishment, and plays a role with regard to the user's external search for information. If a site is poorly designed from either an information architecture or task performance perspective, users may be dissuaded from using the site to search for product information due to perceived time risk. It also refers to the time that a product is expected to satiate the identified need. • Image risk is the degree to which a particular consumer behavior or relationship is congruent with the desired image of the participant. • Nuisance risk is the degree that participating in a given behavior or exchange of personal information will result in irrelevant vendor communication exchanges (Wang et al., 1998). As e-commerce has matured, there is some evidence that even though financial risks are still of concern to online consumers, they are not necessarily predictive of future online purchases (Miyazaki & Fernandez, 2001). The degree that marketers and designers take each of these types of risk into account is contingent on the degree that their target customers find these important (Table 32.4). Certainly, there are those marketers that presume an unlimited number of potential customers and direct market all their products, but most of us cannot ignore the contribution of these irrelevant contacts to consumers' perceived risk. As the Internet matures, there is indication that consumers' concerns about security risks relative to online purchasing (i.e., financial risk) are decreasing (Girard, Silverblatt, & Korgaonkar, 2002). In contrast, findings also suggest that, as the cost of a product increases, risk plays a larger role in consumers' decision process (Lowengart &
602
VOLK AND KRAFT
TABLE 32.4. Perceived Risk Design Recommendations 1. As a means of reducing perceived performance risk and financial, list the credentials of product and site evaluators (Grewal etal., 1994). 2. Designing for efficiency of use decreases the perceived time risk associated with navigating to the desired target (Fogg et. al., 2001). 3. Have professionals design the user experience. Customers recognize an amateurish effort when they see it and are likely to infer both security and information risks associated with establishing a relationship associated with the e-tailer (Fogg etal., 2001). 4. Organize the Web site in a manner that is consistent with the real world. It permits the user to find the information they desire and leverages current user models (Fogg et al., 2001). 5. Do not mislead the user by integrating content and advertisements in a manner that does not facilitate differentiation (Fogg etal., 2001). 6. Never use pop-ups (Fogg et. al., 2001). 7. Demonstrate benevolence by recommending competitors when it is advantageous to consumers (Fogg et al., 2001; Olson & Olson 2000). 8. Convey your expertise to the consumers (Fogg et al., 2001). 9. For technically complex products, provide easy-to-understand information to increase the customers' confidence with the performance of the product. The more technically complex, the higher the perceived risk of purchasing the product (Bhatnagar et al., 2000). 10. For ego-related products such as sunglasses or perfume, provide a money-back guarantee to decrease the image risk for the consumer (Bhatnagar et al., 2000). 11. The design should place emphasis on the reputation of the retailer (Jarvenpaa & Tractinsky, 1999). 12. Provide users an opportunity to have a positive personal interaction experience prior to purchase (Miyazaki & Fernandez, 2001; Olson & Olson, 2000). 13. Share information about shared values that helps the customer identify with the provider (Dowling & Staelin, 1994; Olson & Olson, 2000). 14. Provide risk handling tools on your site, including independent reviews by experts, richer examination of products when possible, or personal interactions (chat or e-mail) with salespeople (Dowling & Staelin, 1994; Olson & Olson, 2000).
FIGURE 32.1. Perceived risk. Tractinsky, 2001).* Although these findings appear counterindicative, these two concerns are really assessing two very different relationships in our model of perceived risk. Lowengart and Tractinsky (2001), through experimental manipulation, address the relationship between cost of the product or service and the evaluation of that cost relative to the cost of other possible outcomes (Fig. 32.1) in contrast to financial concerns over the security of the financial transaction (Girard et al., 2002). *Data were collected in Israel.
A consumer will participate in a behavior when the difference between the perceived risk of engaging in that behavior is meaningfully less than the perceived benefit or when the expected outcome associated with a behavior is the most positive on the range of estimated negative alternatives. That is, people will participate in behaviors that they believe will represent their best interest. A key issue in today's online environment is the level of perceived risk that is associated with participating in the range of consumer behaviors (Table 32.3). Trust and perceived credibility are two factors that play a considerable role in the perception of risk and how that perception affects consumer behavior. Trust is the degree that we permit another an opportunity to harm us by making ourselves vulnerable (Freidman, Kahn, & Howe, 2000). The relationship between trust and the perceived trustworthiness of an online vendor is directly related to customer loyalty and purchase intention (Gefen, 2002; Lee et al., 2000). Trust is founded on consumers' perception of the online provider's (a) skills and competencies, (b) commitment to serve the customer well, and (c) adherence to acceptable behaviors with regard to the proposed relationship (Gefen, 2002). Consumers place trust in a brand and/or retailer in terms of the five risk areas. The degree that consumers are willing to expose themselves to harm by participating in e-commerce behaviors is directly related to how much they trust the provider. The trust that is required is dependent on the behavior of the consumer. In the case of surfing very little trust is required, but in the case
32.
of information search and alternatives comparison a consumer must trust the accuracy and the information representation. In a high-trust scenario, users may use a single site as their source for product information and in a low-trust scenario they will use the information as a single input to a broader information search. A varying degree of trust is also required when users share their personal information during the registering process and provide their credit card number for purchase. Some people are more trusting than others. Perceived comprehensive information, shared values, and communication are antecedents to developing trusting relationships between consumers and online retailers (Gefen, 2002; Grewal, Gotleib, & Marmorstein, 1994; Lee et al., 2000). In a cross-cultural study of consumer trust for an Internet store, Jarvenpaa and Tractinsky (1999)* reported that store reputation and perceived size were reliable antecedents to consumer trust. These findings affect financial and nuisance risk through privacy and security antecedents as opposed to evaluation and information antecedents. An increase in trust has been consistently associated with embeddedness (connectedness to the community), reputation, and positive interpersonal relationships between buyer and supplier and ultimately a decrease in perceived risk (Jarvenpaa, Tractinsky, & Vitale, 2000; Steinfeld & Whitten, 1999). In their study comparing the purchase of computers (high risk) and books (low risk), they reported that the dimensions of the Internet store (e.g., shopping process, reputation, information quality) are more important for the high-risk condition (Lowengart & Tractinsky, 2001). In two comprehensive studies, the Persuasive Technology Lab at Stanford analyzed both qualitative and quantitative data from more than 1,400 and 2,600 participants that identified the attributes of Web sites that contribute to consumers' perceptions of credibility (Fogg et al., 2001; Fogg et al., 2002). In the first study, participants were asked to rate Web site credibility based on a fiftyone-item survey that included seven subscales. Five of the subscales increased perceptions of credibility (real-world feel, ease of use, expertise, trustworthiness, and tailoring), whereas two of the subscales contributed to negative perceptions of credibility (the commercial implications of the site and amateurism; Fogg et al., 2001). This certainly presents some challenges for e-commerce sites, considering that the mere suggestion that a site has a commercial purpose has negative valence. Other items of the Commercial Implications subscale that are relevant to e-commerce include the powerful negative ratings on pop-up ads and the difficulty associated with distinguishing ads from content. On the positive side, personalization, name recognition from other advertising mediums, and Web sites designed for e-commerce transactions contributed positively to the subscale (Fogg et al., 2001). Other subscales that have items with implications for e-commerce designers are the Trustworthiness and Ease of Use subscales. Stating policy regarding content, having links to outside resources, including customers, word-of-mouth recommendation of the site by a friend, professional look and feel, and design consistent with the users' mental model, all contribute positively to credibility,
"Data were collected in Australia, Israel, and Finland.
E-COMMERCE AND CONSUMER BEHAVIOR
603
whereas nonintuitive navigation models contribute to negative perceptions with regard to credibility (Fogg et al., 2001). In their analysis of open-ended comments and rankings of ten different e-commerce sites, Fogg and his colleagues (2002) found name recognition and reputation were the most powerful factors with regard to consumers' evaluation of a Web site's credibility. Alternatively, the appearance of the relatively unknown McMaster-Carr ahead of Best Buy and eBay on the ranking gives hope to lesser known online retailers (Fogg et al., 2002). In addition, comments about customer service were far more prominent on e-commerce sites than Web sites in general (Fogg et al., 2002). Consumer trust and Web site credibility are negatively related to perceived risk and requires a series of risk handling activities by the consumer (Bowling & Staelin, 1994). Primarily, these activities delay purchase by extending the information search process, including looking for a better price or better return policy at another store, consulting other evaluative sources such as friends and family or other advertising media, or merely reconsidering based on internally developed standards (Dowling & Staelin, 1994). Although these strategies are effective in reducing consumer risk, it would obviously benefit e-commerce providers to leverage those relationships that consumers already find comfortable. These include consumers' current relationships with the products, brands, and retailers that currently are perceived as low risk. Alternatively, facilitating risk handling tools, such as online live help, customer reviews, or seller rating, and employing risk reduction policies with regard to returns and fulfillment tracking can discourage customers from comparison shopping at other stores. Framing a product in a manner that will reduce the perceived risk of acquiring that product can make the difference between success and failure. Consumers have a tendency to behave differently under high-risk versus low-risk conditions. Research findings support the notion that under low-risk conditions consumers tend to choose the moderately incongruent option (the one that embodies a higher potential to be inconsistent with the consumers' preferences in terms of performance), but under high-risk conditions they tend to choose the more familiar option (Campbell & Goodstein, 2001). For example, if a user is shopping for a gift for her manager, she is much more likely to pick the gift option that has a performance value that she is confident about (familiar) than the option that she thinks might be better performing, but in which she is less sure (unfamiliar). This is opposed to the scenario where she is shopping for something for her own use and might be willing to choose the less familiar option because her perception of the risk with regard to the negative outcomes is lower. If the online retailer is overstocked with a moderately incongruous option, it is desirable from the retailer's perspective to present that option in lowrisk scenarios or reduce the perceived risk by employing methods such as taste tests, rebates, and free trials. Comparison between in-store experience and remote purchase environments suggests that a lenient return policy is more important in reducing consumers' perceived risk in remote purchase environments
604
VOLK AND KRAFT
(Wood, 2001). Having a positive experience with the Internet is consistently associated with a reduction of perceived risks of purchasing products online (Miyazaki & Fernandez, 2001). This certainly suggests need to attend to design issues that can affect consumers' overall assessment of the Internet session in terms of a general favorable or unfavorable view of the provider. In fact, it is reasonable to assert that those sites that have the lowest perceived risk associated with their customer-provider relationships will flourish in the digital environment.
THE ORGANIZING FRAMEWORK OF E-COMMERCE WEB SITE USE ANALYSIS: THE CONSUMER DECISION PROCESS Problem or Need Recognition The study of the psychological variables discussed plays a major role in the analysis of the consumer decision process (CDP). The CDP is initiated with the consumer's perception of a significant deficit, that is, the realization of a noticeable difference between the person's actual state of affairs and desired state regarding some need. This is required for the initiation of purposeful problem-solving activity by the consumer. The perception of problems may be caused by internal physical or mental state changes, or by stimuli in the external social or physical environment. Marketers have paid special attention to the manipulation of marketing stimuli in the consumers' environment and their role in stimulating problem recognition responses (Engel & Blackwell, 1982). The consumers' use of the Internet during the CDP may be triggered as a purposeful response to problem recognition, or Internet usage for unrelated reasons may provide stimuli that produce problem recognition by consumers. Stimuli at this stage of the CDP include banner and pop-up advertisements and e-mail lists, and may be facilitated by software agents that notify a consumer when a potentially desired product meets a preidentified threshold in terms of price, feature, or availability (Scheepers, 2001). For the most part, all these methods are intended to attract the attention of users, emphasize an apparent need that
the target (product or brand) will satisfy, and increase the probability that users will select the advertisers' product for meeting that need. Automatic and controlled information processing play a significant role in the design of online advertising (Grunert, 1996). Automatic processes fall into two distinct categories, preattentive and automatic (Logan, 1992). Although both preattentive and automatic processes are accomplished quickly, effortlessly, and without conscious attention, preattentive processes are those processes that were never in the conscious attention and operate independently of attention (Logan, 1992). They are certainly automatic but appear to be more closely linked with the physiological potential of the organism (e.g., pattern recognition). Automatic behaviors are those that, upon first effort, require attentional resources, but after practice slowly migrate to requiring less and less attention (navigating a Web page). A behavior would be considered automatic if it is characterized by single-step retrieval from memory (Logan, 1988). Controlled processes are effortful, take a sequence of steps, and are sensitive to attentional manipulations. Such skills are concerned with recognizing and categorizing stimuli, and finding appropriate reactions to them. They are "... processes concerned with the acquisition, storage, and use of information to direct behavior" (Grunert, 1996, p. 89). Grunert (1996) proposed a framework for advertising strategy and suggested a series of strategies to take advantage of experimental findings in attention, memory, and skill acquisition literature. These can be characterized into series of design recommendations (Table 32.5). Need recognition may either occur in the present, within the context of the users' current online session, or prior to the consumer engaging the digital environment (Moe, 2003). Both of these scenarios call for a different approach to Web design. If the user has already entered the information search process, the design and purpose of the advertising may be very different than when a user is merely surfing the net. For example, it makes more sense for the provider to display advertisements for DVDs and DVD player accessories for a user who is searching for information about DVD players, than for a user who is searching across several electronics categories. In terms of human perception and action, it has been found that click-through rate on banner ads can be increased by simply
TABLE 32.5. Consumer Decision Process Design Recommendations 1. Advertisements that are personally relevant to users will be automatically processed. Present users with advertisements that are relevant to their task, characteristics, and targets (Grunert, 1996). 2. Prior experience with a stimulus will result in automatic activation of that experience into conscious attention. For familiar products, designers should activate those experiences that are likely to reinforce the positive aspects of the product experience. 3. Novelty and ambiguity decrease the likelihood of initial attention, but once attention is captured they increase elaboration. Create a perceived knowledge gap in your target users. Although this is easier with new products, it is not uncommon for established products or brands to leverage the "new and improved" label to generate interest (Grunert, 1996). 4. Memory is a function of the number of instances that a stimulus is represented to the user (Grunert, 1996; Logan, 1988). 5. If the brand is unknown, positive attributions can be developed by exposure without engaging controlled cognitive processes (Grunert, 1996). 6. Provide a cue and that is consistent with the goals of the business and the strengths of the product or brand to guide the user to curiosity resolution. 7. Do not use curiosity-based advertising when consumers perceive the risks of investigating and acquiring that product as high. Under high-risk conditions, a product perceived high in novelty will have a higher perceived risk associated with it.
32. E-COMMERCE AND CONSUMER BEHAVIOR
including the imperative "click here" within the ad copy (Hofacker & Murphy, 1998). Although increasing click-through rate is attractive, the generic click-through does not necessarily translate into more sales. It is important that the right type of customers, those that are likely to buy during their current session or at some future time, click-through. For example, if the principle competitive advantage of an offering is price, then the site designer should develop advertising stimuli that will prime users to recognize needs based on criteria on which the firm has a competitive advantage. Most findings regarding advertising do not suggest that a consumer must take immediate observable action in order for that advertising to be effective. Experimental findings indicate that the priming of attributes will result in novice consumers spending more time in their external search on the primed attributes and ultimately influence choice (Mandel & Johnson, 2002). Users who are surfing the net may be gathering information for the development of standards for future evaluation through the unintentional processing of stimuli (Adaval & Monroe, 2002), and the experiential nature of the Web provides opportunity to take advantage of surfer curiosity to improve brand or product position in the information search process (Menon & Soman, 2002). In their study of the role of curiosity in the information search process, Menon and Soman (2002) found that a curiosity stimulus, a teaser for a feature of which the consumer was unaware, resulted in more extensive elaboration and better knowledge acquisition on the trigger topic. Their research also suggests that satisfying this curiosity too quickly inhibits elaboration and, as a consequence, learning (Menon & Soman, 2002). Further, learning is optimized when a moderate period of time is provided between the curiosity-generating stimulus and curiosityresolving stimulus. This allows the time for the user to generate hypotheses regarding the novel product and enhances elaboration (Menon & Soman, 2002). Last, it was reported that curiosity advertising positively affects product evaluation and increased perceived novelty. Using the curiosity trigger can apply across many situations and users, but will likely get the best results for the experiential Internet session. At the same time, for some products it "would be reasonable to suggest that designers can put together an online advertising strategy that may also trigger the curiosity of goal-directed users. None of these findings suggest the effectiveness of an advertisement should be measured by click-through rate, but rather the advertisement should be designed in such a way that is consistent with product positioning and grounded in empirical theoretical findings. It is not enough to design an advertisement to attract attention and encourage click-through. It must be designed to encourage the types of behaviors and cognitions that will result in the positive evaluation of the brand, site, and product. This positive evaluation, of course, depends on the degree that the advertisement enhances users' experiences. The decision on how to enhance the user experience with online advertisements is contingent on the goal of the users' online session, the strategy of the business or agency, the characteristics of the user in terms of their skill, perceptual capabilities, and attitudes (product, risk, and brand), and the position of the online retailer in the overall marketplace. Given that the goal of an ethical business is to provide a beneficial
605
experience for the customer, each of the previous factors must be considered to ensure the ongoing positive consumption experience of the customer and the financial future of the business.
Information Search and Alternative Evaluation After recognizing a problem, consumers require information for determining a solution. They must either retrieve information stored in their memories (internal search) or acquire new information from their environment (external search). The success of internal search depends on the ability to recall relevant solutions, and the related choice criteria and heuristics for choosing among them. As consumers begin their search for information, there is a tendency for them to be overconfident with regard to their own knowledge. In other words, they are poorly calibrated with regard to the facts and their knowledge. They underestimate their knowledge in low-confidence conditions and overestimate their knowledge in high-confidence conditions (Alba & Hutchinson, 2000). Poor calibration develops for a number of reasons: the tendency to remain confident as the facts of previous experience fade and overemphasis on autobiographical experience (Hoch, 2002), poor recall due to misinformation or irrelevant information at encoding and/or recall, the reinforcement and polarization effect of elaboration on currently held beliefs irrespective of their accuracy, and consumers' tendency to overestimate their own performance, to name a few (Alba & Hutchinson, 2000). The role of confidence in the information search process is obvious: The more confident the consumer, the less likely the consumer will engage in an extensive information search for alternatives. The less confident, the more likely the consumer will continue with his or her information search and delay purchase decision. From the solution provider's perspective, the preference is that target customers are confident and well calibrated with regard to their positive evaluation of the information, products, and services offered by the site. As the user gathers information about a set of alternatives for problem solution, an assessment of each of the alternatives is undertaken relative to perceived costs and benefits of attaining one of those options. A user stops the information search process when the perceived cost of gathering further information about the current set of alternatives (depth) or about other potential alternatives (breadth) is greater than the perceived risk associated with choosing one of the available alternatives and the consumers' confidence that an alternative will meet the identified need. Theoretically, the information search and alternative evaluation processes are distinct and sequential. From a behavioral perspective, however, information search and alternative evaluation occur in an iterative loop with each new relevant stimulus producing a new evaluative assessment. Because the iteration from search to evaluation and then back to search can transpire so quickly, they are often temporally indiscernible from a behavioral perspective. As previously discussed, marketers have devised several strategies to improve the likelihood that their brands will be activated in the consumer's memory with each new stimulus instance.
606
VOLK AND KRAFT
TABLE 32.6. Information Search and Alternative Evaluation Design Recommendations 1. Display utilitarian product information in a manner that maximizes efficiency of target acquisition {Lohse & Spiller, 1999). 2. Provide shortcuts for users to minimize the information search process with products that are repeat purchase or consumable items that they have purchased on another occasion (Hoch, 2002). 3. Provide consumer tools to evaluate alternatives at any time during the information search process. 4. Create online advertisements that prime information search and alternative evaluation activities. Emphasize those aspects of your product that position your offering favorably relative to your competition for a well-calibrated consumer. 5. Products that have a competitive advantage on the basis of price are more sensitive to negative word-of-mouth information (Ha, 2002). Do not provide access to user reviews or ratings if you are competing primarily on price. 6. Users will infer a higher-quality product based solely on a lenient return policy (Wood, 2001). Provide a lenient return policy if product quality is one of your competitive advantages.
Internal Search. The effectiveness of the internal search for information is predicated on the consumers' abilities to store information regarding the set of acceptable options for satisfying the identified stimulus deficit and effectively recalling that information accurately. The source of these memories is the sum of information and advertising stimuli and product use experiences of the consumer. Drawing on previous experiences is much more attractive to a user than the costs compared with conducting an external search (Hoch, 2002). Unfortunately, product experience can, in fact, result in a less than accurate diagnostic assessment of a product's potential for meeting consumers' needs. That is, consumers have a tendency to recall the set of their experiences with a product rather than the initial experience. For example, a user may have a particularly positive feeling with regard to her product experience because she has developed a set of skills in using that product that effectively meet her higher order needs. This can occur with products that have a steep learning curve. Experts rarely evaluate a product's usability costs based on the depth of difficulty of their first learning experiences, but rather, on a more global assessment of their most recent experience relative to the ease of use and functional benefit. The nature of personal experience is nearly always exaggerated in terms of accuracy, lack of bias, and recall (Table 32.6; Hoch, 2002). When a consumer recalls product usage, rationalization of that product experience as positive sometimes occurs irrespective of actual experience. Further, they rarely consider their own potentially inaccurate recall of product experience that may be attributable to other experiential factors. Needless to say, the efficacy of our internal search is highly dependent on our ability to recall in an objective manner. Objectivity based on our recollection of experiential and factual information is not always easy for us to achieve. The decision to engage in an external information search is quite different for the user if the need identified is mostly experiential in nature rather than merely utilitarian. That is, in the case of utilitarian products, where perceived risks are low, users are likely to spend little time in the information search process, but with experiential products, where risks are high, consumers are expected to spend a great deal of time in the information search process. External Search. If no satisfactory solutions can be retrieved, the consumer must acquire information externally. It is at this point that consumer behavior becomes observable.
The extensiveness of external search depends on the amount of retrievable knowledge, the degree of risk experienced by the consumer, and the perceived personal costs of acquiring the information. Information search plays a major role in reducing the perceived risk of the online consumer, and the perceived cost of acquiring product information via the Internet is critical to the use or disuse of the Internet as an information source (Ha, 2002). It is important for providers to keep the costs of acquiring product information on the Internet relatively low (Lohse & Spiller, 1999). External search for information can be very brief in cases of spontaneous and repetitive purchases. External resources include personal contacts, online sources, and print or other multimedia sources. The information on which users base their purchases may be integrated with the stimulus that triggered need recognition. For example, an advertisement for a vacation to Jamaica may include the most relevant criteria for the consumer to make the decision (i.e., price, duration). At the same time, deliberate searches for product information may take place across a number of sessions. Ha (2002) found that users rely on brand and word-of-mouth information and customization for registered customers to reduce the perceived costs associated with searching for information and purchasing products online. One way that consumers decrease the cost of searching for product information is through the use of intelligent agents, or shopping bots. These bots range from the very simple price comparison search engines to very complex intelligent agents that consider a user's preferences and history with other agents and operate semiautonomously as a proxy for the consumer (McDermott, 2000). Simple comparison bots are based on search technologies and should be implemented in terms of query and results that are consistent with good user interface design. These include providing the user with control entry of relevant search fields and shortcuts for comparison characteristics and sort order. For example, Best Book Buys, Amazon, and Barnes & Noble all provide the consumer with the opportunity identify the lookup field (Title, Keyword, ISBN, or Author) to limit the potential for users to return an unmanageable number of search results (Rowley, 2000). Further, Best Book Buys provides users a singlestep process for comparing books on price, delivery time, and condition (new or used) from a set of online book retailers. It is also important to provide consumers with price, shipping costs, total charges, a rating of the vendor, condition of the item if it is used, shipping method, expected arrival time, and payment options (Rowley, 2000).
32. E-COMMERCE AND CONSUMER BEHAVIOR
Target-focused agents make recommendations to consumers based on the target of their current search or purchase. For example, when a consumer accesses Amazon.com, a list of books that are consistent with that customer's previous purchasing experiences are displayed. In a study of context, item-specific information, and familiarity of intelligent agent recommendations for music CDs, participants were willing to pay nearly twice as much for an unfamiliar CD when the agent provided a recommendation in the absence of item-specific information about the unfamiliar option than when the unfamiliar option was presented in the context of a set of familiar options with distinguishing characteristics (Cooke, Sujan, Sujan, & Weitz, 2002). In the latter condition, participants evaluated these unfamiliar options negatively on the distinguishing characteristics. There is some indication that users do not give equal weight to agent recommendations, and they are more likely to assign a higher weight to extreme positive recommendations as opposed to extreme negative recommendations (Gershoff, Mukherjee, & Mukhopadhyay 2003). Based on these findings, researchers (Cooke et al., 2002) tentatively suggest alternative design recommendations for agents. The agent should make recommendations in context when (a) the recommendation is known to be preferred by the consumer through extensive purchase history, (b) the recommendation can be presented on meaningful item-specific characteristics that are similar to familiar options, and (c) absent of distinguishing characteristics from familiar options (Cooke at al., 2002). Agents should present recommendations singularly when (a) little is known about the consumer, (b) distinguishing characteristics are available for the consumer to form evaluative contrasts, and (c) the agent can evaluate the item extremely positively (Cooke et al., 2002; Gershoff et al., 2003). The role of word-of-mouth information in the information search process is filtered by the users' prior knowledge. That is, if online consumers base their choice of a vendor on familiarity (as opposed to price), they are less likely to engage in an extensive search for reviews of vendor performance and less susceptible to negative word-of-mouth information obtained on the Internet (Chaterjee, 2001). In addition, familiarity-based shoppers are also more likely to attribute negative information to unstable factors (Chaterjee, 2001). Familiarity with the vendor brand reduces consumers' security risk of doing business with the online vendor. Also, Chatterjee (2001) found that consumers who based their choice of vendor on price were more likely to search for word-of-mouth evaluations of the vendor performance and heed that negative information. As a rule, customers prefer to have all the information associated with any online purchase, including shipping, transaction costs, and taxes. In a study of consumers' expectations of online information for financial products, providing the monetary details associated with transactions (costs associated with an ongoing relationship) and marketing offers, and easy access to that information were found to be the most important factors in provider selection (Waite & Harrison, 2002). These findings suggest that reducing the ambiguity of the financial impact of the customervendor relationship is important to consumers' satisfaction with
*Data were collected in Taiwan.
607
online information. Of course, reducing ambiguity in regard to costs of a relationship may not result in increased sales; it merely reduces the uncertainty with regard to the actual costs of the options. Requiring users' registration during the information search stage has been clearly demonstrated to contribute to consumers' dissatisfaction and rated it as one of the least important attributes of information resources (Chen & Wells, 2001*; Waite & Harrison, 2002). At the same time, if a site has an established relationship with a customer, Ha (2002) found that a customized informational source reduced the antecedents of risk, including privacy, evaluation, and usability (nuisance, financial, performance, and time). This apparently contradictory evidence illustrates the notion that the types of features offered to users depend on the solution, user task, and the customers' desired relationship with the vendor (Chen & Wells, 2001; Ha, 2002; Waite & Harrison, 2002). That is, requiring a user to register early in the relationship for unclear and poorly articulated benefits is presumptuous, but after a trusting relationship is established, registration and customization can create a desirable user experience in terms of efficiency (e.g., one-click ordering) and personalization. Requiring users to register during the information search process will result in the majority of the information search traffic to seek information from another source unless a prior trusted relationship has been established. Of course, this could have some advantages from the provider's perspective in terms of managing bandwidth or controlling access to only the most motivated users. In the case of ongoing relationships, users are more than willing to register to gain access to a customized user experience. When given the choice between text only, graphics only, or a combination of both, users prefer a combination of text and graphics for the display of product information on the Web (Lightner & Eastman, 2002). This may be rooted in the relative importance of accurate product representation to online consumers (Reibstein, 2001). More generally, the design of the site must promote customer centric values such as simplicity, value, trust, fulfillment, and support. Search and Product Class. Girard et al. (2002) reported that product classes contribute to online consumer purchase preferences. Products were categorized into search, experience, or credence. A search product can be described as an item the relevant product attributes of which can be reliably assessed prior to purchase. Experience products are broken down into "... experience durable (low frequency of purchase goods) and experience nondurable (high frequency of purchase goods)" (Girard et al., 2002, p. 3). Intuitively, the product attributes of experience goods cannot be reliably assessed prior to purchase or the costs associated with information search regarding the experience are higher than actual product experience. A credence product is even more unknown to the user, "such that the average consumer can never verify the level of quality of an attribute possessed by a brand or even their level of need for the quality supplied by the brand. That is, consumers will
608
VOLK AND KRAFT
have great difficulty in evaluating the quality level of a product such as vitamins with confidence, or similarly a service such as termite fumigation or surgery" (Girard et al., 2002, p. 3). Girard and his colleagues (2002) found that consumers were more willing to purchase search products than low-frequency durable products. The experience nondurables were least likely to be purchased online. In a second study of product class, Lowengart and Tractinsky (2001) examined the underlying dimensions related to purchasing books and computers. Interestingly, Lowengart and Tractinsky (2001) characterized a computer as experiential product in contrast to Girard and his colleagues' (2002) characterization of a computer as a search product. Although the methodology for inclusion of the computer in the search category in the first study is documented and robust, Lowengart and Tractinsky (2001) found that the underlying dimensions for the purchase of these two items were different in terms of generalized risk and financial risk. Although these findings are preliminary and the products chosen for the categories were limited, this conceptual framework may begin to provide us with a useful way for characterizing and prioritizing design requirements. Considering these two studies, it is apparent that product class is a relevant issue when devising an online strategy for a particular product. At the same time, these studies were focused singularly on online purchase, and it is a reasonable strategy to expect that consumers may in fact use the Internet to engage in prepurchase information search and alternative evaluation or postpurchase customer support or evaluation activities in the future. The Internet as a medium for consumer activity is by no means static, and we can expect providers and consumers to use the technology in unforeseen ways.
Alternative Evaluation. As search proceeds, beliefs about alternative solutions (products or services) are developed and evaluative criteria identified. Researchers believe that consumers are typically unable to collect and evaluate all relevant information and do not have the mental processing capacity to handle all that is available. Further, they may be attempting to meet multiple goals with a particular purchase. As a result, consumers may construct new ad hoc evaluation approaches appropriate in the context of the decision (Arnould, Price, & Zinkhan, 2002). If experienced enough, they may recall and apply previously successful evaluation methods as they go through the decision process. The evaluation methods used are usually classified as compensatory, noncompensatory, or simple choice heuristics. Two thirds of online consumers use the Web as a method to comparison shop (McDermott, 2000). Online consumers can access discussion groups, user evaluations of products and vendors, and/or any number of online comparison sites (Scheepers, 2001). It is important to note that the format of the information that is acquired may affect the resulting evaluation (Blackwell, Miniard, & Engel, 2001). This should be of special concern to Web page designers who may alter the outcomes of evaluation through the method of information presentation. When users have relatively little experience with a target and the purchase task calls for a quality by price trade-off, there is some indication
that a user is more likely to select the middle option relative to price and quality (Prelec, Wernerfelt, & Zettelmeyer, 1997). The implications are certainly evident from both design and business perspectives. For example, in the case where the e-commerce site provides users with recommendations for purchase, this research suggests that a business strategy that displays purchase options in a manner that would highlight those items that are in stock or provide the business with the best margins could be employed through the implementation of display rules with regard to product options for comparison. From a marketing perspective, this also provides some guidance with regard to desirable product placement when negotiating with third-party vendors (Table 32.6). When consumers are aware of a lenient return policy, they infer higher product quality (Wood, 2001). Inferences of quality by brand play a larger role in some product categories rather than others (Kim & Pysarchik, 2000). This suggests that highlighting brand strength in the presentation and evaluation of technology products will result in the positive evaluation of the better brand. There is also evidence that these positive brand evaluations are related to consumers' intention to purchase these brands, and that rinding endures across cultures (Kim & Pysarchik, 2000). Therefore, multinational retailers can expect positive brand image to be more important than the product's country of origin in consumers' brand quality inferences. The Internet provides some unique opportunities with regard to dynamic pricing. Reverse pricing elicitation (name your own price) scenarios are less preferred than scenarios that provide the user with some referent prices, provided by the vendor or even in a reference range generated by the user (Chernev, 2003). Given that the vendor needs to account for the users' desire to have an available referent price range, it is suggested that designers of these interactions provide users with three referents: a referent range of successful similar bids (when the data exist), a user-generated referent range, or optional examples for view by the user. There are many other issues with regard to designing the online user experience that are directly related to the information search process and typically discussed in the context of user interface design in general. Readers are able to identify several chapters in this text that support the notion of providing the users with Web-based interfaces that are easy to use, have the necessary attributes that present a professional image, and offer users an opportunity to participate in their desired modes of interactive experience. Unless a product or service is so essential that consumers will overcome any difficulty to obtain it, a poorly designed Web site poses unnecessary costs on consumers that must be eliminated by competent Web site design.
Choice and Purchase Although consumers do not need to identify themselves during initial interactions with a Web site, they are forced to do so at the choice and purchase stages. At these stages, they must have developed a degree of commitment and trust (at least to the level of the price and importance of the product) associated with as many as five different business entities, including the
32. E-COMMERCE AND CONSUMER BEHAVIOR
credit card company, the product maker, the product seller, the delivery company, and the technical support company (Novak & Hoffman, 1998). With each additional middle entity, the possibility of a breakdown in one of the systems prior to product delivery increases cumulatively. Following the evaluation, when a satisfactory solution to the problem has been identified, the consumer develops an intention to make the purchase. Solution providers must take care to prevent inhibiting factors, such as cumbersome ordering procedures or fear of deception, from deterring the purchase. Shopping cart design, shipping choices, guarantees, and so on, are critical for encouraging the efficient and satisfying consummation of the transaction. Separating the costs of the product and fulfillment, such as shipping, handling, and taxes on different Web pages, increases concern that the vendor is taking advantage of the consumer with potentially hidden charges. Design considerations include enhancing efficiency of the ordering and registration tasks, while reducing the perceived risk of transaction with regard to security of personal and financial information, and establishing a fulfillment relationship with customers (Scheepers, 2001).
Postpurchase Phenomena Purchase and consumption are not the conclusions of the CDP. Several other important phenomena occur following purchasing decisions that influence the long-term effectiveness of marketing efforts. Consumers learn from their purchase experiences and habituate their decision making. Consumers also evaluate the results of their decision relative to their prepurchase expectations. A negative disconfirmation of their expectations results in dissatisfaction, negative attitude change, and a possible reassessment of their search and evaluation processes. In some cases, complaining or negative word-of-mouth activities may occur. Howard (1977) described the development of efficient consumer purchasing processes. When unfamiliar with a product category, consumers engage in extensive information search, concept formation, evaluation, and deliberation. As they gain knowledge, they move from extensive problem solving toward much simpler decision processes that draw on accumulated knowledge of choice criteria, product attributes, brands, and brand availability. When buying behavior has become thoroughly routine or habitual, the consumer does little more than a scan of memory to retrieve appropriate purchasing scripts and form buying intentions. When decision making for high involvement purchases has become habituated to the extent that external search is unnecessary, no new evaluation of alternatives is required, scripted buying intentions exist, and brand loyalty is said to have been established (Engel & Blackwell, 1982). The choice of a particular Web site such as Amazon.com for a consumer's book buying is itself consumer decision. When consumers have found particular Web sites to be productive sources of information and useful as a shopping tool, they may become brand loyal to the site and "bookmark" them for future use, thus embedding them in their behavioral scripts. This makes the selection of a Web site part of a consumer's routine
•
609
decision making, and continual reinforcement of this behavior by maintaining site quality is required to ensure the consumer's satisfaction with the site. In general, consumer research findings point to the conclusion that consistent quality and brand advertising are needed to keep consumers loyal to a brand or service (Mowen & Minor, 1998) The fulfillment of consumer expectations has been widely recognized as the basis for consumer satisfaction or dissatisfaction (Woodruff, Cadotte, & Jenkins, 1983). If businesses fail to deliver on promises, expectations are not met and dissatisfaction results (Olshavsky & Miller, 1972). For example, in December 1999, eToys.com's site was inaccessible to many would-be buyers because of high traffic. For those consumers who were able to place orders, 10% of the orders failed to arrive in time for Christmas (Stankevich, 2000). Undoubtedly, negative word of mouth and customer refusal to give the firm a second chance were partially responsible for the subsequent demise of the business. In fact, in the majority of instances of dissatisfaction, consumers do not complain to the business involved but rather discontinue patronage and/or spread negative word-of-mouth information about the business (Oliver, 1997). The well-known phenomenon of "buyer's remorse" or cognitive dissonance is another important postpurchase phenomenon. Dissonance is an unpleasant emotional state caused by conflicting perceptions such as a consumer's realization that the best purchase was not made and that an attractive alternative choice was forgone (Mowen & Minor, 1998). Dissonance may be relieved by consumers changing their attitude toward a chosen alternative, by undoing the purchase and returning the product, or by deciding never to make such a purchase again. Web sites that provide supportive information to consumers following purchase can help buyers to reduce their dissonance and can reduce the possibility of negative behavior or attitude change on the part of the buyer. Some marketers are not aware of the benefits of analyzing postpurchase consumer behavior, although many have become concerned with measuring consumer satisfaction (Mowen & Minor, 1998). The consumer behavior literature suggests that postpurchase phenomena are of great importance and that postpurchase contact by Web-based sellers, as well as brick-and-mortar retailers, is important in reducing the effects of dissonance, as well as assessing and influencing other postpurchase outcomes, such as satisfaction and repatronage behavior.
CONCLUSION This brief summary of consumer behavior research is by no means a complete work with regard to the range of theoretical works that could be used in the development of online consumer experiences. Instead, it is a sampling of applied and theoretical works that can be used to establish either general, or in some cases very specific, design guidelines. Regardless of the foundation of any single consumer e-commerce solution, those of us who create these online experiences must consider the unique characteristics of our target users, the influence of product class, and how our solution integrates with the broad range of consumer behavior both online and off-line.
610
VOLK AND KRAFT
References Adaval, R., & Monroe, K. B. (2002). Automatic construction and use of contextual information for product and price evaluations. Journal of Consumer Research, 28, 572-292. Alba, J., & Hutchinson, W (2000). Knowledge calibration: What consumers know and what they think they know. Journal of Consumer Research, 27, 324-344. Antil, J. H. (1984). Conceptualization and operationalization of involvement. In T. C. Kinnear (Ed.), Advances in consumer research (Vol. 11, pp. 203-209). Provo, UT: Association for Consumer Research. Arnould, E., Price, L., & Zinkhan, G. M. (2000). "Consumers," Chicago: Irwin/McGraw Hill, ISBN: 0-256-13360-3. Baumgartner, H. (2002). Toward a personology of the consumer. Journal of Consumer Research, 29, 286-292. Beatty, S., & Smith, S. (1987). External search effort: An investigation across several product categories. Journal of Consumer Research, 14, 84-92. Belk, R. W, Wallendorf, M., & Sherry, J. F. (1989). The sacred and profane in consumer behavior: Theodicy on the odyssey. Journal of Consumer Research, 16, 1-38. Bhatnagar, A., Misra, S., & Rao, H. R. (2000). On risk, convenience, and Internet shopping behavior, Communications of the ACM, 43(11), 98-105. Blackwell, R. D., Miniard, E W., & Engel J. F. (2001). Consumer behavior (9th ed.). Fort Worth, TX: Harcourt Brace Publishers. Bloch, P. H. (1982). Involvement beyond the purchase process. In A. Mitchell (Ed.), Advances in consumer research (Vol. 9, pp. 413417). Ann Arbor, MI: Association for Consumer Research. Bloch, P. H., Sherrell, D. L., & Ridgeway, N. M. (1986). Consumer search. Journal of Consumer Research, 13, 119-126. Campbell, M. C., & Goodstein, R. C. (2001). The moderating effect of perceived risk on consumers' evaluations of product incongruity: Preference for the norm. Journal of Consumer Research, 28, 439449. Celsi, R. L., & Olson, J. C. (1988). The role of involvement in attention and comprehension processes. Journal of Consumer Research, 15, 210-224. Chatterjee, P. (2001). Online reviews: Do consumers use them? Advances in Consumer Research, 28, 129-133. Chen, Q., & Wells, W D. (2001). .com Satisfaction and .com dissatisfaction: One or two constructs? Advances in Consumer Research, 28, 34-39. Chernev, A. (2003). Reverse pricing and online price elicitation strategies in consumer choice. Journal of Consumer Psychology, 13,5162. Cooke, A. D., Sujan, H., Sujan, M., & Weitz, B. A. (2002). Marketing the unfamiliar: The role of context and item-specific information in electronic agent recommendations.Journal of Marketing Research, 39, 488-497. Dholakia, U., & Bagozzi, R. (2001). Consumer behavior in digital environments. In J. Wind & V Mahajan (Eds.), Digital marketing (pp. 163-200). New York: Wiley. Dholakia, R. R., & Chiang, K. (2003). Shoppers in cyberspace: Are they from Venus or Mars and does it matter? Journal of Consumer Psychology, 13,171-176. Dowling, G. R., & Staelin, R. (1994). A model of perceived risk and intended risk-handling activity. Journal of Consumer Research, 21, 119-134. Engel, J. E, & Blackwell, R. D. (1982). Consumer behavior. Chicago: The Dryden Press.
Fogg, B. J., Marshall, J., Laraki, O., Osipovich, A., Varma, C., Fang, N., Paul, J., Rangnekar, A., Shon, J., Swani, P., &Treinen, M. (2001). What makes a Web site credible? A report on a large quantitative study. Proceedings of ACM CHI 2001 Conference on Human Factors in Computing Systems (pp. 61-68). New York: ACM Press. Available: http://www.acm.org/pubs/articles/proceedings/chi/365024/p6lfogg/p6l-fogg.pdf Fogg, B. J., Soohoo, C., Danielsen, D., Marable, L., Stanford, J., & Tauber, E. (2002). How do people evaluate a Web site's credibility? Results from a large study. Stanford Persuasive Technology Lab, Stanford University, Stanford, CA. Available: http:// www.consumerwebwatch.org/news/report3-credibilityresearch/ stanfordPTL_abstract.htm or http://credibility.stanford.edu/ mostcredible.html Friedman, B., Kahn, P. H., & Howe, D. C. (2000). Trust online. Communications of the ACM, 43(12), 34-40. Gefen, D. (2002). Reflections on the dimensions of trust and trustworthiness among online consumers. The DATA BASE for Advances in Information Systems, 33, 38-53. Gershoff, A. D., Mukherjee, A., & Mukhopadhyay, A. (February, 2003). Consumer acceptance of online agent advice: extremity and positivity effects. Presented at the Annual Conference of the Society for Consumer Psychology, New Orleans LA. Girard, T, Silverblatt, R., & Korgaonkar, P. (2002). Influence of product class on preference for shopping on the Internet. Journal of Computer-Mediated Communication 8(1) [On-line]. Available: http://www.ascusc.org/jcmc/vol8/issuel/girard.html Grewal, D., Gottlieb, J., & Marmostein, H. (1994). The moderating effects of framing and source credibility on the price-perceived risk relationship. Journal of Consumer Research, 21, 145-153Grunert, K. (1996). Automatic and strategic processes in advertising effects. Journal of Marketing, 60, 88-101. Ha, H. Y. (2002). The effects of consumer risk perception on prepurchase information in online auctions: Brand, word-of-mouth, and customized information. Journal of Computer-Mediated Communication [On-line], 38(1). Available: http://www.ascusc.org/ jcmc/vo18/issue1/ha.html Hirschman, E. C. (1980). Innovativeness, novelty seeking, and consumer creativity. Journal of Consumer Research, 7, 283-295. Hirschman, E. C., & Holbrook, M. B. (1982). Hedonic consumption: Emerging concepts, methods and propositions. Journal of Marketing, 46, 92-101. Hoch, S. J. (2002). Product experience is seductive. Journal of Consumer Research, 29, 448-454. Hofacker, C. E, & Murphy, J. (1998). World Wide Web banner advertising copy testing. European Journal of Marketing, 32, 703-712. Holbrook, M., & Hirschman, E. (1982). The experiential aspect of consumption: consumer fantasies, feelings, and fun. Journal of Consumer Research, 9, 132-140. Howard, J. A. (1977). Consumer behavior: Application of theory. New York: McGraw-Hill. Insko, C. A., & Schopler, J. (1972). Experimental social psychology. New York: Academic Press. Jarvenpaa, S. L., & Tractinsky, N. (1999). Consumer trust in an Internet store: A cross-cultural validation./owraa/ of Computer-Mediated Communication 5(2) [On-line]. Available: http://www.ascusc.org/ jcmc/vol5/issue2/jarvenpaa.html Jarvenpaa, S. L., Tractinsky, N., & Vitale, M. (2000). Consumer Trust in an Internet Store. Information Technology and Management, 1(1-2): 45-71.
32. E-COMMERCE AND CONSUMER BEHAVIOR
Kamali, N., & Loker, S. (2002). Mass customization: On-line consumer involvement in product design. Journal of Computer-Mediated Communication 7(4) [On-line]. Available: http://www.ascusc.org/ jcmc/vol7/issue4/loker.html Kim, D. J., Cho, B., & Rao, H. R. (November 5-8, 2000). Who buys on the Internet? An investigation of consumer characteristics for internet purchasing behavior. INFORMS Conference on Information Systems and Technology, San Antonio, Texas. Kim, S., & Pysarchik, D. T. (2000). Predicting purchase intentions for uninational and binational products. International Journal of Retail & Distribution Management, 28, 280-291. Krugman, H. E. (1965). The impact of television advertising: Learning without involvement. Public Opinion Quarterly, 29, 349-356. Kwak, H., Fox, R. J., & Zinkhan, G. M. (2002). What products can be successfully promoted and sold via the Internet? Journal of Advertising Research, 42, 23-38. Lastovicka, J. L. (1979). Questioning the Concept of Involvement Denned Product Classes. In W L. Wilkie (Eds.), Advances in consumer research (Vol. 6, pp. 174-179). Ann Arbor, MI: Association for Consumer Research. Lavidge, R. J., & Steiner, G. A. (1961). A model for predictive measurements of advertising effectiveness. Journal of Marketing, 25, 5962. Lee , J., Kim, J., & Moon, J. (2000). What makes Internet users visit cyber stores again? key design factors for customer loyalty. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, 305-312. Li, H., Daugherty, T., & Biocca, F. (2002). Impact on 3-D advertising on product knowledge, brand attitude, and purchase intention: The mediating role of presence. Journal of Advertising, 31(3), 4357. Li, H., Kuo, C., & Russell, M. G. (1999). The impact of perceived channel utilities, shopping orientations, and demographics on the consumer's online buying behavior. Journal for Computer-Mediated Communication, 5, 1-23 [On-line]. Available: http://www.ascusc. org/jcmc/vol5/issue2/hairong.html Lightner, N. J. (2003). What users want in e-commerce design: Effects of age, education and income. Ergonomics, 46, 153-168. Lightner, N. J., & Eastman, C. (2002). User preference for product information in remote purchase environments. Journal of Electronic Commerce Research, 3, 174-186. Logan, G. D. (1988). Towards an instance theory of automatization. Psychological Review, 22, 1-35. Logan, G. D. (1992). Attention and preattention in theories of automaticity. American Journal of Psychology, 105, 317-339. Lohse, G. L., & Spiller, P. (1999). Internet retail store design: How the user interface influences traffic and sales. Journal of ComputerMediated Communication 5(2) [On-line]. Available: http://www. ascusc.org/jcmc/vol5/issue2/lohse.htm Lowengart, O., & Tractinsky, N. (2001). Differential effects of product category on shoppers' selection of Web-based stores: A probabilistic modeling approach. Journal of Electronic Commerce Research, 2, 142-156. Mandel, N., & Johnson, E. J. (2002). When Web pages influence choice: Effects of visual primes on experts and novices. Journal of Consumer Research, 29, 235-245. Markin, R. J., Jr. (1974). Consumer behavior: A cognitive orientation. New York: Macmillan. McDermott, I. (2000). Shopping bots: Santa's electronic elves. Searcher: The Magazine for Database Professionals, September, 10-16. Menon, S., & Soman, D. (2002). Managing the power and curiosity of effective Web advertising strategies./oz/raa/ of Advertising Research, 31, 1-14.
611
Miyazaki, A. D., & Fernandez, A. (2001). Consumer perceptions of privacy and security risks for online shopping. The Journal of Consumer Affairs, 35, 27-44. Moe, W. W. (2003). Buying, searching, or browsing: Differentiating between online shoppers using in-store navigational clickstream./owrnal of Consumer Psychology, 13, 29-39Mowen,J. C., & Minor, M. (1998). Consumer behavior (3rd ed.). Upper Saddle River, NJ: Prentice Hall. Netguide (1996). Just say no to graphics. Netguide, July, 32. Novak, T. P, & Hoffman, D. L. (1998). Building consumer trust in online environments: the case for information privacy. Communications of the ACM, 47(12), 41-49. Oliver, R. L. (1997). Satisfaction: A behavioral perspective on the consumer. New York: McGraw-Hill. Olshavsky, R. W, & Miller, J. A. (1972). Consumer expectations, product performance, and perceived product quality. Journal of Marketing Research, 9, 19-21. Olson, J., & Olson, G. M. (2000). i2i Trust in e-commerce. Communications of the ACM, 43(12), 41-44. Pavlou, P. A. (2001). Consumer intentions to adopt electronic commerce: Incorporating trust and risk in the technology acceptance model. DIGIT Workshop. Peter, J. P., & Olson, J. C. (2002). Consumer behavior and marketing strategy. (6th ed.). New York: McGraw-Hill/Irwin. Petty, R., Cacioppo, J., & Schumann, D. (1983). Central and peripheral routes to advertising effectiveness. Journal of Consumer Research, 10, 135-146. Prelec, D., Wernerfelt, B., & Zettelmeyer, E (1997). The role of inference in context effects: Inferring what you want from what is available. Journal of Consumer Research, 24, 118-125. Ray, M. (1973). Marketing communications and the hierarchy of effects. In P. Clarke (Ed.), New models for mass communications (pp. 147176). Beverly Hills, CA: Sage. Reibstein, D. J. (2001). The Internet buyer. In J. Wind & V Mahajan (Eds.), Digital marketing: Global strategies from the world's leading experts (pp. 201-225). New York: Wiley. Rodgers, S., & Sheldon, K. M. (2002). An improved way to characterize Internet users. Journal of Advertising Research, 1, 85-94. Rosa, J. A., & Malter, A. J. (2003). E(embodied>knowledge and e-commerce: How physiological factors affect on-line sales of experiential products. Journal of Consumer Psychology, 13, 6373Rowley, J. (2000). Product searching with shopping bots. Internet Research, 10, 203-214. Schaninger, C. M. (1976). Perceived risk and personality. Journal of Consumer Research, 3, 95-100. Scheepers, R. (2001). Supporting the online consumer decision process: electronic commerce in a small Australian retailer. Proceedings of the Twelfth Australasian Conference on Information Systems. Sherif, C., & Sherif, M. (1967). Attitudes as the individuals' own categories: The social judgment-involvement approach to attitude and attitude change. In C. W. Sherif & M. Sherif (Eds.), Attitude, ego-involvement and change (pp. 105-139). New York: Wiley. Stankevich, D. G. (2000). Was the grinch really online? Discount Merchandiser, 3, 42-44. Steinneld, C., & Whitten, P. (1999). Community level socio-economic impacts of electronic commerce. Journal of Computer-Mediated Communication 5(2) [On-line]. Available: http://www.ascusc.org/ jcmc/vol5/issue2/steinneld.html Stigler, G. (1961). The economics of information. Journal of Political Economy, 69, 213-225.
612
•
VOLK AND KRAFT
Tauber, E. M. (1972). Why do people shop? Journal of Marketing, 36, 46-59. Waite, K., & Harrison, T. (2002). Consumer expectations of online information provided by bank sites. Journal of Financial Services Marketing, 6, 309-322. Wang, H., Lee, M., & Wang, C. (1998). Consumer privacy concerns about Internet marketing. Communications of the ACM, 41(3), 6370.
Weiss, M. J. (2001). Online America. American Demographics, 22, 5360. Wood, S. L. (2001). Remote purchase environments: The influence of return policy leniency on two-stage decision processes. Journal of Marketing Research, 38, 157-169. Woodruff, R. B., Cadotte, E. R., & Jenkins, R. L. (1983). Modeling consumer satisfaction processes using experience based norms, Journal of Marketing Research, 20, 296-304.
33 WEB SECURITY AND PRIVACY E. Eugene Schultz University of California-Lawrence Berkeley National Laboratory
recommendations for improving usability in Web security and privacy.
INTRODUCTION Information security (often also somewhat less appropriately called "computer security") involves protecting the confidentiality of data stored in computers and transmitted over networks, integrity of data, applications, systems, and network devices, and accessibility of data, applications, databases, systems, network services, and so forth (Bernstein, Bhimani, Schultz, & Siegel, 1995). Information security professionals have also become increasingly interested in the goal of nonrepudiation or nondeniability, which means preventing individuals who have initiated electronic transactions from denying that they have done so, and auditability, which means ensuring each user's actions are recorded so all users can be held accountable for their actions. "Privacy" refers specifically to protecting the confidentiality of data referring to individuals, such as employment history, medical information, and credit card numbers. Over the past decades, information security has grown from a small area to a major one within computer science and information technology. Computer and networking technology has expanded to the point that organizations use and depend on computers and networks, that what is stored on and processed by computers and then transmitted over networks can literally be millions upon millions of dollars, as evidenced by empirical surveys. For example, a more recent survey of 223 information practitioners found that reported security-related losses by the respondents' organizations during 2001 totaled more than $455 million (FBI, 2002). This chapter explains the basics of information security—particularly how it is applied to Web security and privacy, describes how Web security and privacy are most often breached, presents an analysis of the relationship between human factors issues and Web security and privacy, and offers
TYPES OF SECURITY-RELATED INCIDENTS A security incident is one in which an adverse outcome due to a breach in a security mechanism has occurred. One of the best-known types of security incidents is unauthorized access to systems (commonly known as a "hacker attack") in which an attacker guesses or cracks the password for one or more accounts on a system or exploits a vulnerability* in a program to gain access. Another common type of incident is a denial-of-service (DoS) attack in which the perpetrator causes a computer or application to crash or causes a computer, application, or network to slow down. In still another, the integrity of a system or data stored on the system or an application is changed without authorization. Web defacements, unauthorized alteration of the content of one more pages on a Web server, are a very common kind of integrity violation attack. Hoaxes, false information about incidents, vulnerabilities in systems, malicious code such as viruses (self-reproducing programs that spread because of actions of users) and worms (self-reproducing programs that spread independently of users), and scams attempts to financially profit by using e-mail or Web sites to convey bogus information (often in the form of some kind of investment opportunity) are other kinds of incidents, as are extortion plots and electronic harassment activity. Given the sophistication of today's computers, networks, and applications; the number of vulnerabilities that can be exploited in them; the range of easyto-use hacker tools freely available on the Internet; and the vast numbers of people (a nontrivial portion of whom are, unfortunately, unscrupulous) who use computers, the possibilities for
*A vulnerability is a flaw in a program that allows someone to exploit it by bypassing one or more security mechanisms.
613
614
•
SCHULTZ
costly and disruptive security breaches have become virtually limitless. WHY PROTECTING WEB SERVERS IS SO DIFFICULT To say that protecting computer systems, networks, data, and applications from attacks is a challenging undertaking that requires considerable planning and resources is an understatement. However, protecting Web servers is in many respects the most difficult of all. Among the most fundamental challenges in keeping Web servers secure is the fact that both the existence and location of most Web servers are deliberately advertised to the entire Internet user community. Most Web servers allow anonymous access, that is, access without any requirement for individuals to establish their identities. Only a few require subscription accounts, accounts that require each user to authenticate Web access by entering a password that is unique to that user's account. Authentication means providing the required level of assurance that a user is who that person claims to be for the purpose of accessing a system, network, and/or application. Although the identities of subscription Web servers can be guarded closely if desired, the fact that anonymous Web servers must be widely known if they are to fulfill their purpose has the unfortunate side effect of alerting a wide range of potential attackers around the world to another potential target. In addition, many automated ways of attacking frequently used Web services and protocols now exist. Many of these methods are embodied in attack tools that often require little more than entering the Internet Protocol (IP) address of the to-be-attacked Web server and a choice concerning the type or scope of attack to launch. Furthermore, new Web attacks are constantly surfacing. The many vulnerabilities in Web servers and applications provide fertile ground for those determined to discover and exploit them. The complexity of Web servers provides yet another challenge in that it makes defending them more difficult. In addition, organizations frequently quickly roll out Web servers to "beat the competition" without considering security needs. The result is Web servers and applications that are too often wide open to attack. Finally, attacking Web servers is very popular within hacker circles right now. Web sites such as http://www.attribution.org post reports of successful Web attacks, thus giving recognition to Web hackers and motivating others to match their feats. THE BASICS OF WEB FUNCTIONALITY Understanding the basics of Web functionality is essential to understanding Web service and privacy concerns. The fundamental protocol to Web services is the hypertext transfer protocol (HTTP), a type of "command and control protocol" that manages all interaction between browsers (clients) and servers.
Interaction involves two fundamental types of transactions, GET (when a browser asks the server to retrieve some information and send it back) and POST, which enables the browser to send some information to the server. In reality, browsers have no control over when there will be a GET and when there will be a POST (Cox & Sheldon, 2000). Furthermore, Web communication is "stateless"; each interaction between browsers (clients) and servers is an independent transaction, meaning that if the browser sends a GET request to a server, the server itself retains no information about any previous GET's or POST'S that may have occurred. A Web server is an implementation of HTTP that allows Web applications to run. The stateless nature of Web interaction presents a problem for Web servers, which often need information from previous interactions to deal with current interactions. Web session management mechanisms have been developed as a solution. The server creates a session identifier early in the user interaction sequence and then transmits the ID to the browser, ensuring the browser will send the identical ID with every future request, thereby linking each transaction with unique session data. Of the ways of maintaining sessions in this manner, "cookies" have proven to not only be the most popular, but also the most secure. A cookie is an object that contains information that a Web server sends to a client for future use.* Cookies are most often used to hold information about user preferences, but they also can contain personal information such as credit card numbers, dates of birth, social security numbers, and so forth. A Web client (commonly known as a "Web browser" or simply as a "browser") is software used to access Web servers. Web clients are by design less sophisticated in functionality than are Web servers, although Web clients typically contain a great deal of the software involved in human-computer interaction. Web clients are a cause of considerable security-related concern because Web servers are in control of client-server interactions, frequently independently of user awareness and consent. For example, executable content languages (XCLs), also called "mobile code," are a class of executables that Web servers frequently download into browsers at certain points in client–server interactions. ActiveX, Java applets, JavaScript, and Visual Basic Script are four of the most common types of XCLs. Web browsers are generally passive while Web servers send XCLs to them, which afterward are executed within the browsers. Although some types of XCLs, particularly Java applets, have built-in security constraints to prevent them from doing something malicious (e.g., launching attacks against other systems on the network), most have few if any constraints. ActiveX, for example, does not limit how its code executes, potentially allowing an ActiveX control to access and start applications on the system that houses a browser, initiate network connections, and so forth. In the World Wide Web today, Web pages are usually built using the hyptertext markup language (HTML). HTML consists of tags (special notations that determine how browsers should display Web pages, including text centering, font selection,
*There are two basic types of cookies, "persistent" and "nonpersistent" cookies. Persistent cookies are usually written to the hard drive of the machine that runs a browser; as such, they are available the next time that machine boots. Nonpersistent cookies are written into the memory of the machine that runs a browser and are generally purged when that machine boots.
33. Web Security and Privacy locations of images, and so forth), code, fields, and other elements needed to organize the data within each Web page. Hidden fields within HTML pages (fields that are hidden from view) are designed primarily to obviate the need for users to reenter data on each form that is presented, but also to keep sensitive information such as users' social security numbers from users. Hiding content in this manner is at best a superficial control measure; most browsers offer options that allow any user to view hidden fields. The extensible markup language (XML) is another type of metalanguage in which Web developers or Web masters can build specialized messages that express sophisticated interactions between clients and services or between components of a service. In short, it enables people to construct their own language tailored to their individual requirements. Although HTML is currently used more than any other metalanguage, XML is the heir apparent because of its very sophisticated capabilities. Unfortunately, XML is also beset with security-related vulnerabilities that allow attackers to engage in unauthorized and malicious actions, such as creating bogus XML pages and pointing to them from a legitimate page (Ilioudis, Pangalos, & Vakali, 2001).
615
Denial-of-Service Attacks Another type of frequently occurring security-related Web incident is a DoS attack in which an attacker or malicious code crashes a Web server or a Web application, or causes the Web server to slow down so much that it seriously disrupts the functionality of the Web site. Although intensive intervention efforts can allow the victim Web site to be running once again in a short period of time, even a small amount of Web site downtime often translates to major business losses, operational disruptions, and embarrassment. Many organizations rely on Web servers as a means of allowing customers to order goods, make financial transactions, and so forth. In addition, if Web sites are down for even a few minutes, self-appointed vigilantes may notice the problem and make negative Web postings, inform the press, and so forth. Patching vulnerabilities that are discovered in Web servers and applications and screening out abnormal input to them that can result in exploitation of vulnerabilities are the best preventative measures for DoS threats.
Privacy Compromise COMMON TYPES OF WEB-RELATED INCIDENTS A wide range of choices regarding the particular Web server to be used—Apache, Internet Information Server (IIS), the Netscape Web Server, Wusage, Domino, Websphere, and many others—is available. Regardless of which is used, that Web server will by default, at least to some degree, be vulnerable to a variety of remote attacks, including Web page defacements, DoS attacks, a variety of privacy infringement attacks, exploitation of cross-scripting, vulnerabilities, session hijacking attacks, password attacks, buffer overflow attacks, and mobile code attacks, all of which are covered in this section.
Web Page Defacements Web page defacements, unauthorized alteration of the content of one or more pages on a Web server, are a very common kind of Web attack. As mentioned earlier, the fact that Web defacements are frequently reported to sites such as http://www.attrition.org not only makes these attacks quite visible, but also motivates attackers to perpetrate these types of attacks. The most common target is homepages, pages that users first see when they reach a Web site. Defacement styles vary considerably—some attackers completely overwrite the content of Web pages (sometimes with a message that attacks the organization that hosts the Web site or one of its employees as well as a range of new graphic images), whereas others simply add a few lines of text to prove that a certain attacker or hacking group did the defacement. One of the most serious downsides to Web defacements is that they can lead to unauthorized alteration of critical information, such as pricing information, leading to all kinds of complications (e.g., lawsuits over prices of commodities advertised via the Web and embarrassment). Web defacements also attract the attention of the press, causing the organization that has a defaced Web server embarrassment and/or public relations setbacks.
Still another type of Web security compromise involves infringement of privacy. Attackers may glean a wide range of information about individuals from poorly secured Web servers and applications. Types of personal information likely to be exposed include credit card numbers, addresses of residences, phone numbers, e-mail addresses, social security card numbers, user preferences, and much more. Privacy infringement can be accomplished in many ways. Virtually anyone may, for instance, read the information contained in other users' cookies, provided that the cookies are in cleartext. Attackers must simply invent a method to access and then copy the information in cookies, which is greatly facilitated by the existence of widely available hacker tools on the Internet. Many of these tools require virtually no skill or knowledge on the part of the attackers, who in many cases needs only to run the tool, enter the IP address of the system to be attacked, and then click on Start. Information in HTML or XML pages is likewise vulnerable to compromise. For example, virtually anyone who knows how to read data within hidden HTML fields can access and read these data. Weak passwords for Web access accounts can also lead to privacy compromises. Attackers are more likely to guess or crack weak passwords, enabling them to access Web accounts and to glean information accessible via these accounts. Furthermore, scripting languages such a Visual Basic may allow users to discover code paths associated with the selection of user options. Attackers can also construct XCLs that once downloaded into browsers can find files that hold personal information, read this information, and send it to an address of the attackers' choice. Finally, well-known vulnerabilities in Microsoft's Internet Explorer browser in Windows systems can allow an attacker to not only achieve unauthorized access to a user's cookies, but also to change values in these cookies without consent or knowledge on the user's part. The attacker needs only to construct a specially formed uniform resource locator (URL) to access other users' cookies in this manner.
616
SCHULTZ
Defending against privacy compromises is not easy because such a wide range of vulnerabilities that result in privacy compromises exists. Solutions such as encrypting all data transmitted between the Web server and the client; encrypting all cookies; avoiding using hidden HTML fields to store sensitive, personal information; using scripting languages that do not allow discovery of code paths; and limiting or stopping altogether the execution of XCLs are commonly used measures in helping defend against privacy compromises. Exploitation of vulnerabilities is not the only way that privacy can become compromised, however. Every cookie that a Web server creates is under the control of the server. As mentioned earlier, cookies contain a variety of information about Web users. Many organizations that host Web sites (and thus have control over Web servers) routinely gather the information contained in cookies for marketing and other purposes. The amount of information concerning individuals that some organizations have collected and are still collecting is alarming; the fact that, in many cases, the systems and databases that store this information are often vulnerable to unauthorized access in many ways only exacerbates the problem. Privacy legislation in many European countries helps protect against the practice of harvesting information contained in cookies, but most countries have no such legislation, and if they have legislation, it is too often inadequate in dealing with the problem (as in the case of the United States, which has a few privacy protection laws, but they do not cover important issues such as harvesting personal information from Web sites). An important caveat about the privacy-related concerns regarding cookies is appropriate at this point. Cookies are not all-powerful programs (they are not even programs, in fact!) that go out and gather all the information about individuals that they can. Cookies are really nothing more than repositories of information that users directly provide when they engage in tasks such as completing fields in Web forms. If users do not enter data such as their social security numbers at some point in interacting with Web servers, cookies will not hold this information. This, of course, suggests that users should be provided with prompts that warn them every time they enter private information in each field in a Web form, although the expediency of overwhelming users with such prompts while they are engaged in Web interaction tasks that require input of a great deal of personal information is at best questionable. In addition, cookies can be transmitted only to and from the Web server or domain (e.g., xyz.net) that actually created the cookie in the first place. This means that another Web server cannot simply request cookies from another to obtain cookies. Thus, although cookies represent a genuine threat to privacy, some inherent mechanisms help contain this threat.
tem that runs the browser. This class of vulnerabilities poses an unusually high level of security risk in that it allows unauthorized access to a such a wide range of user resources—cookies, files, applications, and so forth—thereby potentially exposing a considerable amount of personal information. Patching each cross-scripting vulnerability that surfaces is the proper antidote.
Session Hijacking In another kind of attack, a "session hijacking" attack, attackers monitor network traffic to steal session ID data. In Web interactions, every visit to a Web site results in the creation of a "session" that allows continuous exchange of data between the client and server. Session ID data (as the name implies) include information that Web servers need to uniquely identify and track each session in an attempt to deal with each client separately. If attackers obtain session ID data, however, they can create another Web server connection that goes to the same Web pages with the same access as a legitimate user, allowing them to take over a user's session to the Web server. This is also a very serious problem in that once the user's session has been stolen, an attacker can now do whatever the legitimate user can do. Avoiding the inclusion of session ID data in URLs is an effective way to avoid session hijacking.
Password Attacks Still another similar type of attack involves guessing passwords to accounts on subscription Web servers to gain the same access to the Web server that a legitimate user would obtain. Alternatively, an attacker can use a hacking tool, a "password cracker," to generate possible passwords and then compare them to entries in the password file of a system. As in the case of session hijacking, the attacker who gains access to an account through password guessing or a password cracking tool can do whatever the legitimate user of the account can. Creating an information security policy (a statement of an organization's information security requirements, particularly with respect to what does and does not constitute acceptable use) that requires users to create and use difficult-to-guess passwords is a good countermeasure for password attacks. Tools (called "password filters") that prevent users from entering easy-to-guess passwords are even more effective. One-time-passwords are still another good solution. One-time passwords are, as the name implies, valid for only one logon. Once a user enters a password and it is accepted by the system to which the user is trying to gain access, that password can never be used again. Thus, if an attacker cracks a password, the possibility that the password has already been used is very high if the password is a one-time password.
Exploitation of Cross-Scripting Vulnerabilities A certain class of vulnerabilities (called cross-scripting vulnerabilities) enables attackers to obtain an unauthorized connection to a user's browser via the Web server to which the user is connected. Once connected to the browser, attackers can potentially access a range of data and applications on the sys-
Buffer Overflow Attacks Another kind of Web attack is a "buffer overflow" attack in which an attacker sends an excessive amount of input to a Web server. If the Web server does not have sufficient memory to hold the
33. Web Security and Privacy input, the input can overflow the buffer, possibly causing commands that the attacker has inserted in the overflow portion of the input to be executed without authorization. Results can be catastrophic—in some cases, an attacker can capitalize upon a buffer overflow condition to gain complete control over a Web server. Another possible outcome is DoS in the form of the application or Web server crashing. The best solutions for buffer overflow attacks are to have application developers allocate (reserve) considerably more memory than appears necessary by using memory allocation commands that do this and also to have the Web server reject unusual input, such as an excessive number of characters.
Malicious Mobile Code Attacks The final kind of Web-related attack considered here is an attack in which malicious mobile code (or an XCL) is downloaded into a Web browser. Once the downloaded mobile code executes, a variety of undesirable outcomes is possible. Some types of malicious mobile code, for instance, glean information about users from "Java wallets," objects used to hold information such as credit card and social security numbers to make electronic business transactions easier for users. Although obviating the need for users to enter this information during transactions is advantageous from a human factors perspective, the fact that carefully constructed mobile code can glean this information is highly undesirable from a security perspective. Other types of malicious mobile code open one window after another until the system that houses a Web browser crashes, produce annoying sound effects, initiate long-distance phone calls using a modem within the system that houses a Web browser, and so forth. As mentioned earlier, the fact that the Web browser cannot usually control whether XCLs will be downloaded makes this problem potentially very serious.
HUMAN FACTORS ISSUES IN WEB SECURITY AND PRIVACY The almost universal phenomenon of user resistance to securityrelated tasks such as authentication (proving one's identity such as through entering a password for the purpose of access), setting file permissions, and inspecting system logs for signs of misuse suggests that human factors issues play a large role in the ability to perform tasks that improve security and privacy. Systems with inadequate usability design are apt to cause more user resistance than systems with adequate usability design (AlGhatani & King, 1999). Security control measures may introduce additional usability barriers for system administrators and users, barriers that result in systems that need to be patched and configured for better security being left in an insecure, vulnerable condition—an easy target for attackers. The study of the relationship between human factors and information security itself is, however, in its infancy. Proctor, Lien, Salvendy, and Schultz (2000) showed how task analysis could be applied to different user authentication tasks to obtain an estimate of the difficulty
617
of performing each task. Schultz, Proctor, Lien, and Salvendy (2001) developed a taxonomy of security-related tasks that included a delineation of the major usability hurdles for each type of task. Whitten and Tygar (1999) pointed out usability hurdles in using a popular encryption program, PGP (pretty good privacy). Unfortunately, the few studies and analysis papers that have been published to date do not go very far in addressing the many unsolved issues concerning the relationship between human factors and effectiveness of information security measures. The relationship between human factors and Web security is even less understood. Proctor, Vu, Salvendy, et al. (2002) identified factors that affect the effectiveness of Web content, such as the information that needs to be extracted, how that information needs to be stored and retrieved, and how information should be presented to users. Accessing and using Web sites and securing them are, however, for the most part two completely different types of tasks. Securing Web sites requires rather complex knowledge and skills not required of Web users, as discussed shortly. Web security requires securing Web servers, Web applications, ensuring privacy, securing data sent over networks, securing the operating systems that host Web servers, and ensuring that meaningful options are available to users via interaction with browsers. Human factors issues and challenges in each of these areas are now discussed.
Human Factors Challenges in Securing Web Servers Individuals who set up a Web site seldom engage in all the effort necessary to build a custom Web server "from scratch." Developing a Web server in this manner requires considerable knowledge and effort (and thus ultimately entails considerable delay and expense for the organization that owns the Web site for which the Web server will be deployed). Instead, individuals who need to implement a Web site typically choose from preexisting Web servers, such as the ones mentioned earlier, customizing them to meet their particular needs. Of all the Web servers today, the one that requires by far the least effort to create is the IIS Web server. This Web server is bundled (included) with and enabled by default in Windows 2000 Server and Windows 2000 Advanced Server, two widely used operating systems; it can easily be enabled in other operating systems such as Windows 2000 Professional and Windows XP. From a human factors perspective, IIS might superficially seem almost ideal. Little effort is required to create a Web site using this server—the directories, executables, accounts, and so forth necessary for creating a Web site are built in, requiring almost no human intervention. Default parameters are provided en masse, relieving Web masters of the need to consult manuals and help pages to determine which values are appropriate. To build a minimal Web site, one must create HTML or other pages, select one as the homepage (default), and link the other pages in the desired order. Furthermore, additional related utilities such as Active Server and Front Page eliminate large parts of many task sequences in implementing Web pages as well as in modifying them and maintaining a Web site.
618
•
SCHULTZ
Unfortunately, IIS is also by default the most vulnerable of all Web servers. Although Apache is by far the most prevalent Web server deployed worldwide today, according to attrition.org, more defacements posted on this site are in IIS Web servers than in any other type of Web server. Raising the security level of IIS Web servers to a reasonable level requires that Web masters change a large number of Web- and system-specific parameters, disable accounts, remove certain access rights from groups that are afforded Web access, install a large number of patches (an ever-ongoing task), and so forth (Schultz, 2001). Although the graphical user interface (GUI) that Microsoft has designed for this Web server is for the most part effective from a humancomputer interaction perspective, the sheer number of task steps and knowledge required to increase the security of the IIS Web server to a level needed to resist most attacks is daunting. From a human factors perspective, the Apache Web server stands in striking contrast to the IIS Web server. Apache is also bundled with a number of operating systems (e.g., RedHat, Linux, Solaris, and many others). In comparison to IIS, it is more difficult for an amateur Web master to create a Web site using Apache. The reason is that this Web server requires selection of desired modules (some of which* are far more important to security than others), and then compilation of the source code and configuration of a number of parameters distributed in various files throughout the system that houses the Web server (Schultz, 2002). Interestingly, however, the default parameters in Apache have for many years been more conducive to security than are the default parameters in IIS Web servers.* Apache, for example, by default runs with the privilege level of a user, something that is beneficial for security because if someone exploits an Apache vulnerability, that person will be able to obtain user, not superuser, privileges. In addition, the Apache Web server has special features such as constraints called "directives" that prevent Web users from being able to access certain directories and files, and from running certain dangerous programs and services. The syntax of "directives" is quite straightforward, although the fact that text lines must be formatted in a particular manner is not terribly conducive to usability. Shown as follows, for example, is the format of a directive that prevents Apache Web users from being able to list the contents of a particular directory: Options -Indexes All things considered, therefore, a complex relationship between usability and security in Web servers exists. Usability is high for a default deployment of the IIS Web server, but security is inadequate unless the Web master engages in numerous tasks that require a fair degree of specialized knowledge. Proper
Apache Web server installation is more difficult, but once the Web master gets by the initial hurdles involved in Apache installation, the security level is by default better (although Apache is by no means perfect from a security perspective either). Note that virtually all the other previously mentioned Web servers fall somewhere between the extremes associated with the IIS Web server (easy to create, hard to secure) and the Apache Web server (more difficult to create, fairly secure once created).
Human Factors Challenges in Writing Secure Web Applications Securing Web servers is important, but doing so is only part of achieving total Web security. Another important consideration is securing Web applications that run on Web servers. Web servers can be secure, but if the applications are not secure, malicious users can exploit vulnerabilities in applications to accomplish a range of dire outcomes, including financial fraud, unauthorized gleaning of personal information about other users, application crashes, loss of control of Web servers to attackers, and so forth. The state of the art of securing Web applications has generally improved over the years as the result of organizations having suffered the consequences of running insecure Web applications, as well as the increased availability of information concerning how to secure these applications. Unfortunately, the security of Web applications and usability considerations generally do not go together very well. The main reason is that the majority of Web applications are implemented as common gateway interface (CGI) scripts, which are usually written in the Perl scripting language. Perl is difficult to learn—it has a perplexing syntax and wide range of precise conventions. Consider, for example, the following Perl expression: $ number=~/^/ \d-/+{l, 12} $/ input [0]";
die "Non-allowed characters in
The start of this string ($ number=~) in essence means that phone numbers that are entered must adhere to the rules that follow. \d means that numerals are acceptable input. - means that hyphens in the phone number that users enter are also acceptable. ^/ \d-J + signifies that any permitted characters (in this particular instance, numerals) are allowed, beginning at the start of every line. {1,12} means the amount of input must be between 1 and 12 characters; otherwise, it will be rejected. This length constraint is especially critical because it guards against the potential for buffer overflows due to excessively long input strings. $ means that when the string comparison is complete, the end of the line has been reached. This final constraint stops an attacker from appending commands or other types of dangerous input after the last character of input. If an
*The mod_ssl module, for example, supports secure sockets layer (SSL) encryption, a commonly used type of encryption that helps protect information sent between Web servers and browsers from being read by unauthorized individuals. *This statement is true for all versions of IIS between 1.0 and 5.1. Microsoft, however, completely reimplemented its IIS Web server in IIS version 6.0 (the version that comes with Microsoft's newest server operating system, Windows Server 2003), such that IIS is considerably more secure by default. IIS 6.0 and Apache's newest version, 3.1.27, appear in fact to be very comparable in default security levels.
33. Web Security and Privacy input string does not adhere to all these rules, the program quits and displays "Non-allowed characters in input." The previous example poignantly illustrates the usability problems involved in writing secure Web applications. Given the difficulty in writing applications in Perl (and other languages, too), developing Web applications is usually in and of itself an arduous task. Yet, truly secure Web applications need not only to perform input checking, but they must also (among other things): • Be modular with well-defined modules that rescind any elevated privilege levels when transitioning from one routine to another • Run with the minimum privilege levels that are needed (so if an attacker exploits a vulnerability somewhere in the application, the likelihood that the attacker will gain superuser privileges is lessened) • To a maximum extent, avoid making dangerous system calls that can allow access to system files or memory, or that could allow someone who accesses a Web server to execute commands on the system that supports the Web server • Encrypt sensitive information • Filter out state and environment variables as well as excessively long input strings (as explained earlier) It is little wonder, then, that writing special statements and routines for the sake of security is often pushed aside for the sake of completing applications within an allocated period of time. Not all Web applications are based on CGI scripts, however. Functions such as Active Server in IIS and scripting languages such as PHP (which is an acronym for PHP: Hypertext Preprocessor) in Apache can help a Web developer much more quickly and easily create Web pages and Web site functionality independently of scripts. Interestingly, a large number of security-related vulnerabilities have been found in both Active Server and PHP. In addition, achieving the degree of precision in controlling security as in the statement shown previously is generally not possible with Active Server and PHP. Once again, human factors and Web security appear, to at least some degree, to be orthogonal to each other.
Human Factors Challenges in Ensuring Privacy Ensuring privacy requires a large number of measures, and is thus not trivial to achieve. Cookies, for example, pose a particularly difficult problem for privacy, but it is possible to encrypt cookies, ensuring people other than a particular user can read the information in any particular cookie. Encrypting cookies requires a Web application developer to use one or more command(s) to do so, which is well within the knowledge domain of most Web developers. Other privacy-related challenges, such as protecting information in HTML hidden fields, are addressed by finding other ways to store and retrieve user-related variables, something that generally requires a higher than average amount of programming skill. Privacy enhancement modules
619
for Web transactions with many default values that would protect user/customer data and that would also allow inspection of these values in an optimal manner (e.g., via a spoke display depicting how far from desired values each of a number of critical security parameters is) would be a welcome addition from a human factors standpoint. Unfortunately, no such modules currently exist.
Human Factors Challenges in Securing Data Sent Over Networks Cleartext data going over a network comprise a major security threat. Hardware devices and programs can easily capture the data, making them available to perpetrators. In subscription Web sites, users must enter their passwords to gain access to Web servers. If the passwords traverse the network in cleartext, they are subject to being captured and then used to gain unauthorized access to user accounts. Sensitive information such as credit card and social security numbers can be captured in the same manner. The most frequently used solution for protecting data sent to and from Web servers is encryption. Encryption means scrambling characters using a special algorithm so only individuals who possess a particular object called a "key" can unscramble them (Schneier, 1998). The most often used type of encryption for Web traffic is secure sockets layer (SSL) encryption. Although many newer versions of Web servers and newer operating systems have built-in SSL, this is not always true. Installing either generally requires knowing and entering numerous commands or interacting with a series number of graphic interface screens (which in most cases is the easier way) to perform a range of tasks, such as generating an encryption key to setting parameters within configuration files. Consider, for example, the steps required to enable SSL on a Windows NT system: 1. Bring up the Internet services manager (ISM) tool (which typically requires going through at least three levels of menus, although this depends on how the ISM is configured). 2. Generate a key request from a certificate authority (CA) by filling out the screen that appears. 3. Contact the CA and go through the steps required to apply for the certificate. 4. When the CA approves the application and sends the key, use the ISM to install the new key. Go from Properties to Directory Security to Secure Communications and choose options (e.g., where the key will be saved) that are presented on screens that follows. 5. Open Windows Explorer. 6. Enable security for every folder that users are supposed to reach on the Web server by right-clicking on the folder to Properties, to Security, and then Encryption. Even if SSL is built into a system by default, the system administrator or Web master must nevertheless engage in a number of actions (e.g., by entering commands or interacting via a graphic interface) to configure SSL properly. Just as we have seen previously, Web security, this time in the form of encrypted
620
SCHULTZ
network transmissions, is difficult to achieve from a human factors point of view in that specialized knowledges and entry of precise commands or knowing what components of a graphic interface are both required. Whenever encryption is used, key management (ensuring keys are stored properly and that additional copies are available if needed) becomes a necessity; if a key is destroyed or damaged, encryption and decryption fail, often with catastrophic consequences (e.g., not being able to read business-critical files that have been encrypted). Key management also involves a number of human-in-the-loop tasks, adding substantially to the complexity of using encryption. Someone must at a minimum, for instance, perform a number of tasks such as searching for each user's encryption key, verifying that it is indeed the correct key for that user, copying that key to some medium (e.g., a CD or USB storage device), labeling the medium correctly, and so on. Although SSL has proven effective in protecting data sent over the network, critics have pointed out that there are a number of inherent limitations in SSL that render this encryption protocol less than adequate when activities such as businessto-business (B2B) transactions, monetary transfers, and credit card debits are involved. Several credit card companies, Microsoft, and others created an alternative to SSL, the secure electronic transaction (SET) protocol, to provide an integrated way of handling security for such transactions. Among other things, SET provides strong user authentication, validation of credit card numbers, and authorization of specific transactions. SET not only encrypts all information sent over the network, but it also hides information about individuals from merchants and the nature of any purchases from the banks or credit card companies that process transactions. Although SET in principle provides a much more secure, private, and comprehensive method of handling transactions, SET'S popularity has waned dramatically in recent years. The principle reason is that SET has inherent usability liabilities to the point that many users never learned to use this protocol in the first place, or they learned how to use it, but quit using it because of the difficulty of usage. To simply initiate a SET transaction, for example, the user must request and then complete a certificate (an electronic data structure used to identify individuals and systems and to transmit encryption keys), which requires literally dozens of discrete interaction steps. Listing all the steps in a complete SET transaction would require more space in this chapter than is available, so the interested reader should visit http://www.setco.org/download.htrnltfspec to learn just how complicated SET transactions are from a human-computer interaction perspective. SET is another case in point that security and usability often do not coexist very well. One-time passwords provide another potentially good solution for protecting against the unauthorized capture of passwords sent by users attempting to remotely log on to Web servers. Many different versions of one-time password solutions are available; regardless of the particular solution, human factors problems generally abound in this method of authentication. One of the most frequently used types of one-time password tools is one in which a program generates a list of one-time passwords for a series of logons on a per user basis. This list contains columns of number-password pairs, each of which is good
FIGURE 33.1. A hypothetical one-time password list.
for one and only one logon. Figure 33.1 contains a hypothetical one-time password list for ten successive logons. For each logon attempt, the system to which the user is allowed access displays a prompt consisting of the logon number (corresponding to one of the numbers in the left column in Fig. 33.1). The user must enter the password with which this number is paired. So, for example, referring to Fig. 33.1, it is easy to see that the user needs to enter "SDle$76yF" on the first logon attempt, "yL%5U1VCx" on the second, "3*sAl@z4" on the third, and so on. Remembering passwords that are difficult to crack is typically more difficult for users than remembering trivial (and thus easy to crack) passwords (Proctor, Lien, Vu, Schultz, & Salvendy, 2002). A password such as "safeplace" would thus be considerably easier to remember, for example, than would "4hFd*&bX," although the latter would be considerably more difficult to crack. One-time passwords, however, for the most part obviate the need to choose good passwords in that by the time an attacker or password cracking tool can determine a password, that password is likely to have been already used. One-time passwords thus potentially solve a range of human factors problems related to user memory. At the same time, however, one-time passwords create new, nontrivial human factors problems. One-time password lists are, for example, simple columnar displays plagued with well-documented usability problems, such as proneness to visual vertical transposition errors in which users enter the password for a preceding or succeeding logon on a particular logon attempt. In addition, passwords for each logon attempt almost invariably consist of alphabetical and numeric sequences that do not resemble dictionary words. As such, users tend to enter them more slowly and with more errors than simpler, more meaningful passwords. Still once again, human factors and security are in the opposite direction from each other.
33. Web Security and Privacy
Human Factors Challenges in Securing Systems That Host Web Servers Web servers run on a wide variety of operating systems— Windows NT/2000/XP, Linux, Unix, Macintosh, OS-390, and many others. It is also critical to ensure the operating system on which any Web server runs is secure. Failing to secure the operating system, but making the Web server as secure as possible does not work; attackers will be able to exploit operating system vulnerabilities to reach the Web server without authorization. Security guidelines for the major operating systems that are currently used are posted at http://www.cisecurity.org/ The particular measures needed to secure an operating system to a large degree depend on each operating system in question. Some overlap nevertheless exists. The taxonomy for human factors in information security previously developed by Schultz et al. (2001) applies especially well to operating systems. This taxonomy, depicted in Table 33.1, includes six major types of security-related tasks: identification and authentication, en-
suring data integrity, ensuring data confidentiality, ensuring data availability, ensuring system integrity, and detecting intrusions and misuse. Each task has associated usability issues; some tasks such as identification and authentication present special challenges in that they must be performed by users, who although often insufficiently trained must complete rather complex behavior sequences without any errors to achieve success in a given task. In addition, ensuring system integrity too often involves tasks that require perfectly precise (and thus error-intolerant) entries in configuration files. Certain entries in Unix and Linux systems enable system administrators to learn whether anyone has engaged in actions that have threatened system integrity. Consider, for example, the entries for the critical/etc/syslog configuration file (which controls the type and amount of system logging in Unix and Linux systems) shown in Fig. 33.2. The first line (which begins with "#") is a comment line that does not affect the level of logging that is captured, but appears purely for the purpose of providing context to whomever reads the entries in this file. The second line in essence means that all
TABLE 33.1. A Taxonomy of Security-Related Tasks and Associated Usability Issues Type of Task
Type of Threat Countered
Identification and Authentication
Masquerading as another user; repudiation
Data integrity
Unauthorized deletion and/or changes
Data confidentiality
Unauthorized disclosure and/or possession Unauthorized deletion of data and/or the databases/programs used to store and retrieve them; denial-of-service attacks Unauthorized deletion and/or changes to system data/configuration files; theft; denial of service attacks Unauthorized access to systems; denial of service attacks
Data availability
System integrity Intrusion detection
621
Usability Issues Willingness of users to adopt; ease of using method/device Install and maintain appropriate software; control access rights and privileges Control access rights and privileges Protection provided by system-managed backup media; ease of implementing Inspection by administrators; detection by software Inspection by administrators; detection by software; ease of implementation
FIGURE 33.2. Entries in the /etc/syslog file used to control levels of system logging.
622
SCHULTZ
events that happen in a system that are at the priority of "errors" or higher in an eight-level hierarchy of auditing priorities,* all events that happen in the operating system's kernel (the most basic, innermost part of an operating system) with the priority of "debugging" or higher, and all authentication-related events that have "notice" priority or higher will be sent to the console (the terminal itself). A slight error, such as inserting a colon where a semicolon goes or inserting an extra space in the entries in either the left- or right-hand column, renders an entry completely meaningless. Needless to say, the need for a more usable syntax for system configuration entries, or possibly a graphic user interface with pictorial representations of the results of choosing any particular level of logging, remains a very high human factors-related priority in most operating systems today. The main point in this case is that when operating systems are considered, human factors becomes even more important to Web security because of the need for humans (in this case, system administrators) to intervene in systems to provide reasonable assurance that vulnerabilities and/or faulty configurations in systems will not be the vector through which Web servers that run on them are compromised. Unfortunately, the computing world as a whole relies almost completely on well-experienced system administrators who over time have mastered idiosyncrasies in syntax and other usability hurdles to be able to configure the systems that house Web servers properly for the sake of security. The issue of having beginning or junior system administrators is, however, a completely different one—one in which effective usability design could potentially make a huge difference.
Human Factors Challenges in User Interaction With Browsers Using browsers also presents a number of human factors challenges. Browsers typically notify users of conditions in which security could potentially be threatened or when changes that might affect the user could occur. The Netscape Communicator browser is one of the best examples. This browser frequently displays warnings (in the form of dialog boxes) whenever a Web server attempts to load cookies into it or when certain security features are about to be disabled. Although the content of such dialog boxes superficially may be easy to understand, the pure frequency with which they appear and the nature of their content (which may warn users that yet another cookie is about to be downloaded) tends to wear users out. Furthermore, users generally have little knowledge concerning a "good" and "bad" cookie, or what aspects of a particular XML constitute "dangerous" or "safe." Consequently, users frequently resort to simply turning off all such warnings, a task that requires several nonintuitive steps, but which once completed spares users from further inconvenience, distraction, and confusion at the expense of greater security risk to their systems. Similarly, browsers may
offer users choices of types and levels of encryption. The average user has no idea whatsoever of the meaning of these choices. The same principle applies to types of XCLs that may or may not be downloaded and executed in a user's browser. The typical user has no idea whatsoever of what an ActiveX control or a Java applet is; the fact that dialog boxes warning users of the imminent downloading of an XCL offer virtually no explanation of the relative dangers of each type effectively take controlling the downloading of these executables out of the hands of users. Worse yet, many browsers offer by default a very vulnerabilityridden version of the secure shell program, a program that encrypts all traffic sent between two computers over the network (Cheswick, Bellovin, & Rubin, 2003). Turning this version off, however, requires a multiple set of steps that very few users and novice system administrators are capable of performing without training. Better ease of use would result in considerably more effective user interaction with browsers in matters that affect security. Warnings concerning cookies that are to potentially be downloaded into browsers should, for example, be made more simple and meaningful to users. Giving users a simple desktop option to designate certain "trusted" Web sites from which cookies will always be downloaded without interruption to the user is a step in the right direction. Warnings could then be presented whenever cookies from other sites are about to be downloaded. Microsoft's Internet Explorer browser conies closest to this prescription of any well-used Web browser. If a user brings up this browser (by double-clicking on the icon for Internet Explorer on the desktop) and then pulls down the Tools menu in the menu bar at the top to Internet Options and then clicks on the Privacy tab at the top of the dialog box that appears, the form shown in Fig. 33.3 below appears: Using this form, users can manipulate a slide bar to select privacy levels anywhere between high and low. In Fig. 33.3, a medium level of privacy has been selected. This level not only blocks third-party cookies from sites that do not have a concisely stated privacy policy (something that, unfortunately, most users are not likely to genuinely understand), but also prevents cookies from third-party sites that glean personal information without the user's consent from being downloaded and restricts first-party cookies that use such information without the user's consent. The fact that the menu depth is shallow, not deep, makes these options easier for users to find (Schultz & Curran, 1986). The slide bar method of selecting privacy levels not only provides an intuitive method of setting the privacy level, but it also allows users to explore the outcome of choosing different privacy levels before they make a selection. The main usability limitation appears to be the meaning of the options—many users do not really know what a "cookie" is, how "third party" differs from "first party," and so forth. Still, in general, the interaction methods available to the user are for the most part compatible with principles of effective human-computer interaction.*
*Priorities of logging are (from highest to lowest) emerg, alert, crit, err, warning, notice, info, and debug. *Note also that if a user selects Internet Options and then clicks on Security, the user can also designate "Trusted Sites," Web sites that users believe will not damage their systems, as well as "Restricted Sites," sites that for various reasons (e.g., that site has been known to download malicious code into browsers) cannot be trusted to download anything into the users' browsers. Once again, the human-computer interaction methods necessary to designate trusted and restricted sites are for the most part very simple and intuitive.
33. Web Security and Privacy
•
623
FIGURE 33.3. The privacy control dialog box in the Internet Explorer browser.
Color-coding dialog box messages according to the potential for privacy compromise would also be an effective measure. Alternatively, each user could designate once and for all the types of information (e.g., social security numbers) that can and cannot be contained in cookies, thereby eliminating the need to constantly allow or disallow a new cookie from being downloaded. In addition, instead of having to traverse through several layers of menus to disable XCLs from being downloaded, the desktop should make this option available. Each type of XCL should be color-coded to clearly indicate its relative danger, with ActiveX (the most dangerous) in bright yellow and Java (one of the safer types of XCLs) in a green-yellow color. Better defaults would also obviate or, in some cases, substantially reduce the need for user interaction altogether. Browsers should not, for example, even offer encryption methods that provide weak security, thereby sparing users from having to make choices that are for all practical purposes nonviable.
CONCLUSION This chapter explores how usability and Web security and privacy are interrelated. Failure to consider usability issues can result in tasks that need to be performed for the sake of Web
security and privacy not getting done, or being done improperly. The result is the presence of vulnerabilities that can be exploited by perpetrators of computer crime. Unfortunately, we have seen that security and usability needs can be and are often conflicting—that higher security levels are too often associated with the performance of long, complex, and unintuitive task sequences. One straightforward solution is to simply elevate the default level of security in Web servers. Vendors are reluctant to adopt this solution, however, because raising the default level of security often results in malfunction in Web servers and possibly often systems. Customers generally avoid buying products that do not work correctly out of the box. Yet raising the default settings would obviate the need for Web masters to change so many settings and take other measures that improve security. In addition, Web server vendors could offer simple settings that result in groups of related security parameters settings. So, for example, a Web server could have a setting for overall security level—high, medium, and low. In this scheme, the high setting would result in scores of settings that would tighten file permissions on Web-related files to allow nothing more than Read access to anyone but Web masters and system administrators, lower the privilege level with which the Web server and its
624
SCHULTZ
FIGURE 33.4. A dialog box from Microsoft's IIS Lockdown tool.
applications run, increase the amount of logging of Web transactions to the maximum, keep users from being able to submit any kind of input, and so on. In contrast, choosing the low setting would result in a group of settings that would correspond to outof-the box settings. Ideally, Web masters would have a graphical interface through which they could use a pointing device to select the desired level. This type of solution has been implemented in recent versions of RedHat Linux with enormous success. System administrators can now use a graphical interface to set a firewall (traffic blocking) function to high, medium, or low security, thereby greatly simplifying network security. To do this, system administrators must simply enter setup and then choose the menu option named "Firewall Configuration" and either "High," "Medium," or "Low." User interaction with browsers could be substantially improved if human-computer interaction methods, such as colorcoding of the relative danger of various conditions such as downloading of cookies and elimination of excessive menu depth, were used. In particular, privacy should be put in the hands of each user through improved interaction methods. Allowing simple, up-front specification of what the user will and will not tolerate in terms of potential privacy compromise would be a significant step forward although, as mentioned previously, Microsoft's Internet Explorer browser comes closer to fulfilling these requirements than does any other widely used browser. Finally, more security-enhancing routines that can be easily plugged into current systems and applications and tools that run on Web servers and in connection with applications that run on the Web servers could also help usability considerably
provided, of course, that they minimized the number and complexity of steps needed to elevate security. So, for example, instead of having to write expressions such as the Perl expression shown earlier, a programmer would simply have to locate the appropriate input-filtering routine and integrate it into one or more of the applications that run on any Web server. Microsoft has developed one such tool, the IIS Lockdown Tool, for its IIS Web server. The IIS Lockdown Tool is an easy to install tool that weeds out malformed URL requests, preventing a wide range of attacks and precluding the need to write custom filters (Fig. 33.4). Better yet, having something such as an integrated toolkit from which any particular component tool could be selected from a simple graphic display would be a gigantic step in the right direction. These and other solutions are likely to emerge as individuals and organizations eventually realize just how great the scope of the usability problem is and how much time and resources they could save in creating and maintaining Web servers if security-related task sequences were made more intuitive. So the real message of this chapter is by no means of "gloom and doom," but rather of opportunity. A number of effective security-related human-computer interaction techniques already exist, as pointed out in various parts of this chapter. Security and usability are not in reality inherently incompatible; the problem is that the preponderance of current security-related human-computer interaction tasks have simply not been designed well from a human factors perspective. Devoting the time and effort to reengineer these tasks would result in great benefits, both to organizations and individuals, especially considering the ever-growing worldwide threats to computing systems, data, applications, and networks.
33. Web Security and Privacy
•
625
References Al-Ghatani, S. S., & King, M. (1999). Attitudes, satisfaction and usage: Factors contributing to each in the acceptance of information technology. Behaviour & Information Technology, 18, 277297. Bernstein, T., Bhimini, A., Schultz, E., & Siegel, C. (1996). Internet Security for Business. New York: Wiley. Cheswick, B., Bellovin, S., & Rubin, A. (2003). Firewalls and Internet security: Repelling the wily hacker (2nd ed.). Reading, MA: AddisonWesley. Cox, P., & Sheldon, T. (2000). The Windows 2000 security handbook. New York: McGraw-Hill. FBI Annual FBI and Computer Security Institute Survey on Computer Crime. On-line document, 2002. http://www.gocsi.com/ press/20020407.html Ilioudis, C., G. Pangalos, G., & Vakali, A. (2001). Security model for XML data. Proceedings of the 2nd International Conference on Internet Computing, 1, 400-406. Proctor, R. W., Lien, M. C., Salvendy, G., & Schultz, E. E. (2000). A task analysis of usability in third-party authentication. Information Security Bulletin, 5, 49-56. Proctor, R. W, Lien, M. C., Vu, K-P. L., Schultz, E. E., & Salvendy, G. (2002). Improving computer security for authentication of users: Influence of proactive password restrictions. Be-
havior Research Methods, Instruments & Computers, 34, 163169. Proctor, R. W., Vu, K-P. L., Salvendy, G., Degan, H., Fang, X., Flach, J. M., Gott, S. P., Herrmann, D., Kroemker, H., Lightner, N. J., Lubin, K., Najjar, L., Reeves, L., Rudorfer, A., Stanney, K., Stephanidis, C., Strybel, T. Z., Vaughan, M., Wang, H., Yang, Y, & Zhu, Wenli. (2002). Content preparation and management for Web design: Eliciting, structuring, searching, and displaying information. International Journal of Human-Computer Interaction, 14, 25-92. Schneier, B. (1998). Applied cryptography (2nd ed.). New York: Wiley. Schultz, E. E. (2001). IIS Web servers: It's time to just be careful. Information Security Bulletin, 6, 17-22. Schultz, E. E. (2002). Guidelines for securing Apache Web servers. Network Security, 8, 8–14. Schultz, E. E., & Curran, P. S. (1986). Menu structure and ordering of menu selections: Independent or interactive effects? SIGCHI Bulletin, 18, 69-71. Schultz, E. E., Proctor, R. W., Lien, M. C., & Salvendy, G. (2001). Usability and security: An appraisal of usability issues in information security methods. Computers and Security, 20, 620-634. Whitten, A., & Tygar, J. D. (1999). Why Johnny can't encrypt: A usability evaluation of PGP 5.0. Proceedings of 8th Usenix Security Symposium. Berkeley, CA: Usenix Association.
This page intentionally left blank
Part
XI EMERGING TECHNOLOGIES AND APPLICATIONS
627
This page intentionally left blank
Z34Z WIRELESS COMMUNICATION Mark Pearrow Massachusetts Institute of Technology For a list of all the ways technology has failed to improve the quality of life, please press three. —Kahn
The Evolution of Wireless Technology
INTRODUCTION
Throughout history, certain innovations have had lasting and profound effects on the way human civilization has progressed. In the last century, an exponential explosion of technology has reshaped society and forever changed the way humans do some of the most essential things—such as communicate, for example. Electronic communication has certainly had an immense impact on the way humans do business, manage health, and even fall in love and wage wars. It is hard to imagine that it was not until 1729 that electricity was transmitted over a wire by English chemist Stephen Gray. It took almost another 150 years for Alexander Graham Bell to put the principles of electronic transmission to work in the ubiquitous telephone. In the late 1960s, the development of the ARPAnet heralded a new age of electronic communication, made possible by a small network of computers. Packet radio networks such as ALOHANET made data communications over the airwaves a reality. In 1973, Martin Cooper, then a general manager for Motorola, made the first call on his invention, the cellular telephone. By the 1980s, the use of the Internet, which was formerly the ARPAnet, was growing among academic and research locations around the world, fueled by the NSFNet, a cross-country 56-Kbps network backbone funded by the National Science Foundation in the United States. Applications and services for the Internet, such as ARCHIE, FTP, GOPHER, and TELNET, helped popularize the Internet as a tool for collaboration among academics. However, it was not
There is a growing trend of extending computing capability and network connectivity well beyond the cubicle or family PC, and wireless computing technology is the key to enabling unfettered digital communication. However, wireless computing is still in an embryonic stage, and a host of challenges await human factors specialists if they are to help usher this potentially powerful computing paradigm into the realm of usability. To sharply underscore this point, a 2001 study by the Meta Group showed that as many as 80% of all corporate users that had purchased wireless application protocol (WAP)-enabled phones had completely abandoned the data capabilities of the phone and only used the voice capability. Yet, virtually all cellular operators today are hawking next-generation data capabilities on their cellular networks. AT&T's mLife™service promises to "transform your wireless into a new world of possibilities." Sprint promises that their PCS Vision™service will "expand wireless possibilities beyond anything you've seen or heard before." Are polyphonic ring tones and bitmapped graphics really the keys to such remarkable "transformation"? The experience of usability specialists will be needed to delineate marketing hype and what users really need. In this chapter, we explore the foundations of wireless technology and the problems associated with wireless computing, and recommend guidelines for improving wireless usability.
629
630
PEARROW
until a scientist at CERN in Switzerland named Tim Berners-Lee created a curiously simple protocol, hypertext transport protocol (HTTP), and a markup language, hypertext markup language (HTML), that the Internet became accessible to people with nonacademic backgrounds. Indeed, since its introduction in 1991, the World Wide Web has accelerated the growth of the Internet by phenomenal levels, and made the prefix "www." and the suffix ".com" household words throughout most of the modern world. Today, cellular telephony is beginning to supplant the now-ancient technology of POTS (plain old telephone system) technology, and along the way is accreting more and more of the functionality of once breadbox-size computers, as well as connectivity to the rich information network that marbles the World Wide Web and its countless nodes. Current Trends. It is no wonder that so many power users of computing systems and cell phones, as well as device manufacturers, have longed to marry the power of the Web with the portability of cellular telephony. Even if no one has proposed the true "killer app" of portable, connected computing, it at least seems like the right direction to be going if we are ever to attain the functionality that has long been dreamt up by science fiction authors. The past 5 years have seen many proprietary concoctions by cell phone, personal digital assistant (PDA), software, and computer manufacturers, as these organizations seek to gain the edge in portable, connected computing. The outcome has been largely disappointing, and the resulting miasma of vaporware, incongruent platforms, protocols, and paradigms has made life difficult for developers, user, and usability specialists alike. This digital Tower of Babel has made it difficult to craft specific usability tips, tools, and techniques, which has resulted in a wholly unsatisfying "abstract emptiness" in place of sound usability advice for wireless content developers. The Palm VII PDA, released by Palm, Inc., in May 1999, was the first portable computing device that boasted wireless access to the Internet. It ran the popular Palm OS, and connected to the Internet by way of the BellSouth Wireless Data Network. Initially, users of this wireless data service had to pay anywhere from $9.99 USD for 50 kilobytes of data per month to $24.99 USD for 150 kilobytes of data per month. This pay-per-packet approach to billing tends to accompany the first incarnation of wireless data services. In August 2001, Sprint introduced its PCS Vision™ service for its global system for mobile communications (GSM) network, and the pricing was a whopping $50 USD a month for only 2 megabytes of data, with an extra kilobyte costing an additional $.02. Verizon Wireless also started out of the wireless data gate with a steep $35 USD per month fee for 10 megabytes of data per month or $55 USD per month for 20 megabytes. However, the fact is that pay-per-play plans are simply not popular with users who do not want to have to keep track of usage or suffer the overlimit fees associated with extra transmissions. Unlimited service has become the de facto standard for dial-up and broadband Internet connectivity, so why should it not also be the case for wireless data? Sprint revised its pricing plan in October 2002 and offered unlimited data for
*http://www.wired.com/wired/archive/10.10/wireless.html.
a mere $10 USD a month. It is likely that most 3G data services will come with standard pricing packages that include unlimited access for a flat monthly fee, but history has shown that when a new communication technology emerges, the end user will generally pay per unit until the user base grows sufficiently and the technology is accepted broadly. In addition to cellular-based wireless technologies, portable computing is also enabled by wireless local area network (LAN) technologies such as Wi-Fi or 802.11b. 802.11 is a family of specifications of wireless data communications protocols that have been developed by the Institute of Electrical & Electronics Engineers (IEEE). Perhaps the most popular variant of 802.11 is 802.11b, more commonly known as "Wi-Fi," and easily recognized by the familiar Wi-Fi logo that can be found on just about every commercially available wireless access point and transceiver. 802.11b uses the 2.4-GHz band and provides data rates up to 11 Mbps (with graceful degradation of service to 5.5, 2, and 1 Mbps). Companies such as Apple Computer and Linksys have made 802.11b not only economically feasible for huge numbers of home and business users, but also easy to implement. The Apple Airport Base Station is trivial to configure, even for users who are not technically savvy. The Linksys WAP series wireless access points are also easy to configure, allowing anyone with a broadband connection to extend their network into the wireless domain. The proliferation of these wireless access points and the widespread availability of broadband Internet access—via cable TV infrastructure and digital subscriber line (DSL) from the phone companies—has been a catalyst for the "hot spot" phenomena. Wi-Fi access points allow the inexpensive extension of an Internet connection, often inadvertently, to anyone who happens to pass by and notice the signal. The Linksys BEFW11S4 Wireless Access Point Router with four-port switch, for example, has a published maximum range of 457 meters, or 1,500 feet! In a densely populated neighborhood, if even a small percentage of residents have such an access point, it can create a seamless mesh of signal availability. Businesses that cater to mobile professionals are also beginning to employ Wi-Fi networks, frequently as a magnet service to attract clientele. Nicholas Negroponte, Director and Founder of the Massachusetts Institute of Technology's (MIT's) Media Laboratory, has described his vision of users of future wireless Internet access as "frogs and water lilies,"* in which distributed wireless access points act like lily pads, among which the wireless user—the frog—can hop at will. Negroponte also believes that this sort of grassroots, or peer-to-peer, network could eventually supplant commercial data networks. Regardless of whether we all become frogs in the future, it is very likely that 802.11 and its progeny will play a major role in mobile computing. There is a large difference in these two basic technologies; 802.11 provides wireless access for just about anything that can use a standard PCMCIA (PC card) wireless network card, which means that laptops, PDAs, and even desktop computers with a PC card adapter can participate in 802.11 networks. Cellular data networks generally require special hardware that is
34. Wireless Communication
more expensive to obtain than Wi-Fi cards, which have dropped in price substantially over the last few years. Wi-Fi is a name that has relatively high recognition by people who are likely to use the technology, and as history has shown, technologies that gain a marketing foothold are more likely to survive than superior technologies that have less shrewd marketeering behind them. Most cellular and two-way radio networks have a fraction of the bandwidth of an 802.11 network. For example, a marginal Wi-Fi signal can net a user a 5.5-Mbps connection (of course, the bottleneck is probably not the Wi-Fi access point, but rather the upstream Internet connection), although a full-strength Sprint VisionPCS® signal will only ever yield a 144-kbps connection, with average speeds ranging from 50 to 70 kbps. Generally speaking, less bandwidth means less bulky content. Standard Web pages are designed for browsers on laptop or desktop PCs that are likely to have large amounts of screen resolution, central processing unit (CPU) power, and a large data pipeline. Content that is delivered to a mobile device that uses one of the "narrower" carrier technologies must be optimized somehow to reduce the amount of data needed to transport a page's content. It is also important to note that currently the cellular and other radio networks are more likely to be available in nonresidential and nonbusiness areas, such as along a stretch of highway; Wi-Fi is currently only commonly available in residential and urban, commercial settings. There is no nationwide 802.11 infrastructure. However, there is a possibility that this will change in the future. For example, British Telecom had announced plans to deploy a nationwide network of Wi-Fi hotspots by June 2003.* Device Styles. The current trend in wireless computing devices is to integrate cellular telephony with other sorts of electronic devices, as evidenced by the popularity of the Handspring Treo™line of products. The Treo 300 combines a GSM telephone and a Palm OS-driven PDA into one fliptop unit. It also features an integrated thumb keyboard. Although there has been some backlash against devices like these that cram multiple functions into one, many users also report satisfaction with them despite their shortcomings. The most common complaints against the Treo and similar devices pertain to the small screen size (compared with the larger screen of most PocketPC platform devices). In particular, the Treo 300 lacks ability to send short message service messages, although it can receive them. In addition, in the United States, many users have complained about what they perceive as poor service quality available from Sprint, the first cellular carrier who provided service with the Treo 300.* This is potentially due to the fact that the Treo 300 is a single-band phone with no analog roaming capability. Figure 34.1 shows the Handspring Treo™300. According to Gartner, Inc., a Connecticut-based research and advisory firm, PDA-style devices that are powered by the Palm OS lead the pack with about 55% of the market share,
631
FIGURE 34.1. The Handspring Treo 300. Note. Copyright © 2003 Handspring. Image reprinted with permission.
with Windows CE/PocketPC platform-based designs following at about 26%.* This split is likely to change over time if history is any indication of what lies ahead. Microsoft has made a foray into a new area of portable computing with its Tablet PC line and version of Windows XP that is optimized for tablet PCs.* Other operating systems are dominant in the cell phone market, however, with Symbian OS peaking in 2002 at a 57% share.* Wide and narrow connectivity. These two different broad categories of carrier-level technology are likely to change rapidly over time, in all likelihood narrowing the disparity between the two. At the time of writing, however, the difference means that there are at least two usage paradigms for mobile computing: "wide" and "narrow." Wide connectivity is highly correlated to a higher level of user input/output G/O) for more productive applications: verbose e-mail, Web browsing and document management, file transfer, and so on. Narrow connectivity is highly correlated to more Spartan modes of computing interaction: a quick check on a flight's status, the latest stock prices, the closest bar, and so on. At least one telecom company sees the two disparate technologies as complementary; Wi-Fi provides a means by which congestion on the cellular airwaves can be relieved, whereas cellular data allow greater mobility at the expense of bandwidth. The future is likely to hold much in store for
*http://www.btplc.com/Mediacentre/Agencynewsreleases/2002/an34.htm. *http://computers.cnet.com/hardware/0-2709830-4l8-20252976.html?tag=subnav. *http://www.nwfusion.com/newsletters/mobile/2003/0203mobile2.html. *http://www.microsoft.com/windowsxp/tabletpc/. *http://www.idg.net/idgns/2002/10/21/StudyNokiaBoostsSymbiansHandheldOS.shtml.
632
PEARROW
phenomenal growth in wide area, high-bandwidth wireless data networks, and the line between a "handheld device" and a "desktop computer" is likely to blur, owing in part to the availability of increasingly powerful, yet tiny, processors, and advancement of battery technology. Narrow connectivity is also almost always a temporary mode of operation because it is typically characteristic of early stages of development of new communications technology. As techniques and understanding of a given technology improve, the bandwidth associated with that mode of communication also increases. Only in a few unique cases is narrow connectivity considered a feature, such as in military applications where undetected communication is more important that wide communication. Some experts strongly believe that wide connectivity, a.k.a. wireless broadband, will never be attainable through means such as cellular telephony, and that in fact because of health risks and other factors, wireless broadband will always remain limited to ranges of just a few hundred feet (LeRouge, 2001). If narrow and wide connectivity are truly complementary, then it is obvious that several mobile usage paradigms will likely evolve: widely available, narrow bandwidth connectivity, and narrowly available, wide bandwidth connectivity. Usability experts will need to understand the differences in usage models for these and other potential modes of operation.
Shortfalls. All the advances in processors and memory have completely dwarfed other primordial components of any computing system, however, which means that the total package of mobile computing will have to lag until the more antiquated technologies catch up. Perhaps the most profound stumbling block for mobility is the fact that battery technology is still stuck decades behind the rest of the technology. New innovations in battery manufacturing are going to be essential to the expansion of mobile computing. Companies such as TransMeta have approached the battery-life issue with some novel approaches that make computing devices capable of running longer on the same old batteries. The real issue is that drastic improvements are needed in battery life, not just "band-aid" workarounds. Another major dilemma is the user I/O issue. Any person who has used a PDA for, say, taking meeting notes, is aware of how cumbersome all the options for getting data into the device are. Handwriting analysis, even when assisted by constraints, such as those integrated into the popular Graffiti™handwriting style, is an input mode that is high in portability, but very low in productivity. If a user wants to crank through many words per minute, a separate add-on keyboard—which is often as large as the PDA—is needed. This improves productivity, but reduces portability. Voice recognition, long showcased as the marvel of futuristic technology by sci-fi shows such as Star Trek, turns out to not be so helpful in many cases. Computer speech recognition still cannot effectively attend to one stream of information while ignoring others, including those that might sound like legitimate signal. Alternative keyboards have been popping up since the mobile computing and text messaging craze began. Many teenagers have developed a lightning-fast technique for typing with only
their thumbs on cramped handheld keypads. Alternatives such as the FasTap keyboard and the ThumbScript keyboard have attempted to address the essential problem of packing a fullfeatured keyboard into a small package with a minimal amount of training needed for the end user. Most of these alternate keyboards, however, place unnatural strain on the musculoskeletal systems of the hands and arms, which does not make them attractive options for frequent users. Voice recognition also implies that the user must speak information and commands out loud. This mode of interaction is not appropriate for many tasks, especially tasks that involve the exchange of sensitive information—credit card numbers or banking account and routing numbers, for example. Voice recognition in mobile applications also means that humans must walk around talking to inanimate pieces of equipment, often at the expense of other humans in the area who might not really want to hear all the sordid details of a stranger's life. This sociologic phenomenon is outside the scope of this chapter, however. The point is that the industry has so far failed to yield an input device that is portable, highly productive, safe, and easy to use. Output still lingers in the back ranks as well. Tiny LCD screens are, after all, tiny LCD screens, no matter how bright or fine the resolution. Several companies have proposed alternate output devices that use "microdisplays"—that is, really, really tiny LCD screens—that can be fixed inside a headworn device in front of the eye, giving the illusion of a large screen floating in front of the user. However, because not many day-to-day users are likely to want to wear a Borg-like headpiece just to read e-mail, no immediate solution for the output issue is in sight.
Wireless Computing: An Ecological Model In the field of psychology, ecological models have been useful tools for understanding human and group behavior within the context of the individual's social, political, and environmental surroundings (Bronfenbrenner, 1979). This approach is also useful for understanding mobile technology use. It helps identify elements of usability that fall outside the immediate sphere of contact of the user, which might have an otherwise large but unobserved impact on the user's experience. Palen and Salzman (2002) noted that "The creation, evolution, adaptation, and use of any technology are socio-technical phenomena." Such an approach is particularly indicated in the exploration of mobile computing usability analysis because many more factors can influence the system performance and user experience. This means that the definition of the term "usability" must expand beyond the innermost shell of user experience to encompass other factors, such as "Netware," or elements of the technological system that pertain to the mobile network, and "Bizware," or elements of the system that are sociotechnical in nature, including calling plans, marketing information, manuals, billing information, and customer support (Palen & Salzman, 2002). An integrated approach to usability must consider all these aspects in order to attain a "big picture" of the overall system. Although the same is also true of traditional desktop PCoriented usability, it is more so for mobile usability because there are many other external variables, such as cellular carriers, a
34. Wireless Communication variety of browsers, operating systems, devices, and social norms that can have an impact.
WIRELESS TECHNOLOGIES
633
802.11 family of technologies). Although these two classes fill different needs for mobility at the time of writing, it is possible that the gap will narrow or vanish over the next decade, so the distinction between devices such as smartphones and laptops will diminish, at least insofar as connectivity is concerned.
What Is Wireless, Anyway?
Overview of Key Technologies
Previously, we briefly discussed the differences between some of the popular wireless technologies, as well as some of the typical places where they are used. In this section, we take a closer look at the various wireless protocols and discuss their benefits, drawbacks, and likely direction in the future. Any sort of communication between two parties involves data, which are the bits of information that one party wants to send and that the other party wants to receive, and the transmission medium, which is the thing that the bits are sent through to get to the receiving party. In the case of spoken word, speech forms the "bits" of information, whereas air (or water, steel, or any other substance that is capable of propagating sound waves) is the transmission medium. In data networks, binary digits—ones and zeroes—are literally the "bits," whereas frequently Ethernet is the transmission medium.* This works fine for the vast majority of cubicle-oriented environments, where users sit in front of the same PC in the same place every day. Wires do not tend to move easily, and thus do not foster mobility. However, mobility is becoming increasingly attractive as workspace and workflow habits change, especially in the case of traveling workers who must treat any arbitrary location as an office—the passenger seat of a car, a bench in an airport terminal, and, if the television is an accurate portrayer of reality, a serene, temperate beach on a secluded island. Wireless data communication takes advantage of the principles of radiofrequency (RF) propagation to transmit data through the air, in much the same way that your favorite radio station is transmitted. Different wireless methods employ varying pieces of the RF spectrum—the range of radio waves that have been approved by government regulation, in this case—as well as different encoding techniques to improve speed, range, and reliability. The electromagnetic spectrum includes a vast range of electromagnetic energy, including visible light, X-rays, and our familiar RF waves. RF frequencies range from the very low-frequency band below 100 kHz, up to the extremely highfrequency band in excess of 100 GHz. Typically, portions of the radio spectrum are divided into areas known as bands. A band is generally used for a specific purpose. For example, the FM radio band runs from 88 MHz to 108 MHz. Wi-Fi takes up a part of the spectrum known as the 2.4-Ghz band, which encompasses frequencies ranging from 2,400 MHz to 2,483.5 MHz. This band is further subdivided into fourteen channels, a design that helps reduce interference with neighboring 802.11b sites. Two main classes of wireless access are dominant at the time of writing: cellular network and radio modem technologies such as cellular digital packet data (CDPD), and wireless Ethernet (the
The wireless world seems to have far more than its fair share of acronyms, some of which sound very similar to others. The whole alphabet soup of wireless names and protocols can be baffling, to say the least so we have provided a quick overview of the most commonly used terms in wireless communication and a brief description of each. 802.11. 802.11 is actually a family of related specifications that are maintained by IEEE. The most popular specification in this family is 802.11b, also known as Wi-Fi. 802.11b operates in the 2.4-GHz band, which is an unlicensed ISM (industrial, scientific, and medical) band. In this case, "unlicensed" means that persons operating equipment that broadcasts and receives in this band do not need to obtain a license from the Federal Communications Commission (FCC). The 2.4-GHz band is used for a number of devices, not all of which are used for data communications but which do all impact the usability of the band. For example, other devices that operate in this band include microwave ovens, video/audio transceivers for home and business surveillance, Bluetooth wireless devices, and cordless telephones. Wi-Fi has become popular and inexpensive, which has also made it pervasive throughout many corporate infrastructures, homes, and coffee shops. 802.11b has a maximum bandwidth of 11 Mbps. Two other members of the 802.11 family are growing in popularity: 802.1 la and 802. llg. 802.llg also uses the 2.4-GHz band; however, it is capable of bandwidth up to 54 Mbps, owing mainly to its use of a more modern modulation scheme called OFDM (orthogonal frequency division multiplexing). 802.llg has some major advantages: It is fast, and it is backward compatible with the large 802.11b installed base. This makes it a logical choice for organizations that are not experiencing problems with interference on the 2.4-GHz band, and who want greater performance for some users without cutting off others who are using 802.11b. Nonetheless, the 2.4-GHz band is arguably crowded, and the relatively small choice of channels can complicate the design of large wireless installations. The 802.11b band is divided into a total of fourteen channels, but many of them overlap one another. Furthermore, not all channels are available in every country. In Japan, only one channel has been cleared for use; most of Europe can use channels 1 through 13; and in the United States, the FCC has authorized channels 1 through 11. In the United States, only the combination of channels 1, 6, and 11 provides a fully nonoverlapping complement of channels, so large-scale Wi-Fi networks are typically designed in cells, much like those used for cellular telephone antennae (hence the name "cell phone").
*We are obviously omitting several layers of the network layer model for this simplified example.
634
PEARROW
802.1 la, however, uses the 5-GHz band, which has a total of twelve nonoverlapping channels. The 5-GHz band is also substantially less cluttered (so far) with potential sources of interference. Yet, this amount of RF elbow room comes at a hefty price. Because radio wave propagation (the distance a signal can travel) decreases as frequency increases, the range of an 802.1 la system is significantly less than that of a Wi-Fi station. This means large-scale installation will need a lot more access points to get the same amount of coverage that a smaller number of 802.11b access points would permit. 5-GHz technology is also more costly than the 2.4-GHz technology. Suffice it to say, most experts currently believe that a combination of Wi-Fi and 802.11g is the right way to go for most organizations. GPRS. The general packet radio system (GPRS) is, as the name implies, a packet-switching system that makes efficient use of bandwidth by only tying it up when there are data to transmit or receive. Data are broken into "packets" and then sent out on the packet-switching network, which means that multiple users can use the resources simultaneously. Although the published maximum theoretical throughput of GPRS is about 170 kbps, it is unrealistic for a real-world application to ever reach this level. One of the major advantages of a packet-switched system such as GPRS is that it is "always on"; there is no need for the overhead of initiating a call, such as one would have to do in a circuit-switched network. GPRS is designed to run over GSM networks, but can in fact also run over the IS-136 time division multiple access (TDMA) networks that are frequently found in North and South America. CDPD. CDPD is a data transmission technology that uses cellular telephone frequency bands (800 MHz to 900 MHz) to transmit data in packet form. CDPD has a maximum throughput of 19.2 Kbps, and generally boasts a shorter call set-up than analog modems that use cellular phones for connectivity. CDPD is capable of employing forward error correction to reduce the impact of noise and signal loss on data transmission. It is also deployed over existing cellular networks, meaning the cost to implement is minimized. An end user can use a CDPD modem, often in the form of a PC card for a laptop or even Compact Flash format, to connect to a cellular carrier's CDPD network. The modem acts like any other network adapter so all transmission control protocol/Internet protocol (TCP/IP) applications can transparently work over the link. CDMA. Code division multiple access (CDMA) is a digital wireless technology that was pioneered and commercially developed by Qualcomm. It allows the multiplexing of several streams of conversation along the same frequency band. Unlike GSM and TDMA, each user is not assigned a different frequency. Instead, every channel uses the entire available band. Individual conversations are identified by a pseudorandom sequence. CDMA began as an effort by the Allies during World War II.
TDMA. TDMA is a very old technology that breaks individual conversations up into time chunks, allowing the multiplexing of many calls on a single frequency. TDMA is used by GSM. GSM. The GSM was first introduced in 1991, and today is the most widely used digital mobile telephone system in most parts of Europe and many other parts of the world. More than 120 countries currently use GSM as a primary technology for mobile digital communications. GSM uses a variation of TDMA, and operates in either the 900-MHz or 1,800-MHz frequency band. Many countries have roaming agreements with other countries, allowing a GSM user to roam freely among participating countries. When taken along with other complementary services such as GPRS and EDGE, GSM forms the core of a next-generation suite that features a variety of high-speed data services. 3GSM. 3GSM is a generic term used to refer to the nextgeneration of mobile communications systems. The current vision for 3GSM includes enhanced service for video, voice, and data, and will be globally available. WAP, WML, WMLSCRIPT. This suite consists of a transport protocol, a document markup language, and a scripting language that were all designed from the ground up to be deployed on net-enabled cell phones. WAP was designed with many of the characteristics of cellular phones and other wireless terminals in mind, such as less powerful CPUs, small amounts of memory, restricted power consumption, and smaller displays and input devices. WAP also takes into consideration the characteristics of wireless data networks, such as less bandwidth, more latency, less connection stability, and less predictable availability. WAP content is authored using the wireless markup language (WML). WAP content is compiled by a WAP gateway before transmission. WML is similar to HTML and their grandparent standard generalized markup language (SGML), in that it uses an array of markup tags to provide structure to textual documents. WML is a small language that consists of about 35 tags, around 14 of which have no analogs in HTML. Content is made up of individual pages of information that are called "cards," and many related cards are clustered together into units called "decks." Cards have an upper limit for size, which is approximately 1,400 bytes (very small compared with the average Web page size, which was 10 K as of July 2000* WMLScript is the programmatic portion of this family that allows for dynamic client-side content to be authored and deployed. WMLScript is somewhat syntactically similar to Javascript, although it is really a distinct language that has been optimized for wireless applications. Mobile IP Mobile IP attempts to solve a common problem that may be experienced by users of mobile networked applications. As a user travels from one access point to another, it is likely that he or she will be assigned a new IP address,
*http://www.cyveillance.com/web/newsroom/releases/2000/2000-07-10.htm.
34. Wireless Communication potentially on a completely different network. This is a problem because it makes persistent connections impossible, and interrupts TCP and user datagram protocol (UDP) sessions. Mobile IP uses a type of IP encapsulation that allows a mobile node to permanently keep an IP address on its home network, while transparently using whatever local network is present. When away from its home network, a "care-of address" is associated with the mobile node and reflects the mobile node's current point of attachment.* Bluetooth. Bluetooth is a radio-oriented wireless technology that is intended for use over small areas, primarily as a cable replacement between items such as phones and headsets, computers and printers, and other peripherals. Although Bluetooth operates in the same 2.4-GHz range that 802.11b and 802.11g operate in, it does not replace those wireless networking technologies. Bluetooth is intended to be complementary to wireless networking systems. Because both of these classes of device operate in the same heavily used chunk of RF spectrum, it is likely that there will be instances of interference in environments that use both heavily*
Operating Systems Palm OS. Palm Computing, Inc., released the Palm Pilot 1000 and 5000 in March 1996, in a technological climate that had weathered much disillusionment with handheld computing, owing largely to the unfortunate blunders in marketing the overhyped Newton Message Pad. The Palm Pilot, soon to have the word "Pilot" dropped from its name due to legal infringement on the Pilot Pen company's trademark, had a different approach to handwriting recognition than the Message Pad. It required that the user learn a glyphic alphabet that would allow the handwriting recognition software to use constraints as an aid to letter recognition. Thus, less processing power and software code would need to be dedicated to this task. This power came standard with the Palm's onboard operating system, the Palm operating system (OS). The philosophy behind the Palm OS was also different than those found in other attempts at portable computing. Instead of trying to create a miniaturized version of the desktop experience, the Palm OS created a unique and optimized experience for mobile usage. The Palm OS was designed from the ground up to run on a wide variety of architectures, meaning that manufacturers had greater flexibility in designing systems that would have the Palm OS in its core. Other OSs typically tied the manufacturer to only one or two architectures. Another distinguishing feature of the Palm OS was its focus on usability. For example, details such as the Calendar application always opening up to the current day and hiding unused time blocks to reduce scrolling add up over the course of a day of actual usage by real users to many minutes worth of saved frustration and time. More recently, the Palm OS has shown up in a new market— the cell phone or "smartphone." Several manufacturers have
635
included the Palm OS in their cell phones. However, it is important to note that the Palm OS was not originally designed with cellular telephony in mind, and it may need to do some considerable mutation to keep up with other cellular telephony-savvy operating systems. Symbian OS. Symbian is a software licensing company that was created in June 1998, and is jointly owned by Ericsson, Nokia, Panasonic, Motorola, Psion, Siemens, and Sony Ericsson. Headquartered in the United Kingdom, Symbian has offices in Japan, Sweden, the United Kingdom, and the United States. Symbian is the supplier of the Symbian operating system—an open OS specifically designed for mobile data-enabled phones. The first Symbian-equipped telephone was the Nokia 9210 Communicator, which was released in the first half of 2001. The Symbian OS really differs significantly from other OSs that have been repurposed from OSs for other types of devices into a telephone OS. Symbian has been optimized from the ground up for mobile systems that have limited resources and sporadic connectivity. Code reuse is a critical part of the Symbian design. The C++ programming language is used for all Symbian code from the kernel level upward. Because Symbian was designed for the mobile phone industry, which relies heavily on telephone "personality" to target customers, the OS is highly customizable, allowing developers to reach inside the Symbian user interface code—which is modularized from the core operating system—in order to customize and personalize it for their own device. This level of customization is not possible with the Palm OS or any Microsoft OS. Most OS vendors believe that consistency across all devices is important for branding and usability. However, the mobile phone market is different from the desktop and notebook PC market; users want fashionable phones, or phones with a particular "feel" that reflects their own personalities and lifestyles. Hence, Symbian has a significant advantage over other inflexible options. Symbian is designed so the network layer is sufficiently abstracted to make transition from one type of underlying wireless technology to another seamless. Wireless connectivity connection quality usually varies as a user moves from location to location, probably occasionally losing connectivity altogether for variable amounts of time. Symbian OS is robust against such variability, and instead of taking a "thin client" approach to design, empowers the client side as much as practical, given the resource constraints of mobile devices. Series 60 is a platform that "rides" on top of the Symbian OS. It provides developers an environment in which to create and run applications that use key telephony and personal information management, browsing, messaging clients, and a complete and modifiable user interface. Series 60 developers have access to a rich set of developments tools, such as the Series 60 Software Development Kit, which includes libraries and application programming interfaces (APIs) needed to create content for Java, C++, Browsing (XHTML), and MMS.
*http://www.ietf.org/rfc/rfc3344.txt. *However, experts have recently proposed that RF interference is really a myth, and that the phenomenon is just an artifact of our current, very antiquated transceiver technology. See http://www.salon.com/tech/feature/2003/03/12/spectrum/print.html for more info.
636
PEARROW
Microsoft PocketPC. PocketPC began life in the Ml of 1996 under a different name: Microsoft Windows CE, short for Compact Edition. The name was changed as of version 3.0, and this name is also used to describe a platform and class of mobile device. For example, the Compaq iPaq PocketPCs run the Microsoft PocketPC OS. The philosophy behind PocketPC was very different from the simplified, Spartan approach of the Palm OS. Rather than being a cut-down version of another Windows product, Windows CE was completely written from scratch, and was based around six core functionality modules. Even though the first-generation mobile products that ran Windows CE relied heavily on host computers—which effectively reduced them to "PC Companions"—they were designed to offer more functionality than Palm OS devices while maintaining the familiar Windows look and feel.
WIRELESS WEB BROWSERS One of the primary areas of focus for Web usability experts has traditionally been browser traits. A side effect of the browser wars of the late 1990s was the disparity in feature sets between the most popular browsers. In fact, some gross incompatibilities between the "Big Two" arose in the scramble for ultimate Web dominance. As a result, designers and usability experts have had to allocate resources to making cross-browser compatible content. There have been two approaches that have enabled widely compatible content: the minimalist approach, in which only the lowest common denominator is used, and the Black Magic approach, in which many "browser sniffing" optimizations are rendered in a variety of scripting languages, including Javascript, PHP, and VBScript. It is likely that wireless content developers will again divide into these two camps, but both approaches can be used eclectically in practice. In this section, we examine some of the most popular wireless Web browsers in detail, taking note of the traits, capabilities, and quirks of each. Note that much of this information is likely to change by the time you read it. New browsers are likely to emerge, and some of these may disappear entirely. Be sure to check on updates, and keep in mind that the information contained herein should serve as an outline for your own feature inspection, and not a definitive guide to browser features.
Handspring Blazer General Info. This Web browser is commonly found on wireless-equipped Handspring devices such as the Treo 300. Blazer supports sites written in HTML, WAP (WML/HDML), and cHTML, so a wide range of content is available to users. Blazer is not free, however; at the time of writing, it was available from Handspring at a cost of $20 USD. Blazer requires a device with at least 8 mb of memory to run.
Capabilities • Support for multiple markup languages, including HTML, WAP (WML/HDML), cHTML (I-Mode), and xHTML
• Quick access to most Web pages: Most pages display in 5 seconds or less • Text and images are transformed to fit the screen size and display capabilities of the wireless device (although this can be unpredictable at times) • Support for 16-bit images (65,500 colors) when supported by the wireless device • 128-bit security [secure sockets layer (SSL)] • Full support for cookies • Robust bookmarking support, including the capability to file bookmarks • Support for HTTP proxy servers Known Limitations. Blazer Version 2.0 is known to have a quirky data caching mechanism that can cause usability issues with dynamic content, such as stock quotes, weather information, and other real-time sites. Some users have noted that certain sites simply do not work at all with Blazer, most notably Microsoft's Hotmail Web site. Sites that are not optimized for handhelds can still be viewed with Blazer because it transforms content to fit, although this behavior can make content very hard to read. Finally, several reviews of Blazer have commented about its lack of capability for the Web Clipping feature that is part of the Palm platform. It is important to note that Blazer does not support any kind of media plug-in, nor does it support Java, Javascript, file transfer protocol (FTP), or file downloads. The Blazer proxy server will strip out unsupported content, but the remaining content will only be usable if the unsupported element is not key to the core functionality of the page. This means that Flash-heavy Web sites will be largely unavailable to Blazer users, as will sites that rely heavily on Javascript and DHTML (dynamic menus, for example, will not be available).
Eudora Web Summary. Many PC and Mac users are familiar with the Eudora e-mail client, and the Eudora Internet Suite, which includes the Eudora Web browser, is an extension of this product into the wireless world. The Web browser and its parent suite are available for download for free from Eudora's Web site (www.eudora.com). Eudora Web takes a different approach from many of the other wireless Web browsers. It uses no proxy server and instead connects to remote sites directly through HTTP. This means that existing HTML content on existing HTTP servers can be viewed directly with this client, with no translation required between the two endpoints.
Capabilities • True HTTP/HTML support—no proxy server needed 0 Standard HTML markup such as titles, lists, and links 0 Forms, including all editable fields, buttons, drop-down lists, check boxes, and more 0 Cookies 0 User/password authentication
34. Wireless Communication 0
• • • •
Although HTML frames cannot be displayed, a means for navigating frames is provided Support for https uniform resource locators (URLs) with strong cryptography (SSLv2, SSLv3, and TLSvl) Ability to view server certificate properties (validity dates, issuer, md5 fingerprint, etc.) Ability to override some SSL trust errors Ability to view root certificates
Known Limitations. Eudora Web does not currently support multimedia plugins or client-side scripting (Javascript, etc.).
OpenWave Mobile Browser The OpenWave Mobile Browser is available for many mobile devices. OpenWave has one of the most comprehensive and useful developer Web sites of any wireless/mobile vendors, which is a major reason to consider designing for their platform.*
• Color display: Via the image proxy server, image colors are convered to Web-safe colors. • Frames: A Web site with frames is indicated with a frame icon in the top right corner of the screen, which enables you to choose which frame to view. • Javascript support: Supports most Javascript objects, methods, and properties. A full listing is at http://www.ilinx.co.jp/ en/products/xiino/jscript.html. Note limitations in the following section. • Cookies: Supports cookies. • 128-bit SSL support Known Limitations. There are several core parts of the Javascript language that are not supported, most notably, the for... in control structure. Several HTML tags that normally create a particular formatting effect are implemented without these effects. For example, the tag does not actually make the text appear any different. This is not a bad thing itself, but it could be surprising to designers who incorrectly use these sorts of tags for visual effects.
Capabilities • • • • • • •
XHTML and CSS mobile support CHTML/iMode support Supports off-line viewing (snapshots) Autofill feature for forms Support for GIF, animated GIF, PNG, JPG, and BMP Support for several multimedia formats (audio and video) Full SSL 3.0/ TLS 1.0 support
Known Limitations. No major limitations beyond typical limitations of mobile-size devices (screen size, etc.). Xiino Summary. Xiino is the current incarnation of Palmscape, an older Web browser for PDAs. This is a commercial product that is produced by ILINX, a Japanese company. Xiino is a fairly robust application that integrates into a whole suite of other PDA applications, such as e-mail clients and file compression utilities.
Capabilities • Character display: Font size is user adjustable. • Image display: Image sizes are automatically adjusted for display on the PDA. • Table format: Tables are reformatted to fit the screen.
637
USAGE Usage Trends Do people use their mobile computing devices the same way that they use their desktop computers? The preliminary answer certainly seems to be "no." In an informal study of dozens of PDA usage case studies, we discovered that the majority of the case studies detailed usage modes that could not have been easily performed with traditional desktop PCs. As the name mobile computing suggests, portable computing devices are often used in environments where stationary PCs are not feasible. Desktop PCs are typically used for traditional, extended-session workflow patterns that are associated with tasks such as software development, data entry, accounting, and research. However, portable computing devices are used for a variety of reasons. The most common reasons that organizations deploy portable computing devices include the following (Pearrow, 2003): • Decreased need for training—simpler handheld systems are typically easier to use and learn than desktop PC systems • Rugged—PDAs can be equipped with rugged, industrialstrength housing for less money than traditional PCs or laptops • Inexpensive—PDAs are less of a liability if they are lost, damaged, or destroyed • Mobility—PDA use has grown in the areas of health care and law enforcement because wireless computing enables workers to make rounds in the field and still gain access to info via wireless networking
*A document that is worth reading, but that is obviously a "marketing white paper" (one that an "independent lab" is paid to write for a customer that favors their product in some way) is at http://www.openwave.com/docs/products/resources/mobile_browser_review_2003.pdf. It might give the reader some good ideas about what features are worth comparing between various browsers.
638
•
PEARROW
PDAs are finding their way into more and more locations where PCs would simply be infeasible due to their size. For example, after the September 11 attacks on the World Trade Center (WTC) in New York City, Chief Joseph Pfleffer, the battalion chief who was in charge of the fire department's planning and strategy at the World Trade Center site, made the decision to use PDA devices that were equipped with global positioning system (GPS) functionality to help determine the exact location of pieces of evidence at the WTC site.* Park rangers are using PDAs to report safety hazards in public parks and to arrange for their fast repair* Clearly, portable computing has gained a footing in mainstream use, and so has transcended its status as a toy for wealthy businesspeople. It is crucial for any usability expert to understand usage trends and habits of users because there is no other way to design usable systems. In this section, we briefly survey some aspects of mobile computing use to provide a framework for the reader to explore his or her own users' needs.
Common Wireless Sites What sorts of mobile Web sites are people using? Not a lot of data existed at the time of writing about just what the most popular wireless Web sites were, but one anecdotal metric that can be used is to measure how many sites show up in popular search/categorization Web portals, and which categories seem to be the most popular. Here we list a snapshot in time of the most popular site types. The following numbers were taken from Web searches performed on March 23, 2003. The numbers in parentheses represent the total number of listings in each category. They have been listed here in descending order of total listings: www.yahoo.com Regional* (559) Entertainment and Arts (208) Commerce (137) Society and Culture (120) Travel and Transportation (109) Sports and Outdoors (74) Communication (64) Portals (56) Science and Technology (55) Reference Tools (47) News (39) Finance (33) Health (24) Education (11) www.fonethe.net Travel (10) Entertainment (6) Sport (6) Shop (5)
Finance (4) Games (3) Humor (3) News (3) Music (1) TV (1) Weather (1) Tvww.wireless.co.il Portal (21) News (14) Communication (13) Finance (11) Travel (7) Entertainment (6) Search (5) Sport (5) Weather (3) Leisure (2) An empirical analysis of the most commonly used classes of mobile Web sites is in order, but from this anecdotal data we can infer that current mobile computing network use revolves around entertainment, travel, news,* and communication. However, it is hard to derive any truly meaningful interpretation from these data. In fact, as the face of mobile and pervasive computing shifts daily, it will be difficult for any usability expert to keep pace with the wireless community as a whole. The most important tools for usability experts within the area of mobile computing usability will be frequent interaction with the user community and iterative redesign.
Social Acceptance of Technology A relatively new phenomenon is visible in the mobile computing industry: the "coolness" factor. Computers have generally been regarded as unglamorous, utilitarian devices that sit under a desk. Cubicle denizens have long dealt with the look of ugly beige by adorning their terminals with personal artifacts. Before the Apple Mac™ came onto the computing scene, the computing industry was not generally regarded as one that dealt in high fashion. This has changed more recently, however, as evidenced by a disappearance of beige, and an emergence of bright candy colors, brushed metal, sleek modern designs, and even personalizable computer cases. Fashion sense has also caught on in a major way in the cellular phone market. There are countless vendors of cell phone accessories that are designed to bring a personal touch to one's otherwise utilitarian device. Cell phone vendors offer scores of models that are designed to not just work, but to appeal to the aesthetic sense of users. Siemens, Inc., has launched a line of fashion phones under a new, distinct trade name, XELIBRI. According to President of XELIBRI, George Appling, "XELIBRI products are fashion accessories that make phone calls."*
*http://www.symbol.com/australia/news/btl_2002_01_wtc.html. *http://www.ute.com/home/solutions/portage.pdf. *It is important to note that this category encapsulates entries from all the other categories, but with the additional locale-specific information. *It is possible that the increase in wireless news site presence owes to the more recent military action taken by the United States in Iraq. *http://cellular-news.com/story/8196.shtml
34. Wireless Communication Why should a usability expert care if a phone is fashionable? As in any field of design, form often comes at the expense of function. If usability issues caused by browser incompatibilities, version-related quirks, scripting engines, and flaky connectivity do not provide enough of a challenge, usability experts will now have a new factor to consider—phones and devices that look great, but at the expense of being intuitive or even usable at all. Clearly, there is not a dichotomous relationship between looks and intellect—savvy designers and detail-oriented usability experts can work together using iterative redesign to create the right combination, if both sides respect the work of the other and are reasonable in making compromises. Beyond the immediate scope of whether a device looks "cool" or not, however, is the social propriety of the device, and, more important, its use. Technologies that have long-lasting and widespread social impact are usually plagued with a period of emerging etiquette and public standards of appropriateness before that technology is accepted by the mainstream. To date, there are many different viewpoints that cellular phone advocates and critics might have regarding the use of telephony in public. Palen, Salzman, and Young (2000) followed a group of new cellular service subscribers over the first 6 weeks of their usage. During this study, the researchers charted the behaviors of the participants over several areas of interest, including their rapidly changing perceptions of the social appropriateness of public cell phone use. Most important, they note that "mobile phones are devices that directly serve the individuals who employ them, but their use is influenced by social context" (Palen et al., 2000). Clearly, mobile computing usability must consider these new social aspects when considering modes of use. Factors that lie outside the immediate sphere of contact of the user may well affect usability in ways that are so far poorly studied.
Problems With Mobile Computing Mobile computing is still in its infancy, and as with any new technology, there are many bugs that need to be worked out. Because most wireless mobile computing devices are priced out of the comfort range for novice or occasional users, they have thus far been limited to the domain of wealthy, relatively techsawy professionals. It has been said that wireless PDAs and the like are really just toys for company execs. This is a common trend with new technology: Early adopters are usually people who can afford to drop a lot of money on something that may or may not be very useful (Norman, 1999). The early adopters have an impact, however, on the shape of things to come, as they struggle with half-baked features or real usability problems. User I/O. The PC industry has been riding Moore's law through some massive waves of new and vastly faster CPUs since the mid-1990s. Advances in software techniques and processor speed have allowed the creation of artificial vision, the sequencing of the human genome, and even the creation of some really incredible video games. Amazingly, however, the dominant user input interface to all this technology is the ancient typewriter keyboard—the same mechanical interface that was deliberated designed to slow human down so the typebars would not stick
639
to each other. Even though other keyboard layouts that are optimized for speed have evolved since the invention of the wholly electronic keyboard, the QWERTY keyboard still dominates as the most popular input method. It is bad enough that humans must still transform symbolic language into words and words into musculoskeletal reactions that drive keyswitches, which in turn produce bit patterns, just to instant message a friend about lunch plans. Mobile users suffer more, though, because really usable keyboards cannot be tiny, despite various valiant efforts to engineer a new solution. Over the years, many inventors and manufacturers have attempted to solve the bulky keyboard problem in various ways. IBM took a stab at the problem with their Butterfly, an expanding keyboard that came on the 701 Series ThinkPad laptops. However, the novelty of the Butterfly wore off when LCD panels that were the full width of a full-size keyboard became commercially viable—there was no longer a need to shrink the keys down to the size of the previously tiny LCD displays. An LCD panel cannot fold up per se, so the immediate usefulness of the Butterfly waned. Similar approaches have been taken by companies such as Pocketop, Inc., who makes a fold-up keyboard that uses the Palm platform's built-in infrared capability for connectivity. Still, a fold-up keyboard is at best an awkward compromise that doubles the size of many PDAs, while not necessarily providing the same ergonomics as a regular keyboard. The arrival of computer-like functionality in the cell phone form factor meant that designers had to find a way to incorporate a complete alphanumeric keyboard into such a tiny package. Some approaches put the keyboard on the back of the phone— not a very intuitive or easy-to-use design. Other approaches made use of the existing letter mapping on every telephone keypad, but this meant that user had to press a key four times, in some cases, to register a single character. The Fastap Keypad features a unique approach by David Levy, founder and CTO of DigitWireless. This keypad uses the spaces between the keys, along with a simple algorithm to smooth out errors in actuation. The result is a full alpha keypad in the same space as a regular keyboard. Other companies have created small "thumb keyboards" that require the user to type using only the thumbs. These pack a lot of keys into a small space, and unfortunately this makes the keys difficult to press without error. Fitts's law, a model of human psychomotor behavior, predicts that movement time to a fixed target is a logarithmic function of target size. This relationship can be expressed by:
where MT= movement time, a,b = regression coefficients, D = distance of movement from start to target center, and W = width of the target. What that means is that the bigger the button, the less time it takes a person to make the mental and physical corrections to actuate it. Thus, smaller buttons mean the user must devote either longer time to type, or a lot more mental cycles to do the corrections. Unfortunately, Fitts's law only really considers movement in one dimension; it probably does not quite accommodate movements that require lateral movement. It is still likely, however, that even a representation that accounts for
640
PEARROW
lateral movement would have the same basic outcome: Smaller keys are harder to hit. Small keys also mean that the user needs good eyesight—not always a luxury that older user have. One other disadvantage of small keys is that they are often designed in such a way as to require large amounts of physical force to actuate, which can contribute to repetitive strain injury. Handwriting recognition was once believed to be the input method of the future. The dismal failure of Apple's Newton MessagePad—which was really caused by marketing miscommunications, rather than poor engineering—set back the faith in handwriting recognition several years. It was not until Palm, Inc., popularized handwriting input with its constrained Graffiti™ system that users began to put serious stock in this input method. Handwriting recognition, especially the kind deployed in the Palm OS, is lightweight and allows data input in a small amount of space. However, it is difficult to attain high rates of data entry, even for veteran Graffiti users. Fleetwod et al. (in press) found that veteran Graffiti users could enter an average of 21 words per minute using a stylus and Graffiti, whereas novice user could only manage an average of 7 words per minute. Bailey (1996) found that normal pen-and-paper handwriting can yield about 31 words per minute, so there is still likely a gap in productivity with even veteran Graffiti writers. Nomadism. The utility of a portable computing device is severely reduced if it is not properly suited for dealing with the variances in signal and network availability; regional differences in time, currency, and other standards; and other external variables that can change as a user moves from location to location. Nomadism is the quality of a portable computing device that allows it to deal gracefully and transparently with these variations, insulating the user from needing to know or from making manual adjustments. According to Deleuze and Guattari (1986), "Nomadic space is smooth, without features, undifferentiated from other spaces. Nomadology itself is a line of flight, a process which constantly resists the sedentary and the fixed." However, in art and technology alike, this concept of smooth, transitionless space is hard to realize. Wireless networks fade, cross-talk, interfere, and disappear; cellular networks are sparse, unreliable, and data service is not available in all areas where voice service is. Systems that are even aware of their crude location in the real world, much less systems that take full advantage of this awareness, are still uncommon. Furthermore, systems that do exploit this awareness do not often do so in a way that is usable. For example, at the time of writing, a wireless service was available over the Sprint network* that purported to tell the current weather for the user's locale. However, the service did not actually tell the user where it believed the current locale to be. Furthermore, the information was frequently cached, which was totally nonobvious; there were no icons or status messages to let the user know that the information was just a cached copy. This unaware caching led to an apparent failure of the system
to accurately tell the weather. However, a forced reload of the information from the server generally yielded a report that was close to the current locale's real weather. This behavior clearly violates some of the basic heuristics of usability: Visible system status was not provided. Global Coordinates: GPS. A variety of devices and technologies exist today for general location-determining functionality. In the extreme large-scale arena, system designers can use GPS for coarse location awareness. The GPS has a resolution that varies depending on the type of chip that is used to receive the GPS signal. The most common kind of receiver—the sort found in most commercial devices—uses what is known as C/A-code, which has a resolution of about anywhere between 3 and 300 meters, but appears to be about 100 meters on the average. More expensive models of receivers can use P-code, which can give resolution down to 30 centimeters. A third "flavor" of GPS exists, know as Y-code, which is used exclusively by the military, and incorporates encryption and antispoofing techniques designed to ensure only authorized personnel are able to receive exact data in certain restricted areas. Development kits for GPS systems are available off the shelf. At the time of writing, single-chip and multichip GPS receivers were being produced by a wide variety of manufacturers, including analog devices, Garmin, Motorola, Phillips, and Sony. Garmin, Inc., has also more recently released the iQue 3600, the world's first PDA to integrate GPS technology.* It is likely that this sort of technology will become commonplace in portable computing devices such as PDAs. NTT DoCoMo announced on March 28, 2003 that they would be shipping the world's first GPS-enabled cellular telephone in April 2003. The F66li telephone features full GPS support that can be integrated into services beyond 911 and emergency use. For example, this GPS telephone could be used by parents to keep track of their children by using a service similar to the Wherify GPS Person Locator.* Figure 34.2 depicts the Whereify GPS watch, which incorporates a tiny GPS device used for tracking the location of a child. The availability of such powerful technology has serious implications for privacy and security of users. The ethical considerations and sociotechnical issues that are imminent due to this sort of capability should be addressed by organizations outside the manufacturing world.
Space Aware Versus Location Aware: Cricket Beacons. The GPS does an excellent job of telling a user just where in the world he or she is located. However, GPS is oblivious to any contextual details about the space a user is in. GPS cannot provide details about the type of room a user is in, or notify the user when he or she has moved out of one logical space in an office building and into another. To address these shortcomings, the MIT Laboratory for Computer Science and Artificial Intelligence Laboratory have collaborated on a device
*As well as possibly others, but we only had experience with the service via Sprint. *See http://www.garmin.com/products/iQue3600/ for more info. * http:/www.wherify.com.
34. Wireless Communication
FIGURE 34.2. Note. Copyright © Whereify 2003. Reprinted with permission. known as a "cricket beacon," which is a small, self-contained radiofrequency beacon. The cricket is not tied to any particular architecture or OS; the Cricket Listener, which is the client-side application that receives cricket broadcasts, runs in the Java Virtual Machine (Priyantha, Chakraborty & Balakrishnan, 2000). Unlike GPS, which relies on centralized management of satellites, cricket beacons are managed in a decentralized fashion. They are "glued" together in an experimental framework called Metaglue, which is used to define "societies" of beacons (Coen et al., 1999).
Connectivity and Bandwidth Typically, hard-wired networks do not become unavailable and available sporadically: They stay "on," with little variation in quality of service over time.* Wireless networks, however, are entirely different. The quality of network connectivity in a wireless environment can depend on a number of factors: signal strength, signal-to-noise ratio, interference, environmental RF reflectivity and absorption characteristics, and network congestion can all play an important role in the end user experience. In addition, current wireless technology provides only a fraction of the bandwidth that wired networks provide. The very nature of wireless connectivity presents a new class of challenges to designers and usability experts. It is important to account for all
641
these factors when designing a system if the goal is to present a robust and transparent interface to the user (La Porta, Sabnani, & Gitlin, 1996). Because there is a globalwide deployment of the TCP/IP suite, it is desirable that wireless applications be able to use this same suite, rather than relying on new protocols that are specifically tuned for wireless networking. Developers are familiar with the TCP/IP stack, many software code libraries already exist for it, and most popular applications are built for it. However, TCP/IP does not natively deal gracefully with mobile connections; network conversations are expected to happen between two endpoints with relatively fixed network addresses (CITE rfc 793). Mobile IP, which we defined previously, is a relatively new solution to this problem (CITE rfc 2002). At the time of writing, several major cellular carriers in the United States offer mobile IP support through CDPD service. However, due to a number of roadblocks, most notably security issues, mobile IP is not currently widely deployed in mass market wireless devices.* Several strategies have been suggested for dealing with the sporadic nature of wireless connectivity. Chang et al. (1997) proposed that asynchronous request/response handling can improve the user interaction by caching requests and responses when they cannot be completed so they can complete when connectivity has been restored. They have also suggested that intelligent handling of content caching can effectively mask the intermittent nature of the wireless network, although there are several problems inherent in the caching of dynamic content and form data. Jing, Helal, and Elmagarmid (1999) proposed several new paradigms for client-server interaction in mobile computing scenarios. In particular, they have discussed mobile-aware adaptation, by which a system can adjust to the changing needs of a mobile client and environmental variability. A typical component of such mobile-aware adaptation is a Web proxy system, often implemented as an HTTP agent on the client side, and an HTTP proxy on the server side. Web Express is just such a system (Housel & Lindquist, 1996). It consists of the client-side intercept application and the server-side intercept application, which work in tandem to effect a reduction in data volume and latency of wireless communications by intercepting the HTTP data stream and performing various optimizations, including file caching, forms differencing, protocol reduction, and the elimination of redundant HTTP header transmission (Housel & Lindquist, 1998).
Security, Safety, and Privacy Over the years, many experts have suggested that increased exposure to RF energy may have adverse health effects on humans. Modern electronic equipment produces RF at higher frequencies than older technologies, and in general, the higher the frequency, the more chance there is that the RF can interact with human tissue. The IEEE has developed a guide regarding the
* This is assuming that the local network administrator has done a competent job of prioritizing traffic in environments that tend to be congested. *http://www.birdstep.com/collaterals/wp_mip_deploy.pdf.
642
PEARROW
safe use of RF devices, which has been published by the American National Standards Institute as publication C-95.1-1991. This document covers the safe and acceptable use of nonionizing radiation. Virtually all modern equipment produces RF levels that are far below these published safe levels of radiation; such low levels of energy are typically passed through the human body with no adverse effects because the amount of energy absorbed is extremely low. Nonetheless, there has been much controversy over the use of cellular telephones and their possible link to cancer that results from the heating of tissue by RF waves that radiate from cellular antennae. The overwhelming results of studies performed by health experts and scientists the world over show that there is no correlation between cell phone use and brain tumors. One frequently cited study supposedly shows a correlation between the side of the head that a cell phone user holds a phone to and the use of analog cell phones. However, the correlation is not statistically significant, and relies on participants to self-report data that might have been invalid. It is likely that the issue of cell phone safety will resurface sporadically over time. The topic of computing related health issues is covered more thoroughly in Bertus and Bertus (chap. 37, this volume). Radiation worries aside, portable connectivity means that there are other sorts of risks to consider. Cell phone safety has been a hot topic for many years now, and much data are available now that indicate that even the use of hands-free headsets does not necessarily mean that cell phone usage while driving is safer than without headsets. It appears that diverting attention between two tasks that require a high degree of cognitive processing results in a substantial decrease in the performance of both tasks Gust et al., 2001). These findings should prompt designers and users of portable computing to consider the sorts of risks that might be involved in increased availability of information—information that likely will require at least a minimal level of cognitive processing—in everyday situations. Ergonomic constraints that are typical of current-day portable computing devices are also sources of safety concerns for users. For example, cramped keyboards are typical in many PDAs, even if they have on-board keyboards. In the interest of preserving as much device real estate for display area, manufacturers have trimmed away at the most popular input device, which has mutated into a tiny grid of miniature pill-like keys. Devices such as the RIM BlackBerry interactive communications device have popularized the thumb keyboard. However, it is not clear yet if these keyboards pose any sort of hazard for frequent users; there simply has not been any research on the long-term ergonomics of such devices. Uncomfortable positioning is a de facto issue that vexes many types of portable computing devices. This is often the result of trying to cram many user interface (UI) elements into a small package. Tiny keyboards are part of the problem, but the overall shape and fit of a portable device can vary based on potentially nonusability-related variables such as current fashion trends (colors, shapes, interchangeable/customizable faceplates, etc.) and augmentation for industrial applications (rubberized jacketing or metal case encapsulation, high-contrast screens, headsets, etc.).
FIGURE 34.3. The FasTap keypad. Note. Photo reprinted with permission. Copyright 2003 Digit Wireless.
There have been many attempts to address the issues of user I/O with small devices. Some solutions, like the T9 text entry system, use predictive intelligence or probability functions to accelerate text entry on unmodified keypads such as the standard telephone keypad. Other solutions, like the FasTap keypad from Digit Wireless, put more UI elements into the same amount of space. Figure 34.3 shows the FasTap keypad, which has a full QWERTY keyboard in the space of a normal keypad. One novel approach to the problem comes from Canesta, Inc. They have designed a system that projects a virtual keyboard onto any flat surface, then uses artificial intelligence vision processing to translate a user's "key presses" into actual text on the device. The entire Canesta virtual keyboard system consists of several pieces, including the keyboard pattern generator (a red visible laser that is projected through a patterned reticle), an invisible laser scanning system, and a recognition system. All parts have been designed to work with a minimal amount of resources, and the entire system is small and cheap enough to embed in most portable computing devices. Figure 34.4 shows an artist's conception of the system in action. System integrators
34. Wireless Communication
643
Copyright 2002 Canesta, Inc. FIGURE 34.4. Canesta virtual keyboard. Note. Photo reprinted with permission. Copyright 2003 Canesta. can take advantage of the system's API to register key-click sounds—or other indicators—to help the user know when a key press has been registered. Although the number of alternate input technologies seems to be growing monthly, it is not clear that any one solution is the panacea for the problem of mobile data input. Usability experts and designers will need to take an eclectic approach that synthesizes the best solutions at the moment. Small screens are another fact of life for mobile device users. The literature is full of research into this area because the problem of small screens is not new. Early research from the 1980s into the usability issues in small CRT displays still has relevance today. Much work was done during this time to study the usability issues of small screen devices such as those found in ATMs (automatic teller machines), typewriters, and photocopiers (Buchanan et al., 2001). They have noted that small screen size itself does not constitute a usability problem; rather, problems occur when content is inappropriately constructed for such displays. They use the analogy of the Post-It™note. A Post-It™note is a small piece of paper, but it is perfect for delivering a terse, highly focused message. However, if someone were to try to write a letter on one, the result would be a miserable failure (Buchanan et al., 2001). Trevor, Hilbert, Schilit, and Koh (2001) proposed a technique for optimizing displays with three to ten lines of text that uses a split-mode model. Navigation and acting are separated, creating
in place a navigation interface and an action interface. Kamba, Elson, Harpold, Stamper, and Sukaviriya (1996) suggested a technique that employs transparency in on-screen widgets to allow for more information in the same amount of space. They reason that because on-screen controls can take up substantial amounts of screen real estate, control widgets are appropriate targets for optimization. They also note that gestural controls, such as those available on the Apple Newton Message Pad, are efficient, but require the user to learn control gestures before being able to efficiently use the device (Kamba et al., 1996). This is an example of the trade-off between "knowledge in the head" and "knowledge in the world" (Norman, 1990). DESIGN IMPLICATIONS The alphabet soup of wireless technology can be baffling to developers and users alike. A usability expert must ask, at the end of it all, "How do I ensure usability?" The answer is not particularly straightforward. Devices change constantly; they can go out of fashion, their underlying technology might become obsolete, or the sole company that produces them might succumb to DotCombustion. Whatever the case, it is not practical to try to lay out guidelines that are applicable to a particular sort of browser, device, or even a particular technology, simply because such myopic approaches have no
644
PEARROW
longevity. Conversely, attempts to suggest generic, extensible guidelines that can survive multiple generations of devices might be regarded as "abstract emptiness"; they might not appeal to many people's affinity for easy-to-digest, quick-fix cookbook recipes. This is a dilemma for the designer, the usability expert, and the author of research articles on the topic of wireless usability!
Design Guidelines Usability experts have generally found that a heuristic approach to usability is cost and resource effective, but still helps dramatically eliminate the majority of usability problems. The wellknown list often usability heuristics published by Jakob Nielsen* has shown remarkable longevity given the mercurial nature of technology. To date, most usability issues in any interactive information content, whether wireless or wired, can be traced back to one or more of these ten basic heuristics. Although heuristic evaluation and usability testing are invaluable for ferreting out usability issues in existing systems, it is also helpful to have a list of guidelines and best practices for use during the initial design phase. Such guidelines can also be used retrospectively to analyze the usability of existing systems. In this section, we present an overview of the design guidelines that have been distilled from dozens of research articles. Some guidelines are very similar to well-known guidelines for traditional, "wired" Web content; others are specific to wireless devices. Flach, Bennett, Stappers and Saakes (chap. 22, this volume) provide more indepth coverage of Web usability evaluation. Some readers will note that these guidelines stem from the same basic principle, parsimony, or as it has been called by some, "abstract less-is-more emptiness." Alas, there is nothing glamorous or exciting about designing utilitarian content that works for its intended purpose. 1. Remember, the wireless Web is not a miniature version of the wired Web: Users have a different interaction paradigm when they are "on the go", and wireless content must reflect this shift in usage to be usable. Organizations that simply try to port their wired content to wireless will probably fail, and frustrate users in the process. 2. Make content accessible within one or two links: Research shows that multiple link hops have a significant degrading effect on usability (Buchanan et al., 2001). Deeply linked sites are likely to cause usability problems. 3. Keep relevant content "above the fold": Scrolling can be even more problematic on wireless devices than on traditional Web-browsing setups. It is important to note that horizontal scrolling seems to be less problematic than vertical scrolling (CITE). 4. Allow for user personalization and prefllled forms whenever possible: If the user is likely to type the same information into fields with similar names over many sessions, there should be a way to automate and "remember" information, *http://www.useit.com/papers/heuristic/heuristic_list.html.
at the user's discretion. Note that saved passwords can constitute a major security issue if a device is stolen or lost. 5. Use graphics sparingly: Fewer graphics equate to less download time. Use an image only if it communicates essential information, and even then consider optimizing for the small screen. One-bit (black-and-white) images take up a fraction of the size of the same images in full color, yet color is not typically useful in communicating crucial information. 6. Use highly focused, short pages: The task-oriented nature of wireless access means that people are less likely to need long discourses on a topic, but more likely to need concise and focused information. Use the inverted pyramid style of writing if content must go long for some reason. 7. Do not rely on advanced features: Web sites that rely on client-side scripting languages, multimedia plug-ins, or frame contents will suffer or be entirely unusable with many wireless devices. 8. Minimize necessary keystrokes: Consider replacing text entry with pull-down menus, use predictive techniques, or even use sensible default values for forms. 9. Use the handheld-friendly meta tag in all your pages: Although just including this tag does not magically make your content work with handheld devices, it does identify content as being optimized for small screen devices: 10. Avoid tables that are over 150 pixels wide: Remember that most Palm OS devices have screens that are 160 pixels by 160 pixels, and with screen widgets (scroll bars, etc.) at least 10 pixels will get used. 11. Use handheld-optimized images: Consider posterizing graphics to just 4 or 16 shades of gray. 12. Do not use tiny fonts: Just because you are creating content for a small device, you do not need to use tiny fonts. Most browsers automatically scale fonts for the device. 13. View your content in as many platforms as possible: If your organization cannot afford one of each new device, use an emulator. Most vendors have software emulators for their devices available for free or minimal charge in the developer section of their Web site.
ACCESSIBILITY Accessibility is an integral part of any system design because it is one element of usability. Accessibility can be denned as "the ease with which content, a device, or a system lends itself to be used by persons with a diverse range of physical conditions" (Pearrow, 2002). Mobility is one of the great advantages afforded to most people by portable computing devices, and it is the responsibility of the usability expert to ensure systems are also accessible under a broad variety of conditions. Accessibility is usually categorized as a service for people with disabilities, but this myopic approach does a disservice to many users. The principle of accessibility is extensible to people with any degree of
34. Wireless Communication ability, under any usage conditions. A mobile computing device that is equipped with audible I/O cues might be used by a nonsighted person, but it is just as likely to be used by a sighted person who is operating under conditions in which he or she cannot use the sense of sight to access the device. For example, a police officer might be in the field, performing a routine check on a motorist's vehicle after pulling the driver over for suspicious activity. If the police officer were to divert his or her attention to a handheld device instead of paying attention to a potentially hazardous suspect, he or she would be put at high risk. Assistive technology can be used by people with all levels of ability, and so it is important that accessibility itself come to be understood not as a special case, but as the default case. Although ample literature exists on the topic of Web accessibility, little work has been done to date in the area of accessibility for mobile computing. The World Wide Web Consortium (W3C) publishes a list of techniques and guidelines for Web content accessibility,* but at the time of writing there was no information specifically about portable computing devices or content. Much of the research in this area that does exist is aimed at creating proxy interfaces to "normal" devices (i.e., unmodified or unoptimized designs). For example, Feiner and Shamash (1991) suggested a method for increasing the virtual size of a small screen through the use of a hybrid user interface—in their case, a combination of a high-resolution, small-area display, like a handheld device, and a large-area, low-resolution virtual display, like a eadworn "heads-up" system. Their proposed system consisted of a headworn monocular eyepiece that presented a large virtual space, as well as a head-position tracking sensor. Such a system would probably have sociologic phenomena associated with its use; a headworn system of this sort might conjure up images of the Borg.* An important component of the usability of a mobile system is the social acceptance of the device and its modes of use, so approaches of this sort are less likely to be effective in mobile applications. Abascal and Civit (2001) noted that mobile telephony and computing devices can be used to increase the autonomy of the elderly. They also pointed out that the cliche that "old people don't like to use technology" is likely a misperception that results from the immense enthusiasm that young people have for new technology, coupled with the plain fact that most new technology is poorly designed and unusable (Abascal & Civit, 2001). In fact, most portable computing technology could be used in the form of assistive technology, such as augmentative communicators, emergency location devices, and so on. More information on universal access and designing for users with disabilities can be found in Choong, Plocher, and Rau (chap. 16, this volume) and Zhu, Vu, and Proctor (chap. 18, this volume). In the United States, accessibility is mandated by what is known commonly as "Section 508," or Section 508 of the Rehabilitation Act of 1973. According to www.section508.gov,
645
Section 508 requires that when Federal agencies develop, procure, maintain, or use electronic and information technology, Federal employees with disabilities have access to and use of information and data that is comparable to the access and use by Federal employees who are not individuals with disabilities, unless an undue burden would be imposed on the agency.
Many concrete requirements are spelled out in this document that pertain to services over the Web, as well as requirements that pertain to "telecommunications devices" and "self-contained" or "closed" devices. Some of the requirements have significant implications for portable computing devices. Although only U. S. federal agencies are required to provide Section 508 compliance, these accessibility requirements also benefit organizations that have no such requirement. For example, under Section 508, § 1194.23 (Telecommunications products), we find the following requirements: (k) Products which have mechanically operated controls or keys, shall comply with the following: (1) Controls and keys shall be tactilely discernible without activating the controls or keys. (2) Controls and keys shall be operable with one hand and shall not require tight grasping, pinching, or twisting of the wrist. The force required to activate controls and keys shall be 5 Ibs. (22.2 N) maximum. These requirements are unlikely to be met by the majority of portable computing devices. "Soft keys," or control elements that are represented on a touch screen, cannot really be tactilely discernable at all, much less without activation. Most thumb keyboards put a large amount of strain on the hands, fingers, and wrists, and require a high degree of manual dexterity to operate. Under § 1194.25 (Self contained, closed products), the following guideline clearly has implications for mobile computing products: (c) Where a product utilizes touchscreens or contact-sensitive controls, an input method shall be provided that complies with § 1194.23 (k) (1) through (4). Mobile computing technology has the potential to be a liberating and empowering technology for people of all abilities, but there is a dearth of products that have been designed with accessibility as a high priority. Since the U.S. federal government has mandated accessibility, there is a built-in market for such devices. In addition, accessible devices are suitable for users who are operating them in physically or cognitively taxing environments. It is our hope that as more nations adopt accessibility requirements at the government level, more manufacturers will see the economic benefits of accessible devices, if not the utilitarian or ethical benefits.
*http://www.w3.org/WAI/Resources/. *A cybernetic alien race—with a penchant for monocular eyepieces—depicted in the science fiction series, "Star Trek: Next Generation."
646
PEARROW
References Abascal, J. & Civit, A. (2001). "Bridging the 'Gap between Design for all and Assistive devices'" In Stephanidis, Constantine (ed.), Universal Access in HCI. London: Lawrence Erlbaum Associates. Bailey, R. W (1996). Human performance engineering: Designing high quality professional user interfaces for computer products, applications and systems. Upper Saddle River, NJ: Prentice Hall. Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press. Buchanan, G.Jones, M., Thimbleby, H., Farrant, S., & Pazzani, M. (2001). Improving mobile internet usability. Proceedings of the 10th International. WWW Conference (pp. 673-680). New York: ACM Press. Chang, H., Tail, C., Cohen, N., Shapiro, M., Mastrianni, S., Floyd, R., House, B., & Lindquist, D. (1997). Web browsing in a wireless environment: Disconnected and asynchronous operation in ARTour Web Express. Proceedings of the ThirdAnnual ACM/IEEE International Conference on Mobile Computing and Networking 260-269. Coen, M., Phillips, B., Warshawsky, N.,Weisman, L., Peters, S., & Finin, E (1999). Meeting the computational needs of intelligent environments: The Metaglue system. Proceedings ofMANSE'99. Deleuze, G., & Guattari, F. (1986). Nomadology. New York: Semiotext(e). Feiner, S., & Shamash, A. (1991). Hybrid user interfaces: Breeding virtually bigger interfaces for physically smaller computers. Housel, B., & Lindquist, D. (1996). WebExpress: A system for optimizing Web browsing in a wireless environment, Proceedings of the 2nd annual international conference on Mobile computing and networking, 108-116. Jing, J., Helal, A., & Elmagarmid, A. (1999). "Client-Server Computing in Mobile Environments." ACM Computing Surveys, 31(2), 117157. Just, M. A., Carpenter, P. A., Keller, T. A., Emery, L., Zajac, H., &Thulborn, K. R. (2001). Interdependence of nonoverlapping cortical systems in dual cognitive tasks. Neurolmage, 14, 417-426. Kamba, T., Elson, S. A., Harpold, T., Stamper, T., & Sukaviriya, P. N. (1996). Using small screen space more efficiently. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, (383-390). La Porta, T. E, Sabnani, K. K., & Gitlin, R. D. (1996). Challenges for nomadic computing: Mobility management and wireless communications. Mo bile Networks and Application, 1(1), 3-16. LeRouge, C. (2001). Mobile communications. Alcatel Telecommunications Review, Third Quarter. Norman, D. (1999). The invisible computer: Why good products can fail, the personal computer is so complex, and information appliances are the solution. Cambridge, MA: MIT Press. Palen, L., Salzman, M., & Youngs, E. (2000). "Going wireless: Behavior & practice of new mobile phone uses", Proceedings of the 2000 ACM conference on computer supported cooperative work, 201-210. Palen, L., & Salzman, M. (2002). Beyond the handset: Designing for wireless communications usability. ACM Transactions on ComputerHuman Interaction, 9(2), 125-151. Pearrow, M. (2000). The Web site usability handbook. Rockford, MA: Charles River Media. Pearrow, M. (2002). Wireless web usability. Hingham, MA: Charles River Media. Priyantha, N. B., Chakraborty, A., & Balakrishnan, H. (2000). The Cricket location-support system. Proceedings of the ACM MOBICOM Conference.
Trevor, J., Hilbert, D. M., Schilit, B. N., & Koh, T. K. (2001). From desktop to phonetop: A UI for Web interaction on very small devices. Proceedings of the 14th Annual ACM Symposium on User Interface Software and Technology (UIST2001), 121-130.
FURTHER READINGS Abowd, G. D., Dey, A. K., Orr, R., & Brotherton, J. (1997). Contextawareness in wearable and ubiquitous computing. ISWC, 179-180. Albers, M. J., & Kim, L. (2000). User web browsing characteristics using palm handhelds for information retrieval. Proceedings of IEEE Professional Communication Society International Professional Communication Conference and Proceedings of the 18th Annual ACM International Conference on Computer Documentation: Technology & Teamwork, 125-135. Alesso, H. P., & Smith, C. E (2001). The intelligent wireless Web. Boston: Addison-Wesley. Andersson, C. (2001). GPRS and3G wireless applications: professional developer's guide. Los Angeles Wiley. Buyukkokten, O., Kaljuvee, O., Garcia-Monlina, H., Paepcke, A., & Winograd, T. (2002). Efficient web browsing on handheld devices using page and form summarization. TOIS, 20(1), 82-115. Dillon, A., & Morris, M. (1998). Power, perception and performance: From usability engineering to technology acceptance with the P3 model of user response. Proceedings of the 43rd Annual Conference of the Human Factors and Ergonomics Society. Santa Monica, CA: HFES. Edwards, W K., & Grinter, R. E. (2001). At home with ubiquitous computing: Seven challenges. Ubicomp 2001 (pp. 256-272). Fleetwood, M. D., Byrne, M. D., Centgraf, P., Dudziak, K., Lin, B., & Mogilev, D. (in press). An analysis of text-entry in Palm OS—Graffiti and the Virtual Keyboard. Proceedings of the Human Factors and Ergonomics Society 46th Annual Meeting. Cast, M. S. (2002). 802.11 wireless networks: The definitive guide. Sebastopol, CA: O'Reilly and Associates. Health Council of the Netherlands. (2002). Mobile telephones: An evaluation of health effects (Publication no. 2002/01E). The Hague: Health Council of the Netherlands. Jacob, B., & Mudge, T. (1996). Support for nomadism in a global environment. Proceedings of the Workshop on Object Replication and Mobile Computing. New York: ACM Press. Law, J., & A. Mol (2000). Situating technoscience: An inquiry into spatialities [On-Line]. Available: http://www.comp.lancs.ac.uk/ sociology/soc052jl.html Myers, B. A., Wobbrock, J. O., Yang, S., Yeung, B. Nichols, J., & Miller, R. (2002). Using handhelds to help people with motor impairments. Proceedings of the ACM SIGCAPH's 5th International Conference on Assistive Technologies (ASSETS '02), 89-96. Sacher, H., & Loudon, G. (2002). Uncovering the new wireless interaction paradigm. Interactions (pp. 17-23). New York: ACM Press. Available: http://doi.acm.org/10.1145/503355.503364 Schmidt, A., Schroder H., & Frick O. (2000). WAP—Designing for small user interfaces. Proceedings of the CHI2000 Conference on Human Factors in Computing Systems, Abstracts Volume (pp. 187-188). New York: ACM Press. Ziefle, M. (1998). Effects of display resolution on visual performance. Human Factors, 40(4), 555-568.
•35^ AUGMENTED REALITY IN INTERNET APPLICATIONS Kay M. Stanney University of Central Florida
Roy C. Davies Lund University
INTRODUCTION Augmented reality (AR) offers a unique medium for Web-based interaction, immersing users via an egocentric perspective and allowing traversal throughout three-dimensional (3D) digital worlds, inhabited by tangible objects that can be manipulated and by entities, such as adversaries, instructors, or partners, with which one can communicate. One of the earliest AR applications was Digitaldesk (Wellner, 1993), which provided an augmented desk that supported computer-based interaction with paper that could take on electronic properties. Yet, our desk has become more than the physical artifact in front of us; it has expanded into an interactive portal to a digital world. How can AR be used to enhance access and use of our digital desktop, particularly our gateway into Web-based applications? What does AR mean in the context of the Internet? What potential will such systems provide in how we search, retrieve, visualize, and interact with information? This chapter defines what is meant by AR in terms of Internet design, addresses design and evaluation issues, and provides a number of key questions to consider in the development of AR-based Web design solutions.
Augmented Reality Design Traditional AR has been denned as "the use of transparent glasses on which a computer displays data so the viewer can view 647
the data superimposed on real world scenes" (Blade & Padgett, 2002, p. 17). In other words, the user sees the real world augmented by additional information provided by the computer. Thus, AR can bridge the gap between virtual environments (VEs; i.e., computer-generated worlds) and the real world. However, what does this mean in the context of the Internet? After all, most people do not have transparent glasses for their home PC. Yet, while the current technology does not provide these capabilities, the rate of technology advances means that such systems will very likely be in the home quite soon. It is thus prudent to consider how and when AR should be integrated with Web-based applications. In the context of Web-based applications, AR could allow users to combine Web objects and virtual objects, resulting in a more comprehensive view of presented information. Furthermore, by bringing these objects out of the computer screen and into the real world, the user is able to interact with them in a more natural and intuitive fashion. With the aid of transparent glasses, a video camera, and software to recognize markers in the real world, virtual object references on a Web site could be transposed on the real world in place of the markers (Billinghurst & Kato, 1999; Sinclair & Martinez, 2002). This would allow direct, physical manipulation of 3D objects on the Web site. For example, Milosavljevic, Dale, Green, Paris, and Williams (1998) developed a system that generates Web pages for each object in a museum's collection as a visitor navigates around the system's Web site. These pages are tailored to a user's interests, preferences, goals, and previous interactions with the system. Another
648
STANNEY AND DAVIES
example is that by Fraunhofer IGD, who developed a mobile AR Web-based system that supports the product development life cycle (Mullet, Strieker, & Weidenhausen, 2001). In general, AR applications could provide information based on a combination of users' preferences and a given Web-based context. Augmented information could assist in locating items on the Web and suggest new Web locations to visit, thereby reducing the chance of a user becoming lost. Users could access augmented information as they traversed through a Web site, which could be configured with Web-based location "sensors" that once triggered would display information for a given location. Users could also request augmented information on related objects, comparisons between entities of interest, suggested Web links to follow, and presentation of instructions. The potential for enhancing Web based applications via AR technology are endless.
Augmented Reality Web-Based Applications In this section, four AR Web-based application types that have addressed the challenges involved in combining virtual objects with information from the Internet are considered. The lessons learned from these application types can be used as inspiration for others who are developing AR applications in the context of Internet design. Web site with augmenting virtual objects. Augmenting Web sites with virtual objects is a relatively standard technique today; it can be considered another variation of multimedia where a form of information is provided by 3D objects. This is often used to show products, for example, a car or computer, that the user can rotate, move, and perhaps have nominal interaction abilities with, such as by opening a door or pressing a button to select product preferences (e.g., color, style). Such applications require that there is a standard means of representing virtual objects so they can be properly rendered on the user's computer. To ensure the usability of such applications, there should be a tight coupling between user input and display variant responses and such responses should be rapidly reversible. The latter is important for encouraging user interaction, as well as for assisting users who may have become disoriented while navigating around virtual objects. There are several 3D Web creators that can be used to develop such Web sites, including virtual reality modeling language (VRML), Macromedia Director (http://www.macromedia.com/), and Cult3D (http ://www. cycore. com/). Virtual environment with augmenting web links. Another standard practice is to develop collaborative VEs (CVEs) in which several users can interact together, move around in the VE, and perhaps even bring in virtual objects of their own construction and link to various Web pages and other CVEs. Examples of this are Activeworlds (http://www.activeworlds.com/) and Adobe Atmosphere (http://www.adobe.com/products/atmosphere/). For such collaborative environments, it is essential to develop collective action regulation mechanisms, as well as a communication
etiquette (to support turn-taking, etc.), both of which structure and control interaction (Arial & Schar, 2002; Lockner & Winroth, 1999). In addition, providing an overview option and peripheral views so users can take in more of the entire scene and determine who is currently acting can assist with situation awareness and orientation issues. Further, if speech is a viable interaction option, then communication audio quality is essential to realize effective collaboration. Web site with virtual objects presented using augmented reality. Another option is to employ tangible AR, which tries to eliminate the gap between interaction with a natural environment and a Web site by using real-world objects in the natural environment as an interface to the Web site (Millard, Moreau, Davis, & Reich, 2000). Thus, in essence the real and virtual worlds are combined. In this instance, the computer screen contains a Web page, which has references to virtual objects that are rendered to the user through a see-through display. The user wears a camera that the computer uses to identify markers in the real environment, which are replaced by virtual objects as defined by the Web page. By manipulating the real objects that markers are placed on, the user can also manipulate virtual objects. An interesting variation is where the computer screen is rendered on the side of a flat virtual object, allowing it to also be attached to a marker and manipulated physically. Numerous virtual screens can then be placed around the real environment. Tangible AR provides an intuitive way of interacting with virtual objects, which is especially beneficial for novice users. However, to ensure usability, it is essential to have a clear binding between physical and virtual objects (Fjeld, Lauche, et al., 2002; Fjeld, Schar, Signorello, & Krueger, 2002). More specifically, it is essential to consider physical aspects (how to use the physical object to affect computer operations), handling aspects (conditions for transparency of the physical object such that a user can focus on the virtual object), and subject-object-directed aspects (how to use the virtual object to complete tasks) when designing physical-virtual object couplings. In addition, before users interact with the virtual objects, they should interact with their real-world counterparts (i.e., physical tools) to gain familiarity. Location-Aware augmented reality. With location-aware AR, users have wearable computers that project hypermedia information onto particular locations within the real world (Fritsch, Klinec, & Volz, 2001; Haala, 2001). Such applications can take into account contextual factors, such as location, time, identity, and activity, and use this information to adapt user interaction (Burrell, Gay, Kubo, & Farina, 2002). For example, this concept could be developed in the form of a large room with furnishings overlaid by virtual objects. The virtual objects could then become links to Web sites (or indeed virtual screens containing Web sites). When a user enters the room and accesses a Web site, depending on current contextual factors, this interaction could change the content of the virtual world accordingly. Further interaction could then be dependent on such variables as frequency of interaction or user location-based annotations. Such applications would require some form of tracking system and
35. Augmented Reality gesture registration hardware to support interaction with the actual objects. Spohrer (1998) stretched this concept, suggesting the creation of a WorldBoard, which acts as a planetary AR system that facilitates location-aware association of information with target places. Such a system would allow users to post messages on the faces of the cubic meters of space they traverse during their planetary stay. Once posted, users could see personal Web pages when they look at their (or others') office doors, label interesting artifacts on a nature trail, etc., thereby experiencing any information in any place, co-registered with reality. For specific applications, for example, in architecture or vehicle design, location-aware AR may provide a powerful means of combining real object interaction with virtual object adaptability. Smaller real environments, such as a car interior, could also form the real (tangible) component, with the virtual portraying perhaps different types of interior. With the real and virtual worlds so tightly coupled, it is essential to consider issues such as distractions and coordinated awareness of physical and virtual spaces.
Augmented Reality Design Considerations Although the lessons learned from the design of such AR Web applications are valuable, for those who endeavor to develop AR-based Web sites, there are a number of additional issues that must be considered to ensure effective augmented interaction. In particular, six characteristics of AR design tools have been derived in the context of a classification scheme based on the form and usage of each tool, which can be used to guide the design of AR Web-based applications (Davies, 2002). Task, Users, and Tools. Any Web site (or indeed computer application), including those with AR, should have a purpose (i.e., a task that the user is trying to perform). It is important that the purpose be well defined, as it impacts heavily on the overall usability of an application, particularly when a Web site is not designed so intended users can perform intended tasks with the tools provided. In using AR, and in particular during interaction with virtual objects, consideration of users' levels of expertise with the manipulation hardware is important because the handling of some devices can be difficult for novices. As mentioned previously, designers should allow users to interact with real-world counterparts (i.e., physical tools) for virtual objects when they are available, while also ensuring a clear binding between physical and virtual objects (Fjeld, Lauche, et al., 2002; Fjeld, Schar, et al., 2002). Beyond experience, the spatial ability of users may be particularly important to AR Web-based applications because it can be difficult to maintain orientation within two different frames of reference (i.e., virtual and real worlds). System requirements. Because not all people will necessarily have the hardware required to fully use AR Web-based applications, careful definition of system requirements must be made and perhaps even alternative ways of interaction programmed. This is much the same way as designing Web sites for browsers that can and cannot handle frames. If users have different
649
system configurations, this may lead to perceived differences in architectural space (i.e., layout of augmented environment), semantic space (i.e., intuitiveness of representations and interaction techniques), social space [i.e., awareness of others (either avatars or other users) and their activities], and temporal space (i.e., flow of interaction), which can lead to differences in the effectiveness of the system for individual users (Tromp, 1999). Collaborative potential Currently, users do not have to be alone in a virtual (or augmented) environment. Nevertheless, whether collaboration should be supported depends on the task and purpose of the application. In some instances, it makes sense to allow communication between online users (e.g., for co-work, co-learning, co-discovery, entertainment, etc.), in others, it does not. In general, if an application involves users interacting in a shared space while coordinating multiple activities, then collaboration may be beneficial (Tromp, 1997). For such applications, it is essential to ensure synchronization of events across media and to minimize perceived object/action lag (Steed & Tromp, 1998). As mentioned previously, designers must ensure a tight coupling and rapid reversibility between user input and display variant responses, which becomes complex in collaborative environments because it involves consideration of multiple users, executing and switching between multiple, simultaneous tasks. To support collaboration in AR Web-based applications, users might want to leave comments, messages, or even virtual objects after themselves to allow communication with other users that will later enter the environment (Spohrer, 1998). In some CVEs, for example, it is possible to build your own structures that persist even after you log off. Similarly, members of a design team may want to leave notes to others as to what they thought of the latest version of a design. The most appropriate form of such annotated collaboration likely depends on the context of use, and thus should be specified for any given application. Interaction metaphor. In the application types discussed previously, there are primary and secondary forms of interaction metaphor. The primary metaphor is that with which one controls the overall experience; the secondary metaphor supports the interaction within the primary metaphoric framework. For example, take tangible AR Web-based applications; the primary interaction metaphor is through traditional means to control a Web site on a computer screen, while the secondary metaphor involves interaction with physical objects to which virtual objects are bound. Of course, the boundaries between the two metaphors can become a little fuzzy if, for example, the contents of the screen are shown on a virtual object and perhaps the physical keyboard is replaced by a less tangible form of interaction. Nevertheless, it is vital to ensure the form of interaction chosen is appropriate to the task and target users. In general, the optimal interaction metaphor will depend on the input conditions (i.e., continuous input vs. start-stop vs. automatic) and the interaction task (e.g., travel, selection, manipulation, system control; Bowman, 2002). Content. The content of an AR Web-based application should, of course, be tied to the task (Davies, 2002). Aspects that need to
650
STANNEY AND DAVIES
be considered are (a) what is the level of concreteness (or conversely abstractness) of the information being represented, (b) what sort of behaviors are required of the objects, and (c) what level of realism is required? In a museum, for example, a virtual representation of a real artwork can be said to be quite concrete, whereas a visualization of some numerical data is abstract. Design Methodology. There are various ways to design any application that can impact how "well it meets the needs of users. In some circumstances, it is important to consider the users' own opinions, whereas other situations do not allow enough time for this or the exact users are not known. Participatory design is particularly advantageous when the tacit knowledge invested in the people who work (or play) day to day in a particular situation must be captured (Davies, 2002). In situations where designers regard end users as a collection of typical characteristics, perhaps derived from historical data, subject-matter expert interview, or general reference, expertdriven design works well. In this case, typical end users may be called on for evaluation, but the bulk of the design work progresses with the guidance of a team of experts. AUGMENTED REALITY USABILITY The previous section addressed a number of general AR design considerations. To ensure the usability of such applications, a number of additional issues must be attended to, as designing usable and effective AR-based Web design solutions is a new challenge for system developers and human factors specialists. Usability assessment traditionally involves determining if users can perform tasks supported by a system: (a) efficiently (i.e., the quantity of performance effectiveness); with (b) minimal errors (i.e., the quality of performance effectiveness); (c) minimal support (i.e., intuitiveness in terms of learnability); (d) maximal retention of commands over time (i.e., intuitiveness in terms of memorability); (e) appropriate task demands (i.e., positive subjective perception in terms of perceived mental workload); and (f) sufficient satisfaction (Eberts, 1994; Shneiderman, 1992). Yet, the traditional usability principles listed previously do not consider characteristics unique to AR systems. Beyond these traditional criteria, there are a number of additional system interface and user interface considerations that must be addressed to ensure AR-based Web applications are well received by their users. System interface considerations include interaction and multimodal presentation concerns. User interface considerations include engagement and side effects concerns.
System Interface Considerations Most AR user interfaces are fundamentally different from traditional graphical user interfaces (GUIs), with unique input/output (I/O) devices, perspectives, and physiological interactions. In an AR Web-based application, the computer, displays, and controls in the interface are configured to involve users in an environment containing 3D objects with locations and orientations in multidimensional space, which users can
FIGURE 35.1. System interface usability considerations. interact with in real time (Durlach & Mavor, 1995). Thus, although traditional usability evaluation techniques, do apply to AR Web based applications, they may not be comprehensive enough to characterize usability attributes specific to 3D spatially immersive and interactive environments. Usability assessment of AR Web-based system interfaces (Fig. 35.1) must also consider the interaction techniques and mutlimodal system outputs employed in a given Web and AR application. Interaction, Although interaction in traditional GUI applications is largely restricted to point-and-click-type selection, AR Web-based applications can employ additional interaction techniques such as travel to a destination (i.e., movement of user's viewpoint from place to place) and manipulation of selected objects (i.e., setting the position and/or orientation of virtual objects) both inside and outside the screen. Travel involves both wayfinding (i.e., locating and orienting oneself in an environment) and navigation (i.e., moving from one location to another in an environment; Darken, 1996). Wayfinding. Users must plan how to get from one point of interest to another in an AR Web-based application. The Web provides many sources of navigation assistance, including the http "address" of the current site, a "home" location from which to reorient, history lists, "Back" and "Forward" buttons, the use of color-coding to indicate previously visited links, site maps and table of contents, search mechanisms, and bookmarks (Stanney, Chen, Wedell, & Breaux, 2003). Even given these aids, wayfinding in AR applications can be challenging because it involves cognitive activities (e.g., route planning), rather than simply involving manual maneuvering. If there is insufficient or inappropriate information provided about the spatial structure of such applications, then users are likely to have difficulties locating their current and/or desired destinations (Kaur, 1999).
35. Augmented Reality Thus, such applications must include appropriate sensory cues and navigational aids (e.g., compass, map) to facilitate users' acquisition of spatial knowledge. Without proper design of the navigational space and availability of tools to aid in exploring this space, the overall usability of AR Web-based applications may suffer resulting in ineffective and inefficient human performance or unpleasant user experiences. Specific wayfinding usability concerns include (see Stanney, Mollaghasemi, Reeves, Breaux, & Graeber, 2003): • Do users often feel disoriented or do they often not know where they are (Kalawsky, 1999)? • Does the application facilitate users' acquisition of survey knowledge through maps or other such aids, thus assisting users in determining where they want to go and how to get there (Darken & Sibert, 1996a, 1996b; Lynch, I960; Wickens & Baker, 1995)? • Are appropriate spatial labels, landmarks, and "you-are-here" markers incorporated into the application (Bennett, Chapelle, Zeltzer, Bryson, & Bolas, 1996; Darken & Sibert, 1996a, 1996b)? • Should sensory information other than visual be provided to guide wayfinding, such as the use of aural cues to provide directional cues (Barfield & Danis, 1996; Carter, 1992)? Navigation. Once wayfinding plans have been made, a user traverses or navigates to a point of interest within a given Web site. Such movement should be intuitive to enact, rather than cognitively cumbersome or physically awkward to achieve (Stanney, Mollaghasemi, et al., 2003). Such movement allows users to move into position to perform desired tasks or activities within a Web site. In general, navigation can be aided by tools that display an individual's current position (e.g., current address, highlighting the frame of an active window), display an individual's current orientation (e.g., magnetic compass, a "home" location from which to reorient), log an individual's movements (e.g., history list), demonstrate the surrounding environment (e.g., binoculars, maps, radar, perspective), and guide navigation (e.g., signs, verbal route descriptions, tours, links, and filtering, autopilot systems; Stanney, Chen, et al., 2003). Specific navigation usability concerns include (see Stanney, Mollaghasemi, et al., 2003): • Are the degrees of freedom (DOFs) of user movement streamlined such that it is easy for users to move and reposition themselves within the AR Web-based application (Hinckley, Pausch, Goble, & Kassell, 1994; Jacob, Sibert, McFarlane, & Mullen, 1994; Kalawsky, 1999)? • Does the system effectively use multiple (integral) DOF for coarse user movement, high gain for gross user movement, low gain for fine user movement, and separable DOF for precision user movement (Hinckley etal., 1994; MacKenzie, 1995)? • Is a navigational grid or map included for large environments, and does it adhere to map design principles (Darken & Sibert, 1996a)? • Are a variety of tools available to support navigation, such as indicators of current position and current orientation; logs
651
of previous movements; demonstrations of the surrounding environment beyond maps such as those provided by binoculars, radar, or perspective; and indicators to guide navigation such as signs or verbal route descriptions (Stanney, Chen, et al., 2003)? Objection Selection and Manipulation. One of the main advantages of coupling AR with Web-based applications is that it provides the ability to incorporate virtual objects that can be manipulated. Object selection and manipulation may be denned as the process of indicating virtual objects within an AR Web-based application that one desires to reposition, reorient, or query (Bowman, 1999; Gabbard & Hix, 1997). Which object selection and manipulation methods are incorporated into an AR application, from traditional point and click to more contemporary techniques such as ray casting, speech, and gesture and on to "magical" operations that would never be possible in the real world, has a profound impact on usability, particularly affecting the ease and efficiency of user interaction (Stanney, Mollaghasemi, et al., 2003). The purposes and tasks for which a Web application is designed determine which technique(s) are most appropriate. Specific object selection and manipulation usability concerns include (see Stanney, Mollaghasemi, et al., 2003): • Are object selection and manipulation devices easy to use and control (not too sensitive or too sluggish), thereby making it easy to select, move, and alter basic attributes of objects (color, shape, labels), as well as examine objects from multiple perspectives (Kalawsky, 1999)? • Are object selection points obvious and clear, thereby making it easy to select single or multiple objects? Consider using exaggerations in size, appearance, and interobject distance, transparency, and query formation to assist with object selection (Esposito, 1996; Hinckley et al., 1994; Mine, Brooks, & Sequin, 1997; Zhai & Milgram, 1994). • Are exocentric points of view provided for relative positioning and relative user-object motion control? • Are there any extraneous DOFs? Those not necessary to support tasks should be eliminated because they make object manipulation difficult (Hinckley et al., 1994). • Is ray casting used when objects to be selected are very small or co-located, and cone casting with associated spotlighting used when selecting regions or large, sparse objects (Hinckley et al., 1994; Liang & Green, 1994)? • Is selection based on spatial attributes (location, shape, orientation) supported via direct manipulation (Shneiderman, 1992)? • Is selection based on temporal, descriptive, or relational attributes supported via nondirect manipulation (e.g., queries)? Muttimodal interaction. Using AR, many more sensory experiences can be brought to user interaction beyond traditionally visually dominant Web-based applications. The visual modality can be used to convey spatial relations (e.g., size, location, distance) via graphics and animation; abstract, trend,
652
STANNEY AND DAVIES
or quantitative relationships; and to convey verbal-text information (Baecker, Small, & Mander, 1991). The auditory modality is highly effective at alerting or warning (Sanders & McCormick, 1993), conveying instructions and other relevant information via speech (Hapeshi & Jones, 1992), and providing a source of spatial localization (Blauert, 1996). The haptic modality is advantageous for conveying temporal information, alarming, or warning, and aiding hand-eye coordination tasks (Biggs & Srinivasan, 2002; Posner, 1976). Haptics can be used to stimulate anticipation of a change, provide feedback confirming reception of a user input, provide an indication of current state, guide user interaction toward a desired position or location, and make clear distinctions between orthogonal directions (Miller & Zeleznik, 1999). These sensory modalities can be effectively used together. For example, tangible user interfaces, such as the Massachusetts Institute of Technology's (MIT's) Tangible Bits, can provide physical forms to Web-based digital information, such as by integrating graspable objects and augmented surfaces into Web-based applications or by leveraging "ambient media," such as ambient light, sound, airflow, and water movement to realize digitally mediated awareness of ambient activity (Ishii & Ullmer, 1997). Magic Book leverages visual animations and verbal-text communication, while allowing physical (i.e., haptic) interaction with a storybook (Billinghurst, Kato, & Poupyrev, 2001). While readers view pages wearing a head-mounted display (HMD), animated pictures come to life and readers can interact with 3D virtual scenes. Such multimodal interaction can be leveraged to greatly enhance Webbased applications. When displaying multimodal information, it is essential to consider a number of usability concerns specific to each modality; however, an underlying principle is that any form of information (i.e., visual, auditory, haptic) presented to users should be readily understood, unambiguous, and necessary to complete required tasks (Stanney, Mollaghasemi, et al., 2003).
Visual Output Traditionally, there have been two general types of display technologies used for presenting users with visual information (Vince, 2004). The first type, fully immersive displays [e.g., HMD, arm-mounted (BOOM) display, and virtual retinal display], completely isolates users from the real world. The second type, semiimmersive or "fish tank" displays (e.g., surround screen virtual reality, immersadesk, stereo monitor, see-through HMDs), provides users with the ability to see both the real and virtual worlds simultaneously, and is likely the display mode most suitable for AR Web-based applications. For example, in see-through HMDs, cameras are often mounted onto an HMD, which feed real-world imagery to an image processing system that adds virtual imagery to display in the HMD (Takagi, Yamazaki, Saito, & Taniguchi, 2000). If one seeks to achieve an "immersive" Web application, then a field of view (FOV) of 100 degrees or more is desired (Kalawsky, 1993); however, it is more important to determine what FOV is best suited to a particular application's needs. Other visual modality considerations that affect usability and human performance include when to incorporate stereo displays; what display update or refresh rate is acceptable; and when might simplified, cartoon-
like images be appropriate versus when are high-fidelity displays required (Gabbard & Hix, 1997). Specific visual display usability concerns include (see Stanney, Mollaghasemi, et al., 2003): • Is visual display seamlessly integrated into users' task activities and is the FOV appropriate for supporting such activities (Gabbard & Hix, 1997)? • Does visual output have high frame rates and low latency, with no perceivable distortions in visual images (Kalawsky, 1999; Richard et al., 1996; Ware & Balakrishnan, 1994)? • Is stereopsis effectively used to convey realistic scenes of information provided via an egocentric view? • Is the level of visual scene detail (i.e., photorealistic vs. cartoonlike images) appropriate for the Web application? Auditory Output. Few designers use systematic strategies to implement sound due to a lack of relevant guiding principles, thus usability issues concerning audio interaction can arise once Web applications are deployed (Gabbard & Hix, 1997; Kaur, 1999; Krapichler et al., 1999). One notable exception is Cohen (1992), who provided a conceptual model for organizing and controlling sound within traditional interactive systems. This technique could be modified and extended to realize auditory design strategies for AR Web-based applications. In general, acoustics cues can be used to complement navigational activities (e.g., walking sounds accompanying virtual walkthroughs), thereby enhancing spatial awareness. This could assist users in judging distance and progress while finding their way throughout a Web site. In addition, a combination of visual and acoustic information can be incorporated into Web sites to enhance the intuitiveness of user actions (e.g., pressing a Web link, returning to 'home'). Specific auditory display usability concerns include (see Stanney, Mollaghasemi, et al., 2003): • Is auditory display seamlessly integrated into user task activity, and does it avoid cumbersome display lags (Gabbard & Hix, 1997)? • Are sounds generated in realtime to accentuate user actions, observations, and experiences (Cohen & Wenzel, 1995)? • Does high bandwidth aural information support simultaneous, dynamic presentation of different sounds from different locations, and are these sounds easy to identify and localize (Cohen & Wenzel, 1995; Witmer & Singer, 1998)? • Are headsets used when portable, cost-effective audio is needed for remote single users (Gabbard & Hix, 1997)? Haptic Output. Two types of haptic feedback can be provided in AR Web-based application: (a) kinesthetic— information sensed through movement and/or tension signals originating from sensory receptors in the joints, skin, muscles, and tendons; (b) tactile—information received through nerve receptors in the skin (e.g., with a finger pad), which convey shapes and textures (Biggs & Srinivasan, 2002). When incorporating such haptic interaction, there are a number of usability considerations that must be addressed to ensure the safety and comfort of users. Specific haptic display usability concerns
35. Augmented Reality
653
(i.e., immersion). More specifically, presence may be thought of as the psychological perception of "being in" the Web-based application in which one is immersed (Witmer & Singer, 1998). Immersion can be assessed directly, according to the effectiveness of delivery hardware and software. Presence, being a more cognitively derived construct, is not as directly assessable (although researchers have tried to link it to physiologic measures; see Sadowski & Stanney, 2002, for a review of measures) and thus it is usually characterized via subject report.
FIGURE 35.2. User interface usability considerations, include (see Stanney, Mollaghasemi, et al., 2003): • Is haptic display seamlessly integrated into user task activity, and does it employ incredibly high (spatial) resolution, high frame rates, and low latency, while avoiding cumbersome display lag (Richard et al., 1996; Sturman, Zeltzer, & Pieper, 1989; Ware & Balakrishnan, 1994; Wickens & Baker, 1995)? • Does haptic display provide reliable, intuitive, low fatigue interaction, and does it effectively couple both kinesthetic and tactile information (Kaczmarek & Bach-YRita, 1995; McNeely et al., 1995)? • Does haptic display avoid presenting and semantically binding a large number of haptic intensity levels (Kaczmarek & BachY-Rita, 1995)? • Are users able to actively search and survey the environment via touch and easily identify objects through physical interaction (Witmer & Singer, 1998)? • Can users readily perceive simultaneous presentation of complex haptic patterns, sensations, and objects (Kaczmarek & Bach-Y-Rita, 1995)?
User Interface Considerations User interface considerations are quite different from system interface considerations in that they focus on the user experience engendered by an AR Web-based application (Fig. 35.2). This area of evaluation seeks to determine if the application is engaging and immersive, while avoiding adverse side effects sometimes associated with immersive or wide FOV applications or those that incorporate unwieldy peripheral devices. Engagement. Engagement constitutes both the subjective perception of the experience (i.e., presence), as well as the technology requirements necessary to achieve user captivation
Presence. Particularly when it comes to entertainment or educational AR Web-based applications, the degree of presence (or sense of being engaged in the application) may become paramount in capturing and maintaining users' involvement in the Web-based content. In general, presence is believed to influence both usability and users' task performance (Fontaine, 1992; Zeltzer, 1992). Thus, it is important to assess factors believed to contribute to presence, which include ease of interaction, userinitiated control, pictorial realism, length of exposure, social factors, and system factors (Sadowski & Stanney, 2002; Stanney et al., 1998). Specific presence related usability concerns include (see Stanney, Mollaghasemi, et al., 2003): • Does the AR Web-based application engage a number of sensory modalities, including visual, audio, and haptic aspects (Witmer & Singer, 1998)? • Do users have an egocentric perspective that conveys a compelling sense of self-motion when traversing the Web site, and can they predict responses to their actions? • Is "setting" (or theme) effectively used to enhance presence, and is content similar to real-world experiences (Barfield, Zeltzer, Sheridan, & Slater, 1995; Gabbard & Hix, 1997)? • Do users have control over events during the experience, thereby encouraging them to become emotionally involved in the experience? Immersion. Many technological techniques can be used to achieve a strong sense of immersion in an AR Web-based application. These techniques try to draw the user into the content at the exclusion of the outside "real" environment. This ability to psychologically isolate users is directly affected by the design, particularly the usability of Web applications. Isolation from the physical environment, inclusion in a situation, natural and extensive (i.e., range of sensory modalities) modes of interaction and control, availability and perception of an encompassing (i.e., extent to which scene is panoramic) and vivid stimulus stream, and supporting perception of self-movement are all suggested to increase immersion (Slater & Wilbur, 1997; Witmer & Singer, 1998). Specific immersion-related usability concerns include (see Stanney, Mollaghasemi, et al., 2003): • Are users isolated from and not distracted by outside activities via effective use of immersive technologies (Witmer & Singer, 1998)? • Can users readily concentrate on Web experience and not the use of control devices (Gabbard & Hix, 1997; Kalawsky 1999)?
654
STANNEY AND DAVIES
• Is it difficult for an outside observer to get users' attention when they are engaged in the Web application (Kalawsky, 1999; Witmer & Singer, 1998)? Side Effects. Exposure to AR can lead to adverse effects, such as discomfort, sickness, and aftereffects. It is essential that these issues be addressed to ensure the usability of AR Web-based applications. Comfort. Krapichler et al. (1999) declared the major criteria for acceptance of medical AR applications are system intuitiveness and comfort, thus ensuring user comfort is of paramount importance. Physical discomfort can ensue from issues associated with basic ergonomic concerns, such as the fit of peripheral devices and arrangement of the application environment. Visual discomfort can result from viewing improper depth cues, poor contrast and poor illumination, and/or incorrect interpupilary distance settings (Rushton, Mon-Williams, & Warm, 1994; Warm & Mon-Williams, 2002). In general, one must assess how well AR equipment fits the body to deliver appropriate sense information to users. Specific usability concerns related to comfort include (see Stanney, Mollaghasemi, et al., 2003): • Is AR Web-based application comfortable for long-term use, and does it effectively support users of different statures (e.g., size, weight; Gabbard & Hix, 1997; Kalawsky, 1999)? • Are peripheral devices lightweight, portable, nonencumbering, and comfortable, thereby avoiding issues of limited user mobility and fatigue (Gabbard & Hix, 1997)? • Is excessive eye fatigue experienced (Kalawsky, 1999)? • Do control devices allow for comfortable manipulation of objects via the fingers, or do control devices lead to fatigue by employing large muscle groups (Hannaford & Venema, 1995; Zhai, Milgram, & Buxton, 1996)? • Are real-world props used effectively to reduce body fatigue while sitting or standing and hand/arm fatigue while users interact with the AR web-based application (Stoakley, Conway, & Pausch, 1995)? Sickness. Motion sickness-like symptoms can occur with any immersive application, particularly those that engender vection, an illusory sense of self-motion. In general, sickness associated with interactive applications is believed to be due to sensory conflicts in the neural mechanisms responsible for interpreting and responding to orientation and self-motion, either those immediately present to a user (e.g., visual motion without concordant vestibular stimulation) or between current patterns of input and those anticipated based on experience (e.g., when a visual scene updates later than expected due to lag; Money, 1990). An alternative position is that held by the ecological theory of motion sickness (Riccio & Stoffregen, 1991), which suggests that motion sickness is caused by postural instability associated with environmental situations (i.e., low-frequency vibration, altered specificity) that destabilize the postural control system. AR applications would be anticipated to destabilize postural control through altered specificity (i.e., visually specified accelerations and rotations that lack correlated bodily forces).
Lessons learned from empirical studies investigating the previously stated theories indicate that designers of AR applications should strive to reduce intersensory conflicts in their designs and users should likely be seated or provided with a support bar to assist with maintaining postural control during system exposure. Specific usability concerns related to sickness include (Kennedy, Lane, Berbaum, & Lilienthal, 1993; also see Stanney, Mollaghasemi, et al., 2003): • Are there sensory discordances that lead to sickness, which could be readily redesigned? • Do users experience high levels of general discomfort during interaction with the AR Web-based application? • Do users experience headaches, eye strain, increased salivation, burping, or sweating during interaction? • Do users experience high levels of bodily fatigue or postural instability during interaction? • Do users experience nausea or vomiting during interaction? Aftereffects. Once a user is done interacting with an AR Web-based application, ill effects may linger. Such untoward ailments are known as aftereffects and are a result of the need to readapt to one's normative environment after prolonged exposure to an altered sensory stimulus. Disturbing effects that may linger from minutes to hours or even days postexposure include head spinning, postural ataxia, and reduced eye hand coordination (Stanney et al., 1998). It is of course essential to assess, monitor, and carefully manage such aftereffects upon post exposure. Specific usability concerns related to aftereffects include (see Stanney, Mollaghasemi, et al., 2003): • Do users experience any of the following after exposure to the AR Web-based application: blurred vision, dizziness, nausea, difficulty focusing, or difficulty with coordinated tasks (e.g., walking; Kennedy et al., 1993)? • Do visual, orientation, or coordination aftereffects persist for long periods after exposure? • Do users experience vertigo (loss of orientation to vertical upright) after exposure (Kennedy et al., 1993)? • Are users not likely to desire to reuse the AR Web-based application?
CONCLUSION AR Web-based design solutions have tremendous potential to revolutionize the manner in which users interact with Web content. Future display and interaction technology will bring AR out from the research institute into people's homes and workplaces. Such applications will bring unique levels of interaction, immersion, and collaboration not seen by more conventional media. Applications of this technology can be greatly enhanced through consideration of a number of design and usability concerns, with which Web developers may have varying levels of experience. Thus, it may be prudent to involve human factors practitioners in Web development efforts. By appropriately addressing such concerns during the Web development life cycle,
35. Augmented Reality a more effective application can result that is readily adopted by users.
ACKNOWLEDGMENTS This material is based on work supported in part by the National Science Foundation (NSF) under Grant No. IRI-9624968,
•
655
the Office of Naval Research (ONR) under Grant No. NOOO14-981-0642, the Naval Air Warfare Center Training Systems Division (NAWC TSD) under contract No. N61339-99-C-0098, and the National Aeronautics and Space Administration (NASA) under Grant No. NAS9-19453. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views or the endorsement of the NSF, ONR, NAWC TSD, or NASA.
References Arial, M., & Schar, S. (2002). How do videoconferencing technologies support collaboration? State of the art. Technical report KTI-Project Nr. 4569.3 IMS. Zurich: Communication IT Project. Available: http:// www.iha.bepr.ethz.ch/pages/leute/arial/Communicate JT_repl.pdf Baecker, R., Small, I., &Mander, R. (1991). Bringing icons to life. Proceedings of ACM CHI '91 Conference on Human Factors in Computing (pp. 1-6). New York, LA: ACM Press. Barneld, W., & Danis, E. (1996). Comments on the use of olfactory displays for virtual environments. Presence: Teleoperators and Virtual Environments, 5, 109-121. Barneld, W., Zeltzer, D., Sheridan, T, & Slater, M. (1995). Presence and performance within virtual environments. In W Barneld & T. A. Furness III (Eds.), Virtual environments and advanced interface design (pp. 473-513). London: Oxford University Press. Bennett, D., Chapelle, B. D. L., Zeltzer, D., Bryson, S. T, & Bolas, M. (1996). The future of virtual reality: Head-mounted displays versus spatially immersive displays. Special Panel Session at SIGGRAPH '96, August 4-19, New Orleans, LA. New York: ACM Press. Biggs, S. ]., & Srinivasan, M. A. (2002). Haptic interfaces. In K. M. Stanney (Ed.), Handbook of virtual environments: Design, implementation, and applications (pp. 93-115). Mahwah, NJ: Lawrence Erlbaum Associates. Billinghurst, M., & Kato, H. (1999). Collaborative mixed reality. Proceedings of International Symposium on Mixed Reality (ISMR '99). Mixed reality—Merging real and virtual worlds, 261-284. Billinghurst, M., Kato, H., & Poupyrev, I. (2001). The MagicBook: Moving seamlessly between reality and virtuality. IEEE Computer Graphics and Applications, 2/0), 6-8. Blade, R. A., & Padgett, M. L. (2002). Virtual environment standards and terminology. In K. M. Stanney (Ed.), Handbook of virtual environments: Design, implementation, and applications (pp. 15-27). Mahwah, NJ: Lawrence Erlbaum Associations. Blauert, J. (1996). Spatial hearing. Cambridge, MA: MIT Press. Bowman, D. (1999). Interaction techniques for common tasks in immersive virtual environments: Design, evaluation, and application. Unpublished doctoral dissertation, Georgia Institute of Technology, Atlanta, GA. Bowman, D. (2002). Principles for the design of performance-oriented interaction techniques. In K. M. Stanney (Ed.), Handbook of virtual environments: Design, implementation, and applications (pp. 277-300). Mahwah, NJ: Lawrence Erlbaum Associates. Burrell, J., Gay, G. K., Kubo, K., & Farina, N. (2002). Context-aware computing: A test case. UBICOMP2002 [On-line]. Available: http:// www.hci.cornell.edu/LabArticles/ContextAware-Burrell.pdf Carter, J. P. (1992). The nose have it! Presence: Teleoperators and Virtual Environments, 1, 493-494. Cohen, M. (1992). Integrating graphic and audio windows. Presence: Teleoperators and Virtual Environments, 1, 468-481.
Cohen, M., & Wenzel, E. M. (1995). The design of multidimensional sound interfaces. In W. Barfield & T. A. Furness III (Eds.), Virtual environments and advanced interface design (pp. 291-346). London: Oxford University Press. Darken, R. R (1996). Wayfinding in large-scale virtual worlds. Unpublished doctoral dissertation, George Washington University, Washington, DC. Darken, R. P., & Sibert, J. L. (1996a). Navigating in large virtual worlds. International Journal of Human-Computer Interaction, 8, 49-72. Darken, R. R, & Sibert, J. L. (1996b). Wayfinding strategies and behaviors in large virtual environments. Human Factors in Computing Systems, CHI '96 Conference Proceedings, 142-149. New York: ACM Press. Davies, R. C. (2002). Applications of system design using virtual environments. In K. M. Stanney (Ed.), Handbook of virtual environments: Design, implementation, and applications (pp. 1079-1100). Mahwah, NJ: Lawrence Erlbaum Associates. Durlach, N. L, & Mavor, A. S. (1995). Virtual reality: Scientific and technological challenges. Washington, DC: Academic Press. Eberts, R. (1994). User interface design. Englewood Cliffs, NJ: Prentice Hall. Esposito, C. (1996, April). User interfaces for virtual reality systems. Human Factors in Computing Systems, CHI '96 Conference Tutorial Notes. New York: ACM Press. Fjeld, M., Lauche, K., Bichsel, M., Voorhorst, E, Krueger, H., & Rauterberg, M. (2002). Physical and virtual tools: Activity theory applied to the design of groupware. Computer Supported Cooperative Work (CSCW): Activity Theory and the Practice of Design, 11(1-2), 153-180. Available: http://www.fjeld.ch/pub/cscw.pdf Fjeld, M., Schar, S. G., Signorello, D., & Krueger, H. (2002). Alternative tools for tangible interaction: A usability evaluation. International Symposium on Mixed and Augmented Reality (ISMAR'02) (pp. 157-166). Darmstadt, Germany: IEEE Computer Society Press. Fontaine, G. (1992). Experience of a sense of presence in intercultural and international encounters. Presence: Teleoperators and Virtual Environments, 1, 482-490. Fritsch, D., Klinec, D., & Volz, S. (2001). NEXUS—Positioning and data management concepts for location-aware applications. Computers, Environment and Urban Systems, 25, 279-291. Gabbard, J. L., & Hix, D. (1997). A taxonomy of usability characteristics in virtual environments [On-line]. Available: http://csgrad. cs.vt.edu/~jgabbard/ve/taxonomy/ Haala, N. (2001). Automated image orientation in a location aware environment. In D. Fritsch & R. Spiller (Eds.), Photogrammetric Week '01 (pp. 255-262). Heidelberg, Germany: Herbert Wichmann Verlag. Hannaford, B., & Venema, S. (1995). Kinesthetic displays for remote and virtual environments. In W Barfield & T. A. Furness III (Eds), Virtual environments and advanced interface design (pp. 415-436). London: Oxford University Press.
656
STANNEY AND DAVIES
Hapeshi, K., & Jones, D. (1992). Interactive multimedia for instruction: A cognitive analysis of the role of audition and vision. International Journal of Human-Computer Interaction, 41, 79-99. Hinckley, K., Pausch, R., Goble, J. C, & Kassell, N. F. (1994). A survey of design issues in spatial input. Proceedings of ACM Symposium on User Interface Software & Technology (UIST '94), 213-222. New York: ACM Press. Ishii, H., & Ullmer, B. (1997). Tangible bits: Towards seamless interfaces between people, bits and atoms. Proceedings of CHI '97, 234-241. New York: ACM Press. Jacob, R., Sibert, L. E., McFarlane, D. C., & Mullen, M. P., Jr. (1994). Integrality and separability of input devices. ACM Transactions on Computer-Human Interaction, 1, 3-26. Kaczmarek, K. A., & Bach-Y-Rita, P (1995). Tactile displays. In W. Barfield & T. A. Furness in (Eds.), Virtual environments and advanced interface design (pp. 349-414). London: Oxford University Press. Kalawsky, R. S. (1993). The science of virtual reality and virtual environments. Wokingham, England: Addison-Wesley. Kalawsky, R. S. (1999). VRUSE—A computerized diagnostic tool: For usability evaluation of virtual/synthetic environments systems. Applied Ergonomics, 30, 11-25. Kaur, K. (1999). Designing virtual environments for usability. Unpublished doctoral dissertation, City University, London. Kennedy, R. S., Lane, N. E., Berbaum, K. S., & Lilienthal, M. G. (1993). Simulator sickness questionnaire: An enhanced method for quantifying simulator sickness. International Journal of Aviation Psychology, 3, 203-220. Krapichler, C., Haubner, M., Losch, A., Schuhmann, D., Seemann, M., & Englmeier, K-H. (1999). Physicians in virtual environments— Multimodal human-computer interaction. Interacting with Computers, 11, 427-452. Liang, J., & Green, M. (1994). JDCAD: A highly interactive 3D modeling system. Computers and Graphics, 18, 499-506. Lockner, M., & Winroth, U. (1999). Usability and social interaction in collaborative virtual environments. Technical Report, Project P8097, JUPITER2: Joint Usability, Performability and Interoperability Trials in Europe. Available: http://www.eurescom.de/~publicwebspace/P800-series/P807/results/Usability/R7/D2-T3-UsabilityR7-SocialInteraction.pdf Lynch, K. (I960). The image of a city. Cambridge, MA: MIT Press. MacKenzie, I. S. (1995). Input devices and interaction techniques for advanced computing. In W Barfield & T. A. Furness (Eds.), Virtual environments and advanced interface design (pp. 437-440). London: Oxford University Press. McNeely, W A., Burdea, G. C., Hannaford, B., Hirose, M., Jacobsen, S., Salisbury, K., & Tachi, S. Sinclair & Martinez (1995, March). Panel: Whither force feedback. IEEE Virtual Reality Annual International Symposium (VRIAS '95). Los Alamitos, CA: IEEE Computer Society Press. Millard, D. E., Moreau, L., Davis, H. C., & Reich, S. (2000). FOHM: A fundamental open hypertext model for investigating interoperability between hypertext domains. Proceedings of the '00 ACM Conference on Hypertext, 93-102. New York: ACM Press. Miller, T, & Zeleznik, R. (1999). The design of 3D haptic widgets. Symposium on Interactive 3D Graphics (pp. 97-192). New York: ACM Press. Milosavljevic, M., Dale, R., Green, S., Paris, C., & Williams. S. (1998). Virtual museums on the information superhighway: Prospects and potholes. Proceedings of CIDOC'98: Annual Conference of the International Committeefor Documentation of the International Council of Museums. Available: http://www.dynamicmultimedia.com.au/ papers/cidoc98/
Mine, M. R., Brooks, E P., Jr., & Sequin, C. (1997). Moving objects in space: Exploiting proprioception in virtual environment interaction. SIGGRAPH '97 Conference Proceedings, 19-26. New York: ACM Press. Money, K. E. (1990). Motion sickness and evolution. In G. H. Crampton (Ed.), Motion and space sickness (pp. 1-7). Boca Raton, FL: CRC Press. Muller, S., Strieker, D., & Weidenhausen, J. (2001). An augmented-reality system to support product development, production, and maintenance. CG Topics, 3, 16-17. Available: http://www.inigraphics.net/ publications/topics/200 l/issue3/3-01 a07.pdf Posner, M. I. (1976). Chronometric explorations of the mind. Mahwah, NJ: Lawrence Erlbaum Associates. Riccio, G. E., & Stoffregen, T. A. (1991). An ecological theory of motion sickness and postural instability. Ecological Psychology, 3,195240. Richard, P., Birebent, G., Coiffet, P., Burdea, G., Gomex, D., & Langrana, N. (1996). Effect of frame rate and force feedback on virtual manipulation. Presence: Teleoperators and Virtual Environments, 5, 95-108. Rushton, S., Mon-Williams, M., & Warm, J. P. (1994). Binocular vision in a bi-ocular world: New generation head-mounted displays avoid causing visual deficit. Displays, 15, 255-260. Sadowski, W., & Stanney K. M. (2002). Measuring and managing presence in virtual environments. In K. M. Stanney (Ed.), Handbook of virtual environments: Design, implementation, and applications (pp. 791-806). Mahwah, NJ: Lawrence Erlbaum Associates. Sanders, M. S., & McCormick, E. J. (1993). Human factors in engineering and design (7th ed.). New York: McGraw-Hill. Shneiderman, B. (1992). Designing the user interface (2nd ed.). Reading, MA: Addison-Wesley. Sinclair, P. A. S., & Martinez, K. (2002). Tangible hypermedia using the ARToolKit. Proceedings of the First IEEE International Augmented Reality Toolkit Workshop. Available: http://www.ecs. soton.ac.uk/~pass99r/research/sinclair-art02.pdf Slater, M., & Wilbur, S. (1997). A framework for immersive virtual environments (FIVE): Speculations on the role of presence in virtual environments. Presence: Teleoperators and Virtual Environments, 6, 603-616. Spohrer, J. (1998). WorldBoard: What comes after the WWW? [On-line]. Available: http://www.worldboard.org/pub/spohrer/wbconcept/ default.html Stanney, K. M., Chen, J. L., Wedell, B., & Breaux, R. (2003). Identification of metaphors for interactive training systems. Ergonomics, 46, 197-219. Stanney, K. M., Mollaghasemi, M., Reeves, L., Breaux, R., & Graeber, D. A. (2003). Usability engineering of virtual environments (VEs): Identifying multiple criteria that drive effective VE system design. International Journal of Human-Computer Studies, 58, 447481. Stanney, K. M., Salvendy, G., Deisigner, J., DiZio, P, Ellis, S., Ellison, E., Fogleman, G., Gallimore, J., Hettinger, L., Kennedy, R., Lackner, J., Lawson, B., Maida, J., Mead, A., Mon-Williams, M., Newman, D., Piantanida, T, Reeves, L., Riedel, O., Singer, M., Stoffregen, T, Wann, J., Welch, R., Wilson, J., Witmer, B. (1998). Aftereffects and sense of presence in virtual environments: Formulation of a research and development agenda. Report sponsored by the Life Sciences Division at NASA Headquarters. International Journal of Human-Computer Interaction, 10(2), 135-187. Steed, A., & Tromp, J. (1998). Experiences with the evaluation of CVE applications. Proceedings of Collaborative Virtual Environments Conference (CVE'98). Available: http://www.cs.ucl.ac.uk/ staff/A. Steed/steed-tromp-cve98.pdf
35. Augmented Reality
Stoakley, R., Conway, M. J., &Pausch, R. (1995). Virtual reality on a WIM: Interactive worlds in miniature. Proceedings of CHI'95 Mosaic of Creativity (pp. 265-272). New York: ACM Press. Sturman, D. J., Zeltzer, D., & Pieper, S. (1989). Hands-on interaction with virtual environments. Proceedings of ACM Symposium on User Interface Software & Technology (UIST '89), 19-24. New York: ACM Press. Takagi, A., Yamazaki, S., Saito, Y, & Taniguchi, N. (2000). Development of a stereo video see-through HMD for AR systems. Proceedings of the 3rd IEEE/ACM International Symposium on Augmented Reality (pp. 68-77). Los Alamitos, CA: IEEE Computer Society Press. Tromp, J. (1999). Usability design for CVEs. Technical Report ACTS Project AC040-NOT-CRG-DS-P-029. Available: http://coven.lancs.ac. uk/4/deliverables/del29.pdf Tromp, J. G. (1997). Methodology for distributed usability evaluation in collaborative virtual environments. Proceedings of 4th UK VR SIG Conference. Available: http://www.brunel.ac.uk/faculty/tech/ systems/groups/wr/vrsig97/proceed/0l6/0l6.htm Vince, J. (2004). Introduction to virtual reality (2nd edition). Berlin: Springer-Verlag. Warm, J. P., & Mon-Williams, M. (2002). Measurement of visual aftereffects following virtual environment exposure. In K. M. Stanney (Ed.), Handbook of virtual environments: Design, implementation, and
657
applications (pp. 731-749). Mahwah, NJ: Lawrence Erlbaum Associates. Ware, C, & Balakrishnan, R. (1994). Reaching object in VR displays: Lag and frame rate. ACM Transactions on Computer-Human Interaction, 1, 331-356. Wellner, P. (1993). Interacting with paper on the Digitaldesk. Communications of the ACM, 36(7), 86-96. Wickens, C. D., & Baker, R (1995). Cognitive issues in virtual reality. In W Barneld & T. A. Furness III (Eds.), Virtual environments and advanced interface design (pp. 516-541). London: Oxford University Press. Witmer, B., & Singer, M. (1998). Measuring presence in virtual environment: A presence questionnaire. Presence: Teleoperators and Virtual Environments, 7,225-240. Zeltzer, D. (1992). Autonomy, interaction and presence. Presence: Teleoperators and Virtual Environments, 1,127-132. Zhai, S., & Milgram, P. (1994). Input techniques for HCI and 3D environments. Human Factors in Computing Systems, CHI '94 Conference Proceedings Companion, pp. 85-86. New York: ACM Press. Zhai, S., Milgram, P, & Buxton, W. (1996). The influence of muscle groups on performance of multiple degree-of-freedom input. Human Factors in Computing Systems, CHI '96 Conference Proceedings, 308-315. New York: ACM Press.
36 THE EMERGENCE OF e-HEALTH IN HEALTH CARE Kevin P. Moloney Francois Sainfort ]ulie A. ]acko V. Kathlene Emery Georgia Institute of Technology
INTRODUCTION There is little doubt that the U.S. health care system has its share of problems (Sainfort, Jacko, & Booske, 2003). Although the United States uses the some of the most advanced medical technologies, has the largest medical workforce, and spends the largest proportion of its gross domestic product, the World Health Organization (2000) rated the U.S. health care system worse than most of the Western world, with respect to quality and performance. In addition, there has been extensive documentation of widespread errors that have resulted in avoidable injuries to patients (Institute of Medicine, 2001). In an effort to rectify several of these problems, there has been a push to develop better health care information systems. Properly designed and implemented technologies have the ability to improve quality of health care (Raymond & Dold, 2002), decrease health care costs (Meyer, Kobb, & Ryan, 2002; Vacarro, Cherry, Harper, & O'Connell, 2001), prevent medical errors (Institute of Medicine, 2001), and support the ever-growing demands placed on the health care industry by governmental regulation and health care consumers. Although there has been support for new technologies in health care, these projections need to be tempered by a recognized need for development and evaluation fed by scientific study. e-Health, the conversion of health care services, products, and information to Web-based technologies, has been one
658
proposed solution to many of these problems. However, e-Health technologies (thus far) have failed to deliver on these promises. Although many of the barriers that have contributed to the feilure of e-Health will require larger changes in governmental regulations and policy changes by third-party payers and insurance companies, we believe that Human Factors and human-computer interaction (HCI) research can improve the success of and support for these technologies. HCI, a branch of Human Factors, can improve e-Health technologies by providing valuable insights into the processes and needs of individuals who use these systems. Characteristics of the users, including their abilities, goals, and perceptions, impact the quality and success of their interaction with technologies. HCI focuses on these issues in order to understand how users interact with technologies and how to increase the usefulness and effectiveness of the system (see Dix, chap. 3, this volume). In this chapter, we provide a review of e-Health, discuss how e-Health can contribute to the improvement of the U.S. health care system, and examine barriers to the advancement of e-Health. We focus only on the health care industry of the United States for the sake of brevity, although most of the concepts (particularly with regard to HCI) discussed throughout this chapter can also be applied to health care in other countries. In addition, we discuss the importance of Human Factors and HCI in facilitating the development and implementation of e-Health technologies in health care.
36. Emergence of e-Health
THE EVOLVING NEEDS OF INFORMATION SYSTEMS IN HEALTH CARE Before exploring e-Health and the Human Factors issues that will drive e-Health, it is worthwhile to take a look back at how the health care industry has changed and how these changes have affected the requirements of health care information systems. There have been a number of social, technological, economic, and regulatory factors that have contributed to the current state of health care in the United States and the inadequacy of the information technologies in this industry. The Institute of Medicine (2001) outlined four underlying causes of the inadequate quality of health care in the United States, including a relatively poorly designed delivery system, the increase in individuals with chronic conditions, the growing complexity of science and technology, and constraints on using the products of the revolution in information technology (IT). In addition, Sainfort et al. (2003) proposed increased consumerist behavior as another challenge in health care. Table 36.1 provides an overview of several factors that have contributed to significant changes in the health care industry. These factors have all contributed to the transformation of the U.S. health care system, resulting in increased demands placed on health care information systems and related technologies. The majority of health care information systems currently in use have become insufficient, which has generated interest in creating new applications and systems that take advantage of current technologies. Several of the factors that have created the need for and interest in new technologies are now discussed.
Regulatory and Economic Changes Affecting Health Care There have been a number of regulatory and economic factors that have changed health care, ranging from the transformation of organizational and delivery models of health care to the changing responsibilities and protocols of health care providers TABLE 36.1. Factors That Have Transformed the Health Care Industry Factor Category
Examples
Regulatory and economic
• Effects of the Medicare, Medicaid, and the HMO act • Rising health care expenditures • Managed care business model • HIPAA reform • Paradoxical effect of regulations on technology use
Social
* Increasing number of patients with chronic illness • General aging of U.S. population • Push toward preventative medicine • Social demand for control of and access to information
659
with respect to the handling of electronic medical data. As is seen, these changes have had an impact on the delivery of health care, which has ultimately affected the technological needs of health care organizations. The nature of the today's health care delivery system is much different from that of several decades ago. In the 1960s and 1970s, medical treatment was focused on acute care, the payment structure was fee for service, health insurance was scarcely available for those below the middle class, and hospitals and medical centers (largely associated with universities) were the primary settings for health care delivery (Safran & Perreault, 2001; Shi & Singh, 2001). To solve many of these economic and accessibility of care issues, the federal government expanded the Social Security Act, creating Medicare and Medicaid (Potter & Longest, 1994). However, this dramatically increased governmental spending on health care, which resulted in the creation of the Health Maintenance Organization (HMO) Act of 1973. This encouraged a more competitive, corporate market for health care delivery, which had the overall effect of spawning a new organizational and business model in the health care industry. The managed care model was developed to help manage resource utilization and control costs through the integration of health care functions (e.g., delivery and payment) within a single organizational setting. Managed care organizations (MCOs) accomplish this by organizing providers into coherent networks, sharing risks, leveraging the services of different facilities, extracting discounts based on large patient volumes, and eliminating payer and insurance intermediaries (Shi & Singh, 2001). MCOs naturally transformed into massive integrated delivery networks (IDNs see Safran & Perreault, 2001, for a review), creating geographically dispersed care networks comprised of alliances between care facilities such as ancillary and ambulatory care facilities, health maintenance and insurance companies, larger health care facilities (e.g., hospitals and medical centers), and extended care facilities (Fig. 36.1). Although the transition to the managed care model was largely based on economic reasons, it also created changes in the practices of health care professionals, affecting the workflow and practices of physicians and health care professionals. Primarily, the onset of managed care pushed primary care out
FIGURE 36.1. Organizational structure of the integrated delivery network (IDN).
660
MOLONEY ET AL.
of the hospitals and into outpatient clinics and physician practices. As a result, the major burden of health care, especially with the focus on preventative medicine, has fallen onto the primary care physician. So, primary care physicians now tend to have a higher patient volume, resulting in less time spent with patients and more severe knowledge requirements. In addition, the new IDN organizational model of distributed care facilities requires that health care organizations meet new levels of communication and data sharing in order to support collaboration. As can be seen, these economic and regulatory factors have had longrange impacts on the way that medicine is practiced and how health care is delivered. As a result, these changes to medical practices have also changed the technological needs of health care organizations. Another more recent regulatory factor that has affected health care organizations is the Department of Health & Human Services' (DHHS) Health Insurance Portability and Accountability Act (HIPAA) of 1996. The HIPAA legislation has dramatically altered provider and health care organization liability and responsibility with respect to the integrity, confidentiality, and availability of electronic health information (U.S. Department of Health & Human Services, 2003a). This legislation impacts the protocols used by health care organizations for transaction coding, data storage and transmission, documentation standards, and insurance claims and payments. These regulatory factors have also changed the technological needs of health care organizations, as most current systems do not yet meet the standards set out by these mandates.
Social Changes Affecting Health Care In addition to the regulatory and economic factors that have changed health care practices and technological needs of health care organizations, there are also social changes that have a large impact on health care. Two main categories of these social changes are (a) epidemiologic and demographic, and (b) attitudinal and consumerist. These epidemiologic and demographic factors have restructured the medical needs of the general population, reinforcing the need for the prevention and maintenance paradigms of care. The attitudinal and consumerist factors have reinforced the needs (and demands) for new medical technologies and health care services to provide more convenience and increased quality of care. These two categories of social factors are discussed in the following sections. Demographic and Epidemiologic Factors. Two major demographic and epidemiologic changes in the U.S. population that have affected health care are the increasing number of individuals who are elderly (e.g., 65 years and older) and the increasing prevalence of individuals with one or more chronic ailments. Aging of the population and poor health habits have resulted in increases in the number of individuals with chronic illnesses. As is seen, these factors are interrelated and have had severe economic consequences for the health care industry. The "baby boomer" contingent, numbering 82 million, will dramatically increase the proportion of the U.S. population ages 65 years and older within the next 15 years (Meyer, 2001).
Much of this aging (or elderly) population, as well as millions of younger individuals, have one or more chronic ailments, including arthritis, congestive heart disease, cancer, and Parkinson's disease (just to name a few). The prevalence of chronic conditions increases with age, primarily occurring in individuals ages 45 and older, with increasing rates at ages 65 and older (Partnership for Solutions, 2002). It is estimated that, in 2000, 125 million individuals in the United States had at least one chronic condition. This number is expected to increase 25% by the year 2020 (Mollica & Gillespie, 2003). The cost of health care for individuals with chronic conditions comprises approximately 78% to 85% of all health care spending (Mollica & Gillespie, 2003; Strohecker, 2000). To better explain the severity of this situation for the health care industry, these 125 million (or so) individuals account for about 80% of $1.5 trillion in health care expenditures (2002 estimate) in the United States. Clearly, providing for individuals with chronic conditions will have a strong impact on the economic future of health care. This economic impact has implications on the drive toward technological innovations in health care because health care organizations will seek solutions that can provide cost-effective care for these individuals. The increased probability of these individuals to experience an acute event requiring expensive treatment dictates that alternative forms of care be provided to help prevent excessive hospitalization, encourage behavioral modification to promote compliance with treatment orders, and provide more frequent support. The paradigm shift toward preventative medicine dictates that these individuals should be provided with means to receive extended support and care to improve or maintain noncritical health states. The increasing care needs of these individuals created new technological needs to support this paradigm of continual, active care delivery.
Attitudinal and Consumerist Factors, in addition to epidemiologic changes, there have been other social changes that have created new technological demands. The attitudinal and consumerist changes include the general trend toward increasing consumer demand of health information and control over one's own health care. Namely, the World Wide Web and communication technologies have reformed the way in which individuals conduct business, communicate, manage their assets and information, and further their education (Jadad, 1999; Mittman & Cain, 1999). This, in turn, has changed health care consumer needs, expectations, and behaviors, which has been shown to modify their interaction with physicians and health care professionals (Giorgianni, Grana, & Sewell, 2000) and their knowledge of and demand for new health care services, treatment options, and insurance plans (Mittman & Cain, 1999)- Increased consumerism, in the form of increased demands for information, services, and convenience, has raised a number of technological concerns. Reliance on technologies and the Web for everyday tasks has become commonplace for many individuals. As a result, many individuals seek to use the same technologies that they use for online shopping and investment management for the management of medical records and health insurance and shopping for health care products. Today, because of increased access to
36. Emergence of e-Health information, people generally have more control over the health insurance plan that they choose, the physician that they see, and even the treatment options that they receive. Information technologies have fueled a change in consumer behavior and expectations—individuals want, demand, and seek out more (and better) choices (Sainfort et al., 2003). These technologies empower consumers and will continue to influence their involvement with and decisions about their health care.
The Shifting Technological Needs of Health Care As discussed, there have been a number of factors that have contributed to major changes in the technological needs of health care organizations, ultimately resulting in information systems that are ill equipped to meet these needs. The new technological needs are a result of fundamental changes in the care delivery requirements, financial structure, consumer demands, regulatory forces, or organizational model of health care that dictated changes in the practical needs (e.g., cost savings, extra care for the chronically ill, modified data handling protocols) of the health care industry. The geographically dispersed nature of the IDN organizational model necessitates that health care services entities (e.g., physician practices and insurance providers) become much more collaborative. However, most health care applications, built on proprietary platforms and legacy mainframe systems, primarily handle only the needs of the individual provider or department within the health care organization (Shi & Singh, 2001). Information systems, even within a single health care facility (e.g., a hospital), still rely on complicated and costly interface engines that translate the different data content and formats from the disparate sending and receiving systems (Tang & McDonald, 2001). Thus, although the coordination, sharing, and management of information, resources, and business operations can achieve both increasing quality of patient care and decreasing costs, the information systems currently in place do not support these needs. In fact, even today, it would be more accurate to say that the majority of the IDNs in the United States are still integrating rather than being already integrated (Safran & Perreault, 2001). Moreover, many health care organizations still perform a number of data transactions by phone or fax and still employ the use of extra personnel for medical coding and transcription. More recent regulatory constraints (namely, HIPAA) provide an extreme example of technological inadequacy because health care information systems are actually noncompliant with governmental mandates. This piece of regulation has required health care organizations to become compliant with definitive rules for the privacy, confidentiality, standards, and security of health care data and the transactions performed with these data. To meet these standards, most health care organizations have had to implement entirely new information systems, often with extraordinary costs. The social changes, with respect to increased care needs for the chronically ill and the increased demands for consumer control of and access to health care information, have also created new technological needs. To achieve the reduction in the cost
•
661
of care, increases in the quality of care, and extra education and social support for these individuals, health care organizations and providers need technological systems that can provide continuous, remote, targeted care to monitor health status, enforce treatment protocols, promote healthy behaviors, and provide support. Increased consumer demands for more convenient health services and more control of their own health care dictate that consumers are granted easier and more extensive access to health care information. In addition, as suggested by the Institute of Medicine (2001), the ever-growing complexity of science and medicine also dictates the need for new health care information systems. As Sainfort et al. (2003) stated, "The sheer volume of new health care science and technologies... is large today and has advanced much more rapidly than our ability to use and deliver them in a safe, effective, and efficient way" (p. 809). Advances in the knowledge and research in the genetic, biomedical, biochemical, and other fields have created a need for health care and medicine to come up with ways to store, share, visualize, communicate, and implement this information. Although these advances bring many benefits to improving the knowledge, treatments, drugs, and devices available to health care, they also bring enormous costs in terms of developing new technologies that can support all these new possibilities.
Summary The underlying current that has resulted from all these changes is an increased need for new technologies. Health care organizations are only beginning to apply information technologies and still do not make much use of the Web and other information technologies (Institute of Medicine, 2001; Sainfort et al., 2003). Although health care once had fee-for-service, point-ofcare, acute care delivery model, it now has a financially complex, heavily managed, preventative care model. Consumers once had a passive role in the management of their health care, but they are becoming empowered, demanding decision makers. To meet these increasing technological needs, health care information systems developers have turned to the Web and other networked technologies. These new technological needs spawned the movement toward e-health.
THE e-HEALTH MOVEMENT e-Health has become a highly touted, yet poorly supported, solution to many of these deficiencies with the current information systems in health care. e-Health has yet to be confined to one clear, concise definition (Eysenbach, 2001). Interestingly, some industry professionals believe that "e-Health" was a term created to boost corporate and investor interest in these technologies and now serves as a blanket term covering more traditional fields of telehealth, telemedicine, computer-based learning, remote consultation, and remote disease management, among others (Delia Mea, 2002; Mitchell, 1999). For our purposes, e-Health refers to the transition of health care processes, information, transactions, and the like to Web-based (or Weblike) form factors. Within this working definition of e-Health are all
662
MOLONEY ET AL.
networked health care information systems, information repositories, and medical applications, whether they are implemented in intranets, Extranets, or the public domain Internet. Common e-Health technologies include electronic medical records (EMRs), Web sites providing health care information, and order entry and billings systems, among a wide range of other applications and technologies. The reason that interest in e-health and new health care technologies has so greatly increased is twofold: (a) social, regulatory, organizational, and economic changes have dictated new technological needs in the health care industry (Singer, Enthoven, & Garber, 2001); and (b) these new technologies have shown promise in improving health care services, improving quality of care, decreasing costs, and changing traditional models of health care (Sainfort et al., 2003). Although the former point should follow naturally from the previous section, the potential benefits of these technologies are now discussed.
Why the Web? Although information technologies have been in use in health care since the 1960s, the industry has only more recently begun to really recognize the potential benefits of the World Wide Web and other new information technologies. One major question is "why is the Web being considered the platform on which to build the new technological infrastructure of health care?" It has been suggested that one of the ways in which Web applications are different from the failed health care IT ventures is that Web-based technology solutions engage the most powerful stakeholders in the health care industry—the millions of health care consumers (Kleinke, 2000). There are characteristics of the Web that make it an attractive infrastructure for e-health technologies. Mittman and Cain (1999) identified some beneficial characteristics, describing the Web as (a) inexpensive; (b) easy to use; (c) more democratic (e.g., provides an effective way to share information and viewpoints by many); (d) boundless (e.g., information can come from and be accessed anywhere); and (e) continually increasing useful functionality. Some additional characteristics and their associated benefits are outlined in Table 36.2. The overarching domain of e-Health incorporates a lot of non-Web-based applications and technologies, which have their place and use in health care. For example, health care professionals use self-contained decision aids that are not networked to a larger database, and stand-alone EMRs that do not incorporate interfacing with order entry and billing systems. However, making the application and systems Web based increases the viability of these tools. For example, providing professionals with better (or more) information to aid in decision making can reduce logical errors and increase quality of care by providing caregivers with larger, comprehensive databases that are continually improved through evidence-based medicine, and then collectively updated and distributed. In addition, administrative costs can be reduced by interfacing the functionality of patient interview, billing, and order entry systems in order to decrease the amount of time and resources needed to transcribe physician notes, prepare and send paperwork, or look up and
TABLE 36.2. Overview of Potential Benefits of Web-Based Technologies Characteristic
Example Benefits
Multimedia support/ information visualization
• Better visualization of complex information • Support for multiple media form factors
Universalizability/ cross-compatibility
• Platform/form factor cross-compatibility • Immediate/consistent information updating
Flexibility/customization
• Scalable systems to fit changing needs • Customizable interfaces to suit user needs
Connectivity
• Interfacility collaboration/ communication • Variable access controls
Efficiency/inexpensive
• Decreased data reentry • Cost-effective data storage
Ubiquity/familiarity
• Decreased training/familiarity requirements • Decreased implementation/ maintenance costs
enter diagnostic and procedure coding. From these common examples, one can see how the interconnectivity and flexibility allowed by Web-based technologies can aid in the development of more powerful and useful tools.
e-Health's Brief History Web-based health care applications or information systems, or e-Health, were the natural result of a combination of factors. The relative success of e-commerce and Web-based transaction paradigms in decreasing business costs, the failure in properly coordinating care and resources between health care facilities, and the increasing consumer demands for more access and convenience with regard to their health care likely all contributed to players in the health care industry looking into Web-based technologies for solutions to these problems. Although it is difficult to even approximate a date of e-Health's birth, we can use the history of notable IT ventures in the health care domain as a starting point of Web-based applications in health care. First, it should be noted that e-Health, in the broader sense of technology-enabled health care tools, largely stemmed from the interest in applications of telehealth or telemedicine (which actually started in the early 1900s; Rosen, 1997). The health care industry became interested in remotely delivered health care services (e.g., phone consultation, transfer of radiographic imagery) in order to better leverage resources and provide service to rural areas and remotely located consumers—including astronauts (Bashshur & Lovett, 1977). Interest in these methods of providing care to patients remotely increased with the advent of the Internet, Web-based, and networking technologies in the 1980s and 1990s, spawning new possibilities for remote
36. Emergence of e-Health
•
663
TABLE 36.3. Taxonomy of Current Health Care Web Sites and Web-Based Ventures. Company Type Portal
Common Services • Provide information and links to health topics • Common ground for seeking and providing information for both consumers and providers
Examples • Medscape • drkoop.com • healthgrades.com
Connectivity
• Provides quality assessments of providers • Web-accessible EMRs • Provides services for data and claims transactions
• Healtheon/WebMD • XCare.net • MedicaLogic
Business-to-business
• Provides marketplace for buying and selling medical equipment • Facilitate procurement/selection of group health insurance plans
• Allscripts • Medical Manager
Business-to-consumer
• Customizable insurance plans • Online ordering of medications and health products
• drugstore.com • PlanetRx
consultation and care delivery, including telesurgery, videoconferencing, and remote monitoring through sensor technologies. e-Health naturally resulted from these conditions, as new technological innovations were being applied to health care and medicine. Starting in the mid- to late 1990s, and up to the present, the Web technologies have been used as a springboard for widespread access to "health e-commerce enterprises" (Parente, 2000). Currently, use of the Web-based technologies in health care focuses on Web sites that provide online services for providers and consumers (Parente, 2000), application of Web-based technologies to provide online health management functionality (LeGrow & Metzger, 2001), and use of wireless-enabled mobile devices for clinical and administrative functions (Turisco & Case, 2001). Table 36.3 provides a taxonomy, "Summarized from the work of Parente (2000)," of common health-related online ventures. There are some current examples of the adoption and use of new technologies, such as the use of handheld and portable technologies for reference tools, prescriptions, charge capture and billing, documentation and dictation, and functions of the EMR (Roa, Hoglund, Martucci, & Wilson, 2002). In addition, clinical decision support tools (e.g., protocol and treatment databases, medical research repositories, and analytical decision aid tools) are also becoming popular (Metzger & MacDonald, 2002). Although application of Web-based technologies in health care is still not widespread and does not take full advantage of the available technology, providers and professionals are beginning to see the utility of these technologies.
Benefits of e-Health Although there is some contention about how or when (or even if) e-Health can fix the ailing U.S. health care industry, there is little argument that the industry is ailing: health care costs are outrageous, medical errors are abundant, resources are used inefficiently, and the managed care system (thus far) has generally failed (see Sainfort et al., 2003, for a review). Even though
predictions need to be guarded, e-health has the potential to improve many of health care's problems. Researchers and experts have suggested that application of information technologies can benefit health care in several domains, such as (a) consumer health; (b) professional education, (c) clinical care, (d) public health, (e) administrative and financial transactions, and (f) and research (National Research Council, 2000). We briefly discuss each of these application domains. The following sections outline a number of areas in which e-Health is well positioned to benefit health care in the near future. Consumer Health. We discuss the consumer health domain first because it is the most likely domain in which the application of Web-based technologies can immediately begin to benefit and reshape health care. In this context, consumer refers to individuals who seek, use, or need health care information or services (e.g., a patient). Consumer health refers to efforts in giving consumers a more active, direct role in the management of their own health and health care (National Research Council, 2000). e-Health technologies will contribute to consumer health largely through increasing access to health care and medical information. Harris (1995) defined consumer health information as "any information that enables individuals to understand their health and make health-related decisions for themselves or their families" (p. 23). Although this definition clearly encompasses a lot of non-Web-based information dissemination methods, it should also be noted that these electronic forms of information have a larger potential to be more timely, comprehensive, and accessible to an increasingly larger base of health care consumers. In support of this idea of the relative effectiveness of Webbased technologies as information dissemination and access tools, it has been estimated that 70 to 110 million Americans look on the Web for health-related information (Cain, Mittman, Sarasohn-Kahn, & Wayne, 2000; Giorgianni et al., 2000; Taylor, 2002). With an estimated 100,000 Web sites devoted to health (Ferguson, 2000), health care consumers have access to seemingly unlimited health-related information. According to studies performed by the American Medical Association and the Pew
664
MOLONEY ET AL.
Internet Project, about 6 million Americans seek health information online on an average day, which exceeds the number of individuals that actually visit health care professionals (Fox & Rainie, 2002). Clearly, consumers are taking advantage of the information provided on the Web. In evidence of this behavioral change is the frequency with which patients now present their physicians with printouts of information from the Web (Giorgianni et al., 2000; Kassirer, 2000). Consumers are using the Web and other technologies to research diseases, signs and symptoms, treatment options, new Pharmaceuticals, or even seek advice from other health care specialists, peers, and professionals online. In addition, the Web also allows for the education of the online public about healthconscious behaviors and provides social support (via virtual communities and support groups) for individuals with chronic illnesses. Consumers are realizing, likely through their experience with advent of e-commerce and online shopping, that the Web provides them with a means of getting more information, better information, comparative data, knowledge of alternatives, and even sharing information to leverage collective knowledge.
Professional Education. As suggested by the Institute of Medicine (2001), one key challenge in medicine is the ever-growing volume and complexity of knowledge, tools, and information in medicine and science. It often proves difficult for clinicians and health care professionals to stay current on scientific research findings, new Pharmaceuticals, effectiveness of treatment protocols, and new diagnostic and treatment tools, let alone all the new bureaucratic and legislative regulations that are to be enforced. Although, in some ways, e-Health technologies may contribute to this growing complexity in health care, it can also do much to facilitate the training and education of health care professionals and consumers. The education of health care professionals, ranging from physicians to laboratory technicians, can be greatly enhanced through the use of Web-based technologies. As with consumer education, professional education will be enhanced primarily through the mere availability of more (and better) information through the information superhighway. For example, the Web site of the National Library of Medicine provides a searchable database of more than 10 million references from more than 4,000 medical and biomedical journals (U.S. National Library of Medicine, 2002). Yet, this is just one Web site out of more than 100,000 health-related Web sites (Benton Foundation, 1999; Ferguson, 2000). Information availability has led to increasingly large numbers of health care professionals and consumers who are seeking out health-related information online. The boundless nature of the Web not only provides easy access to more textual information, scientific research results, and clinical outcomes, but it also provides health care professionals with an opportunity to view videoconference sessions of surgical procedures, examine simulations of medical procedures, and more easily interact with their peers. In addition, the Web provides unparalleled abilities of information visualization, which has been shown to greatly facilitate learning and knowledge crystallization (see Card, 2003, for a review). Although it is true
that all these experiences can be achieved with computers and software, without connection to any network, the Web allows for increased information sharing and collaboration. Web-based technologies allow for health care professionals, students, and researchers to better leverage their collective knowledge to improve their education. Clinical Care. One of the most tangible and investigated potential benefits of e-Health is the use of Web-based technologies (e.g., the Web, e-mail) to enhance the provider-patient interaction. This interaction between patient and physician is the key interaction in health care because it involves the two major decision makers in the maintenance and management of an individual's health. As previously discussed, use of the Web as a repository of health-related information has helped consumers achieve a more active role in their health care and health-related decisions. Providing consumers with better, more accessible sources of information has increased the quality of the dialog during encounters and helps patients to better use the limited time they have with their physicians (Giorgianni et al., 2000; MacDonald, Case, & Metzger, 2001). e-Health technologies have also provided health care professionals with clinical support and communication tools, such as EMRs, computerized order entry systems, information databases and query tools, electronic decision support tools, warning/alert systems, and sensor and data analysis tools (Sainfort et al., 2003). These applications of technology decrease medical errors and increase the quality of care by assisting the physician in applying the most appropriate treatment protocol, avoiding drug interactions and allergies, keeping up to date on the most recent information on a patient's condition, and supporting the physician in decision making. Web-based technologies, such as e-mail, have provided health care professionals with new tools to help facilitate effective communication and increase patient compliance to treatment protocols. For example, the use of e-mail between patients and their physicians helps physicians to remain abreast of a patient's current health status, reinforce treatment orders and protocols, and allows for a more efficient, flexible outlet for communication with patients who do not need synchronous, face-to-face consultation (MacDonald et al., 2001). In an age where physicians spend less and less time in direct physical contact with their patients, these technologies help provide new tools that transform care delivery from a paradigm based on discrete interaction dependent on physical proximity to one that is based on continuous, remote care delivery. This more continual contact between patient (or consumer) and practitioner has tremendous potential to increase quality of care by allowing for more frequent reinforcement of treatment protocol and health-promoting behaviors, monitoring of health status, and early notification and resolution of potential health issues. Public Health. e-Health can improve public health, primarily through using Web-based technologies to increase the collection and distribution of information between public health officials, health care providers, and public healthrelated organizations [e.g., the Centers for Disease Control and
36. Emergence of e-Health Prevention (CDC)]. Public health refers to the epidemiologic trends, such as spread of infectious disease, outbreaks of food poisoning, or dense areas of the chronically ill, on the community (e.g., city, county), state, and federal levels. These technologies will contribute to public health primarily through the rapid, easy, continual transmission of relevant public health data from health care practitioners to public health organizations and officials (and vice versa). This transmission of data can help with preventative planning (e.g., targeting education at affected populations), determining resource allocation and planning (e.g., redirecting more resources or supplies to needier areas), and general public health surveillance (e.g., monitoring public health trends for pathogens, communicable disease, and chronic illness). There have been long-standing issues with the communication within the community, state, and federal levels of public health agencies and between these officials and health care providers or clinicians (National Research Council, 2000). The levels of agencies are organized inefficiently (vertically and disease specific), limiting effective, complete, and timely communication. In fact, much of the reporting from clinicians and medical laboratories is still paper based. In support of the need for Web-based technologies to improve this communication, the use of these technologies has, in some cases, drastically limited the delay of information transmission between health care providers or medical laboratories and state- and federal-level public health officials. In addition, this clearly has cost-savings implications to reduce costs by limiting paper waste, streamlining administrative costs, and improving resource allocation. The use of Web-based technologies and electronic forms will decrease the rates of reports being lost, misdirected, or misunderstood (due to poor handwriting), helping to feed those agencies responsible for promoting public health with higherquality information in a more timely fashion.
Administrative and Financial Transactions. Streamlining and reducing the costs associated with administrative process and financial transactions is one application domain in which Web-based technologies could easily produce immediate, tangible improvements. For example, researchers have estimated that the use of electronic claims have reduced the cost of these transactions from several dollars to a few cents (McCormack, 2000). Although the managed care model of health care delivery was originally designed to control costs, the system has spiraled out of control as the majority of health care in the United States is financed by a huge network of third-party payers that handle the insurance, payment, and reimbursement of health services (National Research Council, 2000). The costs of these interactions between consumers with their providers (e.g., health insurance companies, health management organizations) and between these providers with the consumers' care deliverers (e.g., clinicians, hospitals) account for a larger proportion of U.S. health care expenditures. Streamlining these processes through electronic claim filing and processing can reduce extra costs associated with small and private practices having to afford expensive electronic data interchange (EDI) systems, medical practices and payers having to hire extra personnel to
665
simply transfer information by fax or phone, and the extra processes associated with data reentry to ensure claim formats are compatible between medical practices and associated payers. The Web has greatly benefited many nonhealth sectors of the economy by using these technologies to streamline administrative processes and decrease operating costs. It is hoped that many of these advantages will carry over into the health care domain. Although there are a number of ferees aligned against this progress (e.g., traditional players of the "old" health care, federal, and state legislation, and an ailing economy), there is still a clear interest in developing and trying to market and implement Web-based technologies in health care (Kleinke, 2000; Parente, 2000). Only time will tell if e-health technologies will succeed where governmental regulation, managed care, and other IT promises have all failed in taming the uncontrollable costs of health care. Research. Computing technologies have provided the medical and scientific research communities with new mass storage capabilities, methods of visualization and management of large amounts of data, and means of collection and analyzing data. Web-based technologies can continue to build on these strengths by allowing for effective methods of information dissemination and more widespread access to current scientific knowledge. Science and research are primarily performed by individuals or small groups, which can be located around the world. The way that science advances is for the scientific findings of researchers to be shared with other researchers in an infinite self-feeding process. Web-based technologies are well equipped to facilitate this sharing of information, ensuring that new findings and data can be shared in a much more timely and continuous fashion, as opposed to the discrete and lengthy processes of disseminating new results through publication in paper journals and presentations at meetings or conferences. Many domains of scientific research have already realized the benefits of this more rapid and accessible medium for information dissemination because the number of online publications for research has continually increased. As discussed with how Web-based technologies can improve professional education, these technologies can also improve medical research by allowing for effective ways to manage and visualize complex information. The flexible, dynamic nature of the Web allows for multimedia presentation of information (e.g., manipulability of three-dimensional models of proteins or anatomic structures) and the building of large, distributed databases (e.g., the construction and development of "gene banks"). The increasing volume and complexity of data coming from biomedical and scientific research requires that new tools are developed for researchers to share this information. e-Health can help research to advance through the support of these needs.
Summary Although information systems and closed networks have long been used in the health care industry, the Web is beginning
666
•
MOLONEY ET AL.
to make its mark on the reformation of the industry. e-Health has the potential to change the current state of the U.S. health care industry through cost reduction, resource management, and improvements to quality of care, reduction of medical errors, augmentation of provider/consumer education, and facilitation of the preventative model of care. e-Health technologies leverage the power, ubiquity, ease, and familiarity of the Web and other technologies to enhance health care. Although these technologies are well positioned to deliver these benefits, there are a number of barriers to their proliferation and success.
BARRIERS TO THE PROLIFERATION OF e-HEALTH As demonstrated in previous sections, the Web-based technologies are being targeted at, and even already being used for, a variety of applications in health care. However, use of Web-based technologies in health care is still limited when compared with other sectors of the economy that use these technologies, as well as when compared with the broad potential for application in the health care domain. There are a host of issues that contribute to this industry's sluggish pace in adopting, implementing, and using Web-based technologies. These issues include problems stemming from constrictive regulations and policies; economic issues with the costs of development, maintenance, and implementation; organizational barriers; reservations with the usefulness of technology in the medical community; remaining technological limitations; and societal issues concerning consumer demands and equitable access. Table 36.4 provides an overview of these issues.
It should be noted that all these issues must be addressed if e-Health is to reach its potential in aiding health care in those application domains outlined in the previous section. We now provide a brief discussion of these different classes of issues affecting the development and implementation of e-health technologies in health care.
Regulatory Issues Although several other domains have been able to effectively implement Web-based technologies, health care has been largely unsuccessful. However, health care may very well be the most regulated industry (Fried, Weinreich, Cavalier, & Lester, 2000), which has likely contributed to the health care industry's constant lag in adopting technology. Conversely, it has also been suggested that federal and state regulations, and not the recognized needs and opportunities for growth, have driven the more recent surge of interest and adoption of new technologies by health care providers (Sainfort et al., 2003). The effects of the regulation and policies governing health care delivery and business practices is far too complicated to sufficiently cover in this chapter. However, we will provide an overview of some these issues that impact the ability of the health care industry to use Web-based technologies. The effects of the health care regulatory environment go beyond the much discussed HIPAA put forth by the DHHS in 1996. Regulatory directives have put the health care industry in a "catch-22." Pieces of regulations, such as HIPAA, require that health care organizations develop and implement new information systems in order to become compliant with definitive rules for the privacy, confidentiality, standards, and security
TABLE 36.4. Overview of Issues That Affect Health Care and the Progress of e-Health Category
Examples
Regulatory
• Health Insurance Portability and Accountability Act (HIPAA) • Antikickback, self-referral, and beneficiary inducement law • Difference in state-level regulations
Economic
• Previous investment in legacy systems • Large investments for development and implementation • Financial disincentives for care providers
Cultural
• Long-standing history of empty IT promises • New tools are not more efficient or easier to use
Organizational
• Uncertainty associated with changing roles/relationships • Difficulty determining resulting costs/organizational needs
Societal
• Lack of ethical/legal assurance of online information • Limited access to technology (the "digital divide") • Contrary demands of increased access and increased security
Technological
• Limitations with bandwidth, latency, and network stability • Insufficient data security, privacy, access controls • Lack of standardization of medical data coding schemes
36. Emergence of e-Health of health care data and the transactions performed with this data. There are severe monetary and punitive (including years in prison) consequences for noncompliance, negligence, or misuse of patient information (Phoenix Health Systems, 2003). This has actually encouraged (or forced) health care organizations to develop new information systems, including utilization of Webbased technologies. However, providers are also under the jurisdiction of state-level control. States have some control over the coding standards and implementation specifications used for health care information transactions, privacy of patient information, and regulation of health insurance plans (U.S. Department of Health & Human Services, 2003b). Differences between statespecific policies, such as professional licensure laws, limit the ability of health care organizations to deliver health care services and transfer information across state lines. In addition, on top of this complexity, the policies of the third-party payers and insurance companies may not allow for payment or reimbursement for Web-based services. This leaves little incentive for care providers and health care organizations to invest in these technologies if they will not be allowed to full use them and/or not get paid for the services rendered through these systems. Moreover, there is existing legislation originally proposed to stave off many of the potential harms inherent in a competitive health care market, which also inhibits the development and success of e-Health. For example, the antikickback, self-referral (a.k.a. the Stark law), and beneficiary inducement laws hinder the ability of e-Health in areas such as supporting the integration of health care organizations within an IDN and fostering the relationships between technology developers and the health care industry (Fried et al., 2000; Kleinke, 2000). These vague, outdated laws are problematic for the open access, flexible, integrated nature of online health care information systems. Clearly, these regulatory issues affect the ability of health care organizations to take advantage of the power of Web-based technologies.
Economic Issues There are economic issues affecting the use and acceptance of e-Health technologies. A number of groups have conducted studies and analyses that argue both for (e.g., Centers for Disease Control and Prevention, 1999; Littell & Strongin, 1996) and against (e.g., Cassell, 1993; Drake, Jaffe, & Fitzgerald, 1993; Newhouse, 1992; Nitzkin, 1996) the financial impact of implementing and using e-Health technologies. Economic obstacles to e-Health include massive implementation, maintenance, and training costs associated with revamping legacy information systems; complex payment mechanisms and economic relationships; and the lack of payment/reimbursement rules for the delivery of services now available through these new technologies. As previously alluded to when discussing regulatory issues, health care organizations have a need and/or desire to implement new information systems. However, most e-Health technologies are not "ready to use" and require a substantial, initial investment in terms of development, implementation, and training. In addition, these organizations will need to establish the in-house facilities for proper service, maintenance, and further
667
development of these systems. This investment, especially in light of the disincentives created by regulation and policy, is objectionable for health care organizations that want to leverage their previous investments in technology. Also, there is an inherent paradox with using new technologies as health care organizations implement them in an effort to streamline costs and improve resource utilization because these same technologies create the possibility of new products and services, which tend to increase expenditures on supporting new services, training personnel, and maintenance of systems (Starr, 2000). The general payment structure in health care creates a host of issues that need to be addressed. For example, physician use of the Internet, e-mail, and similar applications has suffered because there are no provisions for these new forms of care in either the government programs (i.e., Medicare and Medicaid) or the private sector (Kleinke, 2000; Landa, 2002). In addition, because care providers are typically rewarded for treatment rather than prevention and based on actual office visits, it leaves little incentive to invest in technologies that promote overall illness prevention and facilitate remote consultation (National Research Council, 2000). As with some of the legislation discussed, service payment and reimbursement models were not created with these new care paradigms in mind. All these issues create economic barriers to the use and adoption of e-Health technologies.
Cultural Issues Somewhat surprisingly, there are also cultural issues in the medical industry that have limited the adoption and use of e-Health technologies. In the medical community, there has been a longstanding lack of faith that technology will provide viable solutions to solve problems within the health care industry. This is true for two reasons: (a) there has been a history of highly touted IT solutions that have been oversold and underdeveloped; and (b) the new IT tools that have been developed often have not meet the needs of practicing clinicians and hospital administrators. As a result, health care providers and professionals have often reverted back to paper records and legacy systems that have served as trusted solutions (Kleinke, 2000). The health care industry commonly views new technology as too expensive, not supportive of current work practices, unusable (or not "user friendly"), difficult to implement, and maintenance intensive (Sainfort et al., 2003; Schoen, Davis, Osborn, & Blendon, 2000). Although physician interest and acceptance of technology is changing, particularly with respect to physician adoption and use of the Web and handheld and wireless technologies, there is still a widely held view that an information system that is not 100% reliable is 0% useful (Kleinke, 2000). If physicians require too much time and effort or have to endure too much frustration when trying to use new technologies to do the same things that they have been doing relatively efficiently and effectively for years, they tend to discard these technologies as useless or overly complicated. To increase the use and adoption rates of these new technologies, these tools need to be designed to support the natural workflow of the user.
668
MOLONEY ET AL.
Organizational Issues There are also a number of organizational issues that have prevented use of e-Health technologies from becoming more widespread. The major organizational issue affecting e-Health is how these new technologies will alter the business relationships and the power structure between consumers, practitioners, and payers. The National Research Council (2000) discussed several sources of uncertainty that health care organizations might face -when considering how to implement a Web-based business strategy, including organizational and industry structures for Web-based care and service, the technological capabilities of Web-based systems, the internal policies and procedures that will guide use of these technologies, and the human resources needs dictated by these new systems. All these sources of uncertainty affect if and how health care organizations will use e-Health technologies. The complex structure of relationships in the health care industry has created problems for the implementation of e-Health technologies. The use of Web-based technologies will likely cause highly unpredictable changes in the business relationships between and practices of payers, insurance companies, health maintenance organizations, and health care organizations. For example, the increased connectivity between care providers (e.g., physicians) and benefits managers may influence the treatment protocols chosen. In addition, the empowerment of consumers by providing them with access to more information through the Internet and Web sites may tip the balance of power by granting consumers the power to determine medications and procedures covered by the managed care organization or the power to directly access specialists without first going through a generalist. The reality is that health care organizations and insurers/payers cannot forecast all the changes that may result from Web-based systems, which creates something highly undesirable for businesses: uncertainty.
Societal Issues There have also been a number of societal issues that have affected the development of e-Health technologies. These issues include the lack of established ethical standards to govern online health-related information, uneven access to information technologies and the Web, and assurance of security and confidentiality of patient information. All these societal issues will have an enormous impact on the extent to which e-Health technologies will be used (especially in the public domain) and the future of governmental regulations governing these development, use, and implementation of these technologies. There are growing concerns about the lack of governing bodies controlling the quality of health-related content on the Web. The health care information available to consumers can affect their decisions and behaviors, creating a real need for ensuring the accuracy and quality of this information. Several e-health ethics initiatives have already begun to explore and propose standards for control of information quality, accuracy, and privacy. For example, eHealth Ethics Initiative (2000) established
the "eHealth Code of Ethics," including general guidelines, such as candor, honesty, quality, informed consent, privacy, professionalism in online health care, responsible partnering, and accountability. Although these guidelines for online health information have received some support, enforcement is likely to be achieved only through governmental regulation (as HIPAA legislation was needed to enforce the handling of patient information). Limited access to the Web and other technologies for many health care consumers, especially the numerically large Medicaid and Medicare recipient population, is another societal concern. The availability of access to Web-based and computing technologies is a major societal barrier. The well-documented decrease in access to these technologies for various demographic groups, including certain minority groups and those with low income, has been termed the "digital divide." This has been a topic of general debate with respect to equitable access for education and related domains. The widespread use of e-Health technologies may well increase the impact of the digital divide by putting these individuals with limited access to technologies at risk of having limited access to health care information and actual medical care. This also raises issues of equitable access for other individuals, such as those with disabilities who may have limited access due to problems with being able to use these technologies. A large problem for e-Health technologies that has been created by contrary consumer demands for increased access to information, while still demanding the highest levels of privacy and data security. The capabilities of Web-based technologies with respect to privacy and security continue to increase, as data encryption, access controls, and authentication security functions continually improve and are tested in more commercial Web-based domains. However, the intrinsic personal and stigmatic nature of health care or medical information brings the seriousness of confidentiality, data integrity, and security to another level. Although, currently, these societal issues truly impact only health care consumers, they may eventually affect health care providers and insurance companies as governmental regulation spills into this domain of health care.
Technological Issues There are a host of technological issues that have raised serious questions about the capability of Web-based technologies to meet the constraints of regulatory mandates, the needs of health care providers, the demands of consumers, and the policies of payers and insurance companies. Many of these technological issues have been the source of other issues, such as the lack of sufficient data and transaction security causing societal issues or the lack of system stability and the organizational concern for limited downtime and data loss. Although we do not discuss many of these technological issues in depth [e.g., how extensible markup language (XML) works and why it is being proposed as a good alternative to various coding standards], we do discuss the overall impact of how these technological limitations continue to impact the fate of e-Health. The National Research Council (2000) suggested five aspects of Web-based
36. Emergence of e-Health technologies that they believe will impact the ability of these technologies to support application domains of health care (e.g., clinical care or public health). These factors include bandwidth, latency, security, availability, and ubiquity. Although this list addresses many of the technological concerns that will impact many of the issues described previously, we would also like to add coding standardization and implementation with legacy systems. We believe that issues of bandwidth, latency, and availability will become less and less important as technology continues to advance and become cheaper. Ubiquity, or access to Webbased technologies, will likely continue to be a problem until highly affordable technological solutions become available or governmental regulations require that the digital divide be narrowed. Security for Web-based transactions has already greatly improved, as previously discussed, largely through the efforts for transaction and information security and privacy in the e-commerce domain. Despite improvements in these issues, e-Health must still face issues of implementation of these new systems with legacy systems (also an economic issue) and the continual problem with data coding and structure (also a regulatory issue). The lack of data standardization (or the overabundance of standardization schemes) in health care is an enormous problem for e-Health developers. There is a long history of efforts to develop data and coding standards in health care (see Hammond & Cimino, 2001, for a review). In this case, coding standards refer to the syntactical structure of procedural, diagnosis, and billing events. HIPAA legislation sets some new coding standards for data interchange and transaction, largely based on the established EDI standards. Technologies, such as XML, have greatly improved the ability to both meet HIPAA compliance and support the integration of transactions with EDI (the HIPAA standard) and non-EDI systems (Yang & Chang, 2002). However, it is unclear if these data structuring techniques can truly integrate all the different coding standards in health care, which number in the dozens. Developing systems that are truly integrated, which can transfer data seamlessly between disparate entities and facilities, requires that these systems be able to handle all the various coding standards for health care data. This comes with heavy development, implementation, and maintenance costs, which is cause for health care organizations to hesitate making a complete transfer to a Web-based system. Related to this idea is the desire of many health care organizations to salvage their initial investments in information systems. As a result, e-Health developers will have to develop frontend systems and integration engines that will both consolidate the information from these disparate systems, organize it for use, keep the data structure consistent for transfer back to the original system, and translate the data for transactions with systems with inconsistent coding schemas. Despite these technological woes, there is hope. The travel industry has succeeded in integrated multiple databases, services, and payment systems fairly seamlessly (Parente, 2000), the big three players in the automobile manufacturing industry have established standards for interchange and ordering of parts (Lumpkin, 2000), and the e-commerce infrastructure is a
669
successful means of making online transactions using sensitive information. These examples of success in applying Web-based technologies toward issues representative of those to be faced in health care do hold some promise for the future of e-Health. However, the costs (economic and otherwise) of system downtime and data loss or piracy in the health care domain are unparalleled. The reality is that the severity of many of these technological issues will persistently recede as technology continues to advance and become cheaper. However, the extent to which these technological issues hinder the advancement of e-Health will be largely determined by the resolution of issues in the regulatory, economic, cultural, organizational, and societal domains.
Summary These issues are all barriers to rapid, wide acceptance and implementation of e-Health technologies. For e-Health technologies to help the U.S. health care system, these barriers will need to be broken down. Unfortunately, many of these issues will take time to resolve, requiring new legislation, new business and delivery of care models, new payment and reimbursement policies, and new research and development groups to guide standardization rules. Fortunately, research focusing on how consumers, health care professionals, and other users interact with these new systems can help with some of these roadblocks. Human factors and HCI research can help improve the viability of e-Health technologies by helping to ensure the applications and systems developed support the needs of users, effectively help users to complete their work, and provide insight into the faults of unsuccessful designs.
HUMAN FACTORS AND HUMAN-COMPUTER INTERACTION ISSUES As can be seen in the previous section, there are still many barriers in the path of e-Health. However, we believe that research in Human Factors and HCI can help improve the chances for e-Health to become more widely implemented by health care and health-related organizations. Realistically, Human Factors and HCI will not be able to help resolve (at least directly) all the issues previously discussed. For instance, the changes to the payment policies of insurance companies and third-party payers, as well as governmental regulations, will likely be required to allow for remote consultation and Web-based care to become viable care practices. However, Human Factors and HCI research can help feed these processes of change by illustrating how the tools can be used and how the use of these technologies can improve health care, in terms of increased patient satisfaction and decreased medical errors. In addition, HCI research should and can involve cost-effectiveness analysis in order to contribute directly to the improvement of organizational and investor interest in implementing these technologies. HCI research can also reveal changes in procedures, increased productivity, and/or decreased errors that can directly translate into
670
MOLONEY ET AL.
cost savings. The real value of e-Health technologies, as well as the ways to design these systems and applications, can be revealed through HUMAN FACTORS and HCI research efforts directed at this domain. As with all domains in which technologies are used, the human element is vital to the success of e-health systems. Developers must address Human Factors and HCI issues, such as user capabilities and needs, natural user workflow, and contextual design, when designing these new systems. However, as e-Health is still in the developmental stage, much of the focus is on establishing the e-Health technological infrastructure and battling many of the regulatory and technological issues that prevent the widespread use of these technologies. As a result, there has been limited research focused on HUMAN FACTORS and HCI aspects of these technologies. Jacko, Sears, and Sorensen (2001), reported an early investigation into the perceptions of health care professionals and students regarding the effectiveness of the Web as an information retrieval tool. The study examined one set of potential e-Health users (practitioners) and one type of behavior (information retrieval). Clearly, more research is needed to build on this basis so broader segments of the system are studied, including other stakeholders and additional interaction behaviors. Fortunately, researchers can draw on much of the research that has already been conducted in other work domains, such as computer science, industrial engineering, psychology, education, and cognitive science, to name a few. When discussing the importance of applying HUMAN FACTORS and HCI principles to e-Health, the following research areas are particularly relevant: (a) human information processing and cognitive modeling, (b) models of work and system use, (c) universal accessibility, (d) computer-mediated collaboration, and (e) testing and evaluation. The next section focuses on these topics, while emphasizing their implications on the development and proliferation of e-Health.
Human Information Processing and Cognitive Modeling Effective HCI occurs when this interaction coincides with inherent human processing capabilities (Proctor & Vu, 2003). The human information processing approach to understanding HCI has traditionally focused on developing models of human information processing, based on empirical data collected from users performing tasks, which can be used to characterize and predict human behavior or performance. There has been a long history of efforts in HUMAN FACTORS and HCI to develop representative models and cognitive architectures that characterize human cognitive activities and behavior during interaction with computer systems (see Proctor & Vu, 2003; Yoshikawa, 2003; and Byrne, 2003, for reviews). Some of these models include the Model Human Processor (Card, Moran, & Newell, 1983); the Executive-Process Interactive Control (EPIC) architecture (Meyer & Kieras, 1997); the Adaptive Control of Thought (ACT) model (Anderson, Matessa, & Lebiere, 1997); and the GOMS family of models (John & Kieras, 1996).
Historically, a variety of problems with the design and implementation of information systems and technologies have stemmed from a failure to properly consider the cognitive needs of users (Tang & Patel, 1994). These problems include decreased productivity, increased user frustration, increased error rates, increased user stress, and poor decision making, among a host of problems. In addition, designers must recognize that human information processing behaviors and capabilities fluctuate with internal and environmental changes that affect the cognitive or sensory behavior of the user. Although there is a large base of HCI research examining cognition and information processing, further research is needed in the health care domain to understand how these theories and models apply to use of e-Health technologies.
Modeling Work and System Use The inherent complexity and required flexibility of health care information has contributed to the failure of past IT solutions and partially accounts for why physicians have often reverted back to paper-based charts (Kleinke, 2000). Part of the reason that many e-Health ventures have failed (especially EMRs) is the fact that these systems have not properly supported the work of health care professionals. The concept of context and workflow support is the idea that systems and technologies should be designed with consideration for the actual parameters of a user's work in the context that this work actually occurs. Unfortunately for e-Health systems designers, they need a solid understanding of the actual working conditions of health care professionals and consumers in addition to an understanding of their information processing and cognitive capabilities. Fortunately, however, there are already a number of methods for understanding actual user work. HCI research has developed a number of methods involving user-centered design or participatory design (see Muller, 2003; Norman & Draper, 1986, for reviews), which focus on getting representative users involved in the design and development process. In addition, methods of contextual design or ethnography (see Holtzblatt, 2003; Beyer & Holtzblatt, 1998; Blomberg, Burrell, & Guest, 2003, for reviews) focus on observing how users behave in their actual work environment. These overall "good practices" during the design life cycle of systems development help gather valuable insights into user characteristics and behaviors in the actual work environment to help guide system requirements needed for design (Mayhew, chap. 19, this volume). HCI research has produced several modeling techniques and methodologies that model the constraints of the user, work, and environment, and that have been shown to have the potential to guide system design, lessen user workload, minimize errors, and improve performance in various domains such as process control, aviation, and manufacturing. These techniques include the hierarchical task analysis (Shepherd, 1989) and the abstraction-decomposition space (Rasmussen, 1985), among others. These various methods for investigating the work practices of e-Health systems users in a real work environment will help designers and developers to better understand the needs of
36. Emergence of e-Health these users, the constraints of the -work environment, and the work practices that define the actual use of these technologies. Examining the wide variety of users, usage scenarios, workflow practices, and contexts of use will help e-Health technologies designers and developers to produce applications and systems that can be successfully implemented, achieve high user adoption rates, and benefit the health care industry.
Universal Accessibility Given the wide array of users and their respective needs, knowledge, and abilities, an effective e-Health system must adhere to the principles of universal design and access. e-Health technologies, especially those used by consumers, need to be designed to support varying user functional capacities. The concept of universal accessibility goes beyond the issues of physical accessibility or opportunity to use technologies (as with the digital divide). Universal accessibility refers to the design of information technologies that can be used by all individuals, including individuals with sensory, motor, and cognitive impairments, older and aging individuals, and individuals from different cultures (see Stephanidis, 2001, for a review; see also Stephanidis & Akoumianakis, chap. 13, this volume, and Choong, Plocher, & Rau, chap. 16, this volume). This issue has been the focus of countless studies in HCI. An in-depth discussion of the relevance of universal accessibility and "design for all" in health care is outside the scope of this chapter. However, suffice it to say that health care has a seemingly infinite variety of users, with unique needs and abilities. To ensure these individuals can use e-Health technologies and take advantage of all the benefits (e.g., increased care, social support, education) that these technologies can provide, designers and developers must ensure these systems are accessible to everyone. The importance of this issue to the prescribed problems of health care, one can consider the need to make e-Health tools accessible to individuals with chronic illness and impairment. As previously discussed, the majority of health care expenditures in the United States can be attributed to individuals with chronic illness. Some of these individuals suffer from various cognitive or sensory disorders, which impact their abilities to interact with technology. For example, people with visual impairments or blindness (one class of chronic illness) number in the millions. If e-Health tools are not accessible to, and/or usable by, these individuals, the potential economic and quality of care benefits brought about by the use of these technologies is effectively diminished. On a related note, it has been proposed that many situationally induced impairments result in similar performance decrements as disability-induced impairments (Sears, Lin, Jacko, & Xiao, in press). This further supports recognizing the unique needs of different classes of users and use scenarios, as even users with normal abilities are sometimes compromised. For example, a normally abled physician working in a highly distracting environment, such as an emergency room, may not be able to effectively use an e-Health tool as would have been possible in a less distracting environment. Research in HCI will help address many of these issues, ensuring all users have consistent, equal, and reliable access to vital health care information.
671
Computer-Mediated Collaboration Study of collaborative work has been conducted for a few decades, as the advent of networking for information technologies has enabled new methods of interactive working (Grudin, 1991). Computer-mediated collaboration, also known as computer-supported cooperative work (CSCW), is the study of the development and use of software that allows for geographic and/or temporal flexibility in the collaboration of individuals to accomplish work (Olson & Olson, 2003; van Tilburg & Briggs, chap. 30, this volume). The health care domain is an extremely complicated amalgamation of participants, agents, and stakeholders, each with their own goals, knowledge, strategies, and needs. This makes the much-needed collaboration between players in health care, including professionals, consumers, and organizations, extremely important, yet very difficult. Although some research has found that the use of Web-based technologies (e.g., e-mail) in health care can enhance interactions and collaboration, particularly through patient involvement in the health care process (MacDonald et al., 2001; Safran et al., 1998), other research has noted that much more work needs to be done in the understanding of how technology can be used to support this collaboration (Patel et al., 1999; Patel & Kushniruk, 1998). The support of this distanced (either by time or space) collaboration, in terms of business transactions, care delivery from clinician to patient, or problem solving by a group of health care professionals, is a fundamental aspect of e-Health technologies. The onset of the managed care model of health care, as well as the push toward preventative medicine, has made the communication among health care entities and between practitioners and consumers a crucial issue. Currently, patients are typically required to see a primary care physician, who will then refer this patient to a specialist as need be. It should also be noted that during this process, insurance companies and third-party payers are at work behind the scenes to handle payment, transaction, and benefits issues. This complex system of information sharing for physician referrals is just one example of the need for researchers to examine how e-Health technologies will change the social dynamics and communication of work in health care.
Testing and Evaluation Finally, the penchant for testing and evaluation is another benefit that HCI research can offer to the development of e-Health systems. As long as health care information systems are being built, there will always be a need (and a desire) to evaluate and improve user interaction with these systems. e-Health technologies have suffered from lack of user acceptance because these new tools have largely not supported users or provided users with usable tools. Research in HUMAN FACTORS and HCI, directed at e-Health technologies, has the potential to remedy many of these issues. Testing and evaluation are particularly important for e-Health applications because of both the potential consequences of poor design and the history of undertesting these technologies before implementation. The so-called "discount usability methods" such as heuristic evaluation (Nielsen, 1992), cognitive
672
MOLONEY ET AL.
walkthrough (Wharton, Bradford, Jeffries, & Franzke, 1992), and heuristic walkthrough (Sears, 1997) have also been shown to have value in diagnosing potential usability problems with systems and design prototypes (see Cockton, Lavery, & Woolrych, 2003, and Zhu, Vu, & Proctor, chap. 18, this volume, for reviews). These techniques can be used by e-Health developers, in tandem with representative users and subject-matter experts, to locate issues with the design or functionality of the systems or application that may decrease its usability. These formative evaluations can help to improve e-Health products and reverse the longstanding view by health care professionals that these new technologies are unreliable or not useful. Human error identification is a huge realm where testing and evaluation during the design life cycle (e.g., prior to implementation) can help decrease the well-documented problem of medical errors. Techniques such as Systematic Human Error Reduction and Prediction Approach (SHERPA; Embrey 1986) and Task Analysis for Error Identification (TAFEI; Baber & Stanton, 1994) are two methods for analyzing work activity and humanmachine interaction in order to analyze work tools and processes that lead to errors. These and other models of human error identification have been shown to have predictive power in assessing the likelihood for errors during interaction (Baber & Stanton, 1996; Hollnagel, Kaarstad, & Lee, 1999). These techniques can be used to help decrease the number of medical errors that occur during interaction with new or poorly designed technologies. Although a detailed view of the methods and metrics that are most appropriate for the evaluation of e-health technologies is outside the current scope, Sainfort et al. (2003) provide a more extensive discussion. As can be seen, the testing and evaluation methods used in HUMAN FACTORS and HCI research will help reveal problems with the design of these technologies. This information can ultimately be used to improve the viability of e-Health technologies by ensuring these new tools actually do what they have been designed to do—decrease medical errors, improve work proficiency, support user needs, improve care, and decrease costs.
Summary Work in Human Factors and HCI is important to ensure e-Health systems and applications are usable, accessible, efficient, and supportive as these technologies advance. As with many other domains, such as with automation in aviation (Parasuraman & Riley 1997; Wiener & Curry, 1980), the technical capabilities and economic benefits of technology are often the determinants of employing new technologies, whereas the effects on the human users are often ignored. HUMAN FACTORS and HCI research of e-Health technologies will bring the focus back to the users of these new systems and how these technologies can be designed and implemented in order to best support the needs of these users. We have suggested that the focus of this research should be on supporting human information processing needs, modeling and supporting the natural workflow practices, ensuring universal accessibility, examining the inherent collaborations in the health care process, and developing solid testing and
evaluation method. This work will be important in increasing the acceptance rates of technology and ensuring users leverage the power and efficiency of these tools. In early stages of design, and especially as the e-Health infrastructure develops further, attention to HUMAN FACTORS and HCI issues will become paramount in keeping e-Health from simply becoming another health care IT pipedream.
PREDICTIONS FOR e-HEALTH TECHNOLOGY IN HEALTH CARE e-Health technologies are becoming more widely used despite the many barriers to the development and implementation of these systems. For example, many health care providers are already allowing some consumer control and management of health insurance claims over the Web (Faulkner & Gray, 2000). We predict that the use of Web-based health care information systems and applications will continue to increase. Increasingly more health care organizations, health-related commercial enterprises, individual practitioner, and insurance companies will begin to implement these technologies because they will no longer be able to ignore their potential benefits and the inadequacies of their current information systems. These health care entities will recognize the potential for these technologies to improve quality of care, decrease medical errors, provide administrative and process streamlining, simplify and speed transactions, and provide health care professionals with the technological support they need to function in spite of an increasingly complex, information- and process-laden field. The increasing development and ubiquity of Web-based technologies, combined with the ever-increasing numbers of individuals seeking health information and services online, will feed this movement toward a Web-based health care technological infrastructure. e-Health developers and investors will naturally fall in line, feeding on the potential for a financial windfall. We predict that the following applications of Web-based technologies will have an immediate impact in the next several years: online consumer information services, online support groups and virtual communities, remote consultation and communication (primarily via e-mail), the online transaction infrastructure between health care provider groups, third party payers, and insurance/benefits companies, EMRs and other Web-based charts, and Web-based remote monitoring via sensor technologies. Despite the excitement about cutting-edge technologies, such as biomedical nanotechnologies, we believe that widespread application of these technologies will not impact medicine in the immediate future. Right now, the health care industry has to handle "first things first." This is likely so because the U.S. health care system has been on the verge of doing a major overhaul of information systems for nearly a decade. Web-based technologies, although still limited by various barriers, are well positioned to make an enormous impact on several of the problems outlined by the Institute of Medicine (2001) and other researchers. The interconnectivity generated by a Web-based infrastructure will help the health care industry support the dispersed
36. Emergence of e-Health nature of the IDN organizational model, support the collaboration required by the managed care model, and enable practitioners, providers, and public health organizations to provide preventative care and educational materials. The ability for consumers to easily seek continuous, remote care from providers, learn more about their condition or illness, and find social support from virtual communities on the Web will help reinforce the preventative care and health maintenance necessary to better manage the costly health care of the chronically ill masses. The open, flexible nature of the Web will allow the ever-increasing body of health care and medical knowledge and research to be easily manageable and accessible by health care professionals, providers, and consumers. This will help increase consumer education, improve professional education and training, and keep providers and payers informed of the latest research and clinical trials results. The open, public nature of most Web-based technologies will also begin to quell the increasing consumer demand discussed by Sainfort et al. (2003). Web-based technologies can finally provide consumer with the increased convenience to health care information and services. Finally, as Web-based technologies become more ubiquitous, less cost prohibitive, and more secure, the constraints on using these technologies in health care will start to dissolve.
CONCLUSION This chapter provides an overview of e-Health, from the historical factors that transformed the nature of health care in the United States to the applications of e-Health that have been envisioned for widespread use in the near future. We discuss the factors that have led to changes in the technological needs of the U.S. health care industry. We then provide an overview of e-Health, including why the Web has been the proposed platform, and several application domains that have been suggested as likely targets for implementation of Web-based technologies. Then an overview was provided of the various barriers, including regulatory, economic, organizational, cultural, societal, and technological issues, which currently stand in the path of e-Health developers and proponents. Finally, we discuss the role the research and theory in HUMAN FACTORS and HCI will play in improving the viability of e-Health technologies.
673
Despite the controversy surrounding the ability of these Webbased systems and applications to break through the barriers and provide viable solutions to the industry's problems, e-Health is well positioned to become the new technological infrastructure of health care in the United States. There are several hospitals, physician offices, e-commerce health care ventures, and insurance companies who are testing Web-based information systems and applications. As encouraging evidence of e-Health's benefits, such as cost savings, increased patient satisfaction, and fewer medical errors, continues to permeate the health care research, business, and marketing literature, these technologies will begin to receive more interest from investors and lead decision makers in the industry. One thing is clear: A need exists for new technologies to deal with the continually increasing demands of consumers for new and better services and information, the increasingly stringent regulations by the government about the accessibility, security, format, and storage of health care and medical data, and the continuing economic drive to decrease health care costs and support managed care, public health, and the chronically ill. The mere availability of new technologies will not ensure e-Health's success. Rather, the ability of e-Health technologies to deliver on their promises will be a result of attention from researchers in the fields of engineering, computer science, cognitive science, psychology, medicine, and biological sciences to ensure these systems are designed to support the natural work practices of users, ensure universal accessibility for all user groups, support the cognitive and information processing needs of users, enhance the collaboration between users, and ensure all these benefits through proper testing and evaluation during the engineering design lifecycle. This will be a job for you, the readers— the researchers and practitioners in HUMAN FACTORS and HCI.
ACKNOWLEDGMENTS The preparation of this chapter was largely supported by the William W. George Professorship of Health Systems held by the second author, and by the National Science Foundation through grants awarded to the third author (BES-0196030 and 0106248). The authors also gratefully express their appreciation to Paula J. Edwards who provided important feedback and input on iterative drafts of this chapter.
References Anderson, J. R., Matessa, M., & Lebiere, C. (1997). ACT-R: A theory of higher level cognition and its relation to visual attention. HumanComputer Interaction, 12,439-462. Baber, C., & Stanton, N. A. (1994). Task analysis for error identification: A methodology for designing error tolerant consumer products. Ergonomics, 37(11), 1923-1941. Baber, C., & Stanton, N. A. (1996). Human error identification techniques applied to public technology: Predictions compared with observed use. Applied Ergonomics, 27(2), 119-131. Bashshur, R., & Lovett, J. (1977). Assessment of telemedicine: Results
of the initial experience. Aviation Space and Environmental Medicine, 48(1), 65-70. Benton Foundation. (1999). Networking for better care: Health care in the information age. [On-line]. Available: http://www.benton.org/ Library/health/ Beyer, H., & Holtzblatt, K. (1998). Contextual design: Defining customer centered systems. San Francisco: Kaufmann. Blomberg, J., Burrell, M., & Guest, G. (2003). An ethnographic approach to design. In J. A. Jacko & A. Sears (Eds.), Human-computer interaction handbook: fundamentals, evolving technologies and
674
MOLONEY ET AL.
emerging applications (pp. 965-986). Mahwah, NJ: Lawrence Erlbaum Associates. Byrne, M. D. (2003). Cognitive architecture. In J. A. Jacko & A. Sears (Eds.), Human-computer interaction handbook: Fundamentals, evolving technologies and emerging applications (pp. 97-117). Mahwah, NJ: Lawrence Erlbaum Associates. Cain, M. M., Mittman, R., Sarasohn-Kahn, J., & Wayne, J. C. (2000). Health e-people: The online computer experience. Oakland: Institute for the Future, California Health Care Foundation. Card, S. (2003). Information visualization. In J. A. Jacko & A. Sears (Eds.), Human-computer interaction handbook: Fundamentals, evolving technologies and emerging applications (pp. 544-582). Mahwah, NJ: Lawrence Erlbaum Associates. Card, S. K., Moran, T. P., & Newell, A. (1983). The psychology of humancomputer interaction. Hillsdale, NJ: Lawrence Erlbaum Associates. Cassell, E. J. (1993). The sorcerer's broom: Medicine's rampant technology. Hastings Center Report, 23(6), 32-39. Centers for Disease Control and Prevention. (1999, June). New data show AIDS patients less likely to be hospitalized. [On-line]. Available: http://www.cdc.gov/od/oc/media/pressrel/r990608.htm Cockton, G., Lavery, D., & Woolrych, A. (2003). Inspection-based evaluations. In J. A. Jacko & A. Sears (Eds.), Human-computer interaction handbook: Fundamental, evolving technologies and emerging applications (pp. 1119-1138). Mahwah, NJ: Lawrence Erlbaum Associates. Delia Mea, V (2002). What is e-health (2): The death of telemedicine? [editorial]. Journal of Medical Internet Research, 3(2):e22. Available: http://www.jmir.Org/2001/2/e22/ Drake, D., Jaffe, M., & Fitzgerald, S. (1993). Hard choices: Health care at what cost? Kansas City, MO: Andrews and McMeel. eHealth Ethics Initiative. (2000). eHealth code of ethics [On-line]. Internet Healthcare Coalition. Available: http://www.ihealthcoalition. org/ethics/code0524.pdf Embrey, D. E. (1986). SHERPA: A systematic human error reduction and prediction approach. Paper presented at the International Meeting on Advances in Nuclear Power Systems. Knoxville, TN. Eysenbach G. (2001). What is e-health? [editorial]. Journal of Medical Internet Research, 3(2):e20. Available: http://www.jmir.org/2001/ 2/e20/ Faulkner & Gray. (2000). Faulkner & Gray's health data directory, 2000 Edition. New York: Faulkner & Gray, Inc. Ferguson, T. (2000). From doc-providers to coach-consultants: Type 1 vs. type 2 provider-patient relationships. The Ferguson Report [serial online]. Available: http://www.fergusonreport.com/articles/ tfr07-01.htm Fox, S., & Rainie, L. (2002). Vital decisions: How Internet users decide what information to trust when they or their loved ones are sick [On-line]. Pew Internet and American Life Project. Available: http:// www. pewinternet.org/reports/pdfs/PIP_Vital_Decisions_May2002. pdf Fried, B. M., Weinreich, G., Cavalier, G. M., & Lester, K. J. (2000). E-health: Technological revolution meets regulatory constraint. Health Affairs, 19(6), 124-131. Giorgianni, S. J., Grana, J., & Sewell, S. (2000). e-Health care: The Internet information revolution. The Pfizer Journal, 4(2). Grudin, J. (1991). CSCW: The development of two development contexts. Proceedings of the ACM Conference on Human Factors in Computing Systems (CHI'91), 91-97. Hammond, W E., & Cimino, J. J. (2001). Standards in medical informatics. In E. H. Shortliffe & L. E. Perreault (Eds.), Medical informatics: Computer applications in health care and biomedicine (pp. 212256). New York: Springer-Verlag. Harris, J. (1995). Consumer health information demand and delivery: A
preliminary assessment. Partnerships for networked health information for the public. Washington, DC: Office of Disease Prevention and Health Promotion, U.S. Department of Health and Human Services. Hollnagel, E., Kaarstad, M., & Lee, H-C. (1998). Error mode prediction. Ergonomics, 42, 1457-1471. Holtzblatt, K. (2003). Contextual design. In J. A. Jacko & A. Sears (Eds.), Human-computer interaction handbook: Fundamentals, evolving technologies and emerging applications (pp. 941-963). Mahwah, NJ: Lawrence Erlbaum Associates. Institute of Medicine. (2001). Crossing the quality chasm: A new health system for the 21st century. Washington, DC: National Academy Press. Jacko, J. A., Sears, A., & Sorenson, S. J. (2001). Framework for usability: healthcare professionals and the Internet. Ergonomics, 44(\\), 989-1007. Jadad, A. R. (1999). Promoting partnerships: Challenges for the Internet age. BMJ, 319,761-764. John, B. E., & Kieras, D. E. (1996). The GOMS family of user interface analysis techniques: Comparison and contrast. ACM Transactions on Computer-Human Interaction, 3(4), 320-351. Kassirer, J. P. (2000). Patients, physicians, and the Internet. Health Affairs, 19(6), 115-123. Kleinke, J. D. (2000). Vaporware.com: The failed promise of the health care Internet. Health Affairs, 19(6), 57-71. Landa, A. S. (2002). Telemedicine payment expansion sought. [Online]. Amednews.com. Available: http:// www.ama-assn.org/scopubs/amnews/[ick_02/gvsc0708.htm LeGrow, G., & Metzger, J. (2001). e-Disease management. Oakland: California Healthcare Foundation. Littell, C. L., & Strongin, R. J. (1996). The truth about technology and health care costs. IEEE Technology and Society Magazine, 75(3), 10-14. Lumpkin, J. R. (2000). E-health, HIPAA, and beyond. Health Affairs, 19(6), 149-151. MacDonald, K., Case, J., & Metzger, J. (2001). E-Encounters. Oakland: California Healthcare Foundation. McCormack, J. (2000). Group practices find their way to the Internet. Health Data Management, 8(1), 46-53. Metzger, J., & MacDonald, K. (2002). Clinical decision support for the independent practice physician. Oakland: California Healthcare Foundation. Meyer, D. E., & Kieras, D. E. (1997). A computational theory of executive cognitive processes and multiple-task performance: Part 2. Accounts of psychological refractory-period phenomena. Psychological Review, 104,749-791. Meyer, J. (2001). Age: 2000 [On-line]. Census 2000 Brief, October 2001 (C2KBR/01-12). U.S. Census Bureau. Available: http://www.census. gov/prod/2001pubs/c2kbr01-12.pdf Meyer, M., Kobb, R., & Ryan, P (2002). Virtually healthy: Chronic disease management in the home. Disease Management, 5, 87-94. Mitchell J. (1999). From telehealth to e-health: The unstoppable rise of e-health. Canberra, Australia: National Office for the Information Technology. Available: http://www.noie.gov.au/projects/ ecommerce/ehealth/rise_of_ehealth/unstoppable_rise.htm Mittman, R., & Cain, M. (1999). The future of the Internet in health care: Five year forecast. Oakland: California Healthcare Foundation. Mollica, R. L., & Gillespie, J. (2003). Care coordination for people with chronic conditions. Portland, ME: National Academy for State Health Policy. Muller, M. J. (2003). Participatory design: The third space in HCI. InJ. A. Jacko & A. Sears (Eds.), Human-computer interaction handbook:
36. Emergence of e-Health
fundamentals, evolving technologies and emerging applications (pp. 1052-1068). Mahwah, NJ: Lawrence Erlbaum Associates. National Research Council. (2000). Networking health: Prescriptions for the Internet. Washington, DC: National Academies Press. Newhouse, J. P. (1992). Medical care costs: How much welfare loss? Journal of Economic Perspectives, 6(3), 3-21. Nielsen, J. (1992). Finding usability problems through heuristic evaluation. In R Bauersfield, J. Bennett, & G. Lynch (Eds.), Proceedings of the ACM Conference on Human Factors in Computing Systems (CHI'92) (pp. 373-380). New York: ACM Press. Nitzkin, J. L. (1996). Technology and health care—Driving costs up, not down. IEEE Technology and Society Magazine, 75(3), 4045. Norman, D. A., & Draper, S. (1986). User centered system design. Mahwah, NJ: Lawrence Erlbaum Associates. Olson, G. M., & Olson, J. S. (2003). Groupware and computer-supported cooperative work. In J. A. Jacko & A. Sears (Eds.), Human-computer interaction handbook: Fundamental, evolving technologies and emerging applications (pp. 583-595). Mahwah, NJ: Lawrence Erlbaum Associates. Parasuraman, R., & Riley, V (1997). Humans and automation: Use, misuse, disuse, abuse. Human Factors, 39(2), 230-253. Parente, S. T. (2000). Beyond the hype: A taxonomy of e-health business models. Health Affairs, 19(6), 89-102. Partnership for Solutions. (2002). Chronic conditions: Making the case for ongoing care. Baltimore, MD: Johns Hopkins University. Patel, V L., Kaufman, D. R., Allen, V G., Shortliffe, E. H., Cimino, J. J., & Greenes, R. A. (1999). Toward a framework for computer-mediated collaborative design in medical informatics. Methods of Information in Medicine, 38(3), 158-176. Patel, V. L., & Kushniruk, A. W (1998). Interface design for health care environments: The role of cognitive science. Proceedings of the 1998 AMIA Annual Symposium, 29-37. Phoenix Health Systems. (2003). HIPAA primer [On-line]. Available: http://www.hipaadvisory.com/regs/ HIPAAprimerl .htm Potter, M. A., & Longest, B. B. (1994). The divergence of federal and state policies on the charitable tax exemption of nonprofit hospitals. Journal of Health Politics, Policy, and Law, 19(2), 393419Proctor, R. W., & Vu, K-P. L. (2003). Human information processing: An overview for human-computer interaction. In J. A. Jacko & A. Sears (Eds.), Human-computer interaction handbook:Fundamentals, evolving technologies and emerging applications (pp. 35-51). Mahwah, NJ: Lawrence Erlbaum Associates. Rasmussen, J. (1985). The role of hierarchical knowledge representation in decision making and system management. IEEE Transactions on Systems, Man and Cybernetics, 15, 234-243. Raymond, B., & Dold, C. (2002). Clinical information systems: Achieving the vision. Oakland, CA: Kaiser Permanente Institute for Health Policy. Roa, R., Hoglund, D. H., Martucci, J., & Wilson, S. C. (2002). The future of wireless personal digital assistants (PDAs) in healthcare [Presentation]. Healthcare Information & Management Systems Society. Available: http://www.himss.org/content/nles/proceedings/ 2002/sessions/sesslides/sessl008.pdf Rosen, E. (1997). The history of desktop telemedicine. Telemedicine Today, 5(2), 16-17, 28. Safran, C., Jones, P. C., Rind, D., Bush, B., Cytryn, K. N, & Patel, V (1998). Electronic communication and collaboration in a health care practice. Artificial Intelligence in Medicine, 12(2), 139153Safran, C., & Perreault, L. E. (2001). Management of information in integrated delivery networks. In E. H. Shortliffe & L. E. Perreault (Eds.),
675
Medical informatics: Computer applications in health care and biomedicine (pp. 359-396). New York: Springer-Verlag. Sainfort, E, Jacko, J. A., & Booske, B. C. (2003). Human-computer interaction in health care. In J. A. Jacko & A. Sears (Eds.), Humancomputer interaction handbook: Fundamentals, evolving technologies and emerging applications (pp. 808-822). Mahwah, NJ: Lawrence Erlbaum Associates. Schoen, C., Davis, K., Osborn, R., & Blendon, R. (2000). Commonwealth fund 2000 international health policy survey of physicians' perspectives on quality. New York: Commonwealth Fund. Sears, A. (1997). Heuristic walkthroughs: Findings the problems without the noise. International Journal of Human-Computer Interaction, 9, 213-234. Sears, A., Lin, M., Jacko, J., & Xiao, Y. (2003). When computers fade... Pervasive computing and situationally-induced impairments and disabilities. Proceedings of the Tenth International Conference on Human-Computer Interaction (HCII 2003), 12981302. Shepherd, A. (1989). Analysis and training in information technology tasks. In D. Diaper (Ed.), Task analysis for human-computer interaction (pp. 15-55). Chichester, England: Ellis Horwood. Shi, L., & Singh, D. A. (2001). Delivering health care in America: A systems approach. Gaithersburg, MD: Aspen. Singer, S. J., Enthoven, A. C., & Garber, A. M. (2001). Health care and information technology: Growing up together. In E. H. Shortliffe & L. E. Perreault (Eds.), Medical informatics: Computer applications in health care and biomedicine (pp. 663-696). New York: SpringerVerlag. Starr, P. (2000). Health care reform and the new economy. Health Affairs, 19(6), 23-32. Stephanidis, C. (2001). User interfaces for all: Concepts, methods, and tools. Mahwah, NJ: Lawrence Erlbaum Associates. Strohecker, J. (2000). HealthWorld Online: Promoting healthy living via the Internet. HealthWorld News [serial online]. Available: http://www.healthy.net/media/1iwnews/faulknergray. htm Tang, P. C., & McDonald, C. J. (2001). Computer-based patient-record systems. In E. H. Shortliffe & L. E. Perreault (Eds.), Medical informatics: Computer applications in health care and biomedicine (pp. 327-358). New York: Springer-Verlag. Tang, P. C., & Patel, V L. (1994). Major issues in user interface design for health professional workstations: Summary and recommendations. International Journal of Bio-Medical Computing, 34, 139148. Taylor H. (2002). The Harris Poll #21, May 1, 2002 [On-line]. Cyberchondriacs Update. Harrislnteractive. Available: http://www. harrisinteractive.com/harris_poll/index.asp?PID=299 Turisco, E, & Case, J. (2001). Wireless and mobile computing. Oakland: California Healthcare Foundation. U.S. Department of Health and Human Services. (2003a). Health insurance reform: Modifications to electronic data transaction standards and code sets. Federal Register, 65(34), 8381-8399U.S. Department of Health & Human Services. (2003b). Health insurance reform: Standards for electronic transactions. Federal Register, 65(160), 50312-50372. U.S. National Library of Medicine. (2002). Pub Med: Medline® retrieval on the World Wide Web [Fact sheet]. Available: http://www.nlm.nih. gov/pubs/factsheets/pubmed.html Vaccaro, J., Cherry, J., Harper, A., & O'Connell, M. (2001). Utilization reduction cost savings, and return on investment for the PacifiCare Chronic Heart Failure Program, "Taking Charge of Your Heart Health." Disease Management, 4(3), 1-10.
676
MOLONEY ET AL.
Wharton, C, Bradford, J., Jeffries, R., & Franzke, M. (1992). Applying cognitive walkthroughs to more complex user interfaces: Experiences, issues, and recommendations. In P. Bauersneld, J. Bennett, & G. Lynch (Eds.), Proceedings of the ACM Conference on Human Factors in Computing Systems (CHF92) (pp. 381-388). New York: ACM Press. Wiener, E. L., & Curry, R. E. (1980). Flight-deck automation: Promises and problems. Ergonomics, 23(10), 995-1011. World Health Organization. (2000). The world health report 2000: Health systems: Improving performance. [On-line]. Avail-
able: http://www.who.int/whr2001/2001/archives/2000/en/index. htm Yang, D., & Chang, C-H. (2002). Web services & security: A better approach to HIPAA compliance. [On-line]. Available: http://www.cysive.com/news/Web_Services_And-Security.pdf Yoshikawa, H. (2003). Modeling humans in human-computer interaction. In J. A. Jacko & A. Sears (Eds.), Human-computer interaction handbook: Fundamentals, evolving technologies and emerging applications (pp. 118-146). Mahwah, NJ: Lawrence Erlbaum Associates.
Part
IXIIE RETURN ON INVESTMENT AND STATE OF THE ART USABILITY GUIDELINES FOR THE WEB
677
This page intentionally left blank
37 DETERMINING THE VALUE OF HUMAN FACTORS IN WEB DESIGN Eugenie Bertus BMC Software, Inc.
Mark Eertus Auburn University
INTRODUCTION Bias and Mayhew (1994) edited a pivotal book on usability and quantifying return on investment (ROI). In the last chapter of their book, they predicted that in 10 years there would no longer be a need for such a book because usability would have become generally accepted and practitioners would know how to provide ROI information to their management as common practice. Ironically, Bias and Mayhew are currently working on a second edition of their book. Today's designers may be new and the interfaces glitzier, but the ROI concerns are just as relevant as they were 10 years ago. Everything old is new again. One does not have to be particularly insightful or be much of a visionary to know that when money and profitability are involved, it will always be safest to know how to justify your existence. It would be nice if the value provided by human factors professionals were so self-evident that people would be overwhelmingly impressed with the interfaces resulting from their input. People would note that the interfaces are so easy to use that their grandmother could use them like a professional programmer. Unfortunately, this is not reality, and it likely never will be. The progression of computing in the past several decades has been astounding (see Bernstein, chap. 2, this volume). In a relatively short period of time, the computing industry has gone from huge expensive machines used only by government agencies, universities, and large industries to smaller, relatively
679
inexpensive machines that are pervasive throughout every aspect of our lives. Computing has so deeply entrenched itself into our society that yesterday's technophobes have become today's mainstream users of technology. Software interfaces now aim to be usable to every conceivable buyer type in the market. The World Wide Web has enabled people from a wide variety of backgrounds to access information, communicate, and purchase merchandise—all from the comfort of their homes. Thus enters the human factors professional. Human factors professionals are called on to take businesses with an interest in reaching broader audiences and make them accessible to every segment of the market. The human factors professionals are called by many different titles: human factors professionals, usability engineers, interface designers, and usability testers. Regardless of their title, though, the goal is the same: to improve interface designs for human use. There are several other chapters in this handbook that discuss how to make the interfaces better and easier to use. This chapter concentrates on an issue that almost every human factors professional will have to address at one time or another in his or her career, that of justifying the investment in human factors. It emphasizes the benefits that human factors professionals can provide to a company and how much more profitable the company can become by using their services. One might question why there is a need to focus on the benefit human factors as a profession brings to the development process. The answer is simple. Corporations exist for one basic purpose: to make money. If human factors analyses are to be
680
BERTUS AND BERTUS
regarded favorably, then there must be demonstrable proof of how human factors contributes to the bottom line. When it comes to improving the user interface, everyone thinks they are experts. One of the greatest challenges that a human factors professional will face in a corporation is the idea that usability is just common sense and, therefore, that any developer should be capable of designing a highly usable interface. In some cases, the argument is that developers and technical architects are experts and as such know the needs of the users better than the users do themselves. In addition, when usability work is done correctly and well, people look at the outcome and think, "Of course, it was so obvious." When a user interface is done well, it is nearly transparent to the users. The best way to combat these beliefs is to provide ROI measures. Comparing a user interface's ROI metrics before usability intervention and then after usability redesign can be persuasive. Even comparing product interfaces that have had usability input to those that have not can be insightful. When these measures are provided, people will notice the benefits. The following quote from the Chief Technology Officer ofStaples.com after the company redesigned their Web site sums it up nicely: "The people who are developing the applications are the least capable of making them usable for customers" (Leon, 2001). This statement came after enhancements to their site, which included major usability work that resulted in increased sales of 491% (Roberts-Witt, 2001). Such a remarkable increase makes a very convincing case for human factors intervention. Gathering the data and calculating ROI information can be time consuming. There are some corporations, such as Intel, that employ staff full time for the sole purpose of providing ROI data. However, it is more likely that a human factors professional will be asked to provide this information in addition to his or her regular work. To determine how much data to collect, a quick estimate should be made for the proposed ROI. One could attempt to gather the ROI metrics at many different levels. Very accurate measures can be obtained by spending large quantities of time measuring multiple metrics, such as task completion time, user satisfaction, documentation preparation time, and training material preparation time, and comparing them with the revenue numbers and the development budget. More approximate measures can be obtained by simply measuring user satisfaction and comparing it with the revenue numbers. The human factors professional must assess the level of support that is appropriate for making the argument that fits the goal for which the data are being obtained. There are numerous reasons to quantify the human factors benefit to the company. One reason might be to justify the current positions and be able to influence when and how human factors is integrated into the development process. Alternatively, there may be a need to seek funding for more positions; in this case, it is necessary to demonstrate a proposed ROI for additional headcount. Or the justification might simply be to gain "social" acceptance within the company and other development areas. The amount of effort for collecting data will likely be higher when the goal is to justify hiring for a new position versus to demonstrate the value of human factors and gain acceptance. After establishing why there is a need to provide the ROI data, it is important to identify the target audience and their
most important goals. The cost-benefit analysis should draw a connection from the usability metrics (e.g., task completion time, error rates, user satisfaction numbers, etc.) to the benefit goals of the group receiving the analysis (Donahue, Weinscheck, & Nowicki, 1999). Keeping in mind the goals of the target audience and the business, one is now ready to begin thinking about gathering the data for the ROI calculations.
HISTORICAL STUDIES Where to begin? As stated previously, everything old is new again. To begin to develop a cost-benefit argument, draw on some historical work in the field of human factors. There have been many papers and essays written on cost-justifying human factors. The quintessential book, "Cost-Justifying Usability" (Bias & Mayhew, 1994), brought together practitioners who shared their experiences and statistics on cost justification of products on which they worked. Although they wrote the book prior to the Web boom, the lessons learned from their experiences are still relevant to the designer of a Web user interface. One chapter in the book lays out a compelling argument on how usability can impact the development of software, stating that the user interface can be up to 60% of the system code (Karat, 1994). In addition, the interface can require 40% of the development time (Wixon & Jones, 1992, as cited in Karat, 1994). Budget estimates show that software projects will frequently exceed their budget and time estimates. The most common reasons given for these overruns are related to usability (Lederer & Prassad, 1992). Poor communication and understanding of user needs, or overlooked user tasks are two examples. Given the level of commitment for the development of the user interface, it stands to reason that appropriate effort should be applied to design it well. A widely cited estimate is that for every dollar invested in usability at the definition phase of a project, it would cost $10 to implement the same feature during development and $100 to implement after release (Boehm, 1981; Pressman, 1992). In addition, Pressman claims that 80% of the cost of software development occurs postrelease, in the form of fixes to missed or incorrect requirements. Incorporating human factors into the development process early in the development cycle would help alleviate some of these issues, by uncovering user tasks and needs earlier in the development life cycle. Overall, using human factors methods can reduce the product development cycle by 33% to 50% (Bosert, 1991). Although most arguments for incorporating human factors into the software development life cycle center around introducing the practices early, studies have demonstrated that even methods that occur late in the cycle can add benefit. LaPlante (1992) found that on average a usability test pinpoints many issues and results in 70 to 100 usability recommendations. If only one half of these recommendations were addressed, this would improve the usability of the interface by more than 50% (Landauer, 1995). This information is useful for any development of user interfaces, including a Web interface. In a paper describing how
37. Human Factors in Web Design Web sites can achieve a competitive advantage, Rhodes (2000) claimed that the first tenet of survival is to have the better user experience. He stated that if all else is equal, then the company with the easier-to-use site will win the market. How much of a market is that? In 1998, it was estimated that more than 44 million people in the United States had already made purchases online, and 37 million more said they would be making purchases soon (Wildstrom, 1998). By 2002, the number of online purchases increased from those in 2001 by 37% to 358.6 million totaling $47.98 billion (Fitzgerald, 2003). According to Nielsen (2003), a mere 10% investment of the development budget into usability efforts can improve site ease of use by 135%. This translates into conversion rates (the number of people who enter the site and then make a purchase) improving 100% and user productivity increasing 161% on average. There are various studies that have demonstrated how the inclusion of usability into the development process can achieve these high levels of benefits. On an internal project, IBM spent $68,000 on usability improvements. In the first year alone, the system saved the company an estimated $6.8 million (Karat, 1990). A Forrester report (Souza, 2001) described how Skechers changed its Web site to move product selection closer to the homepage. By doing so, there was a 400% increase in holiday sales. Another article (Black, 2002) asserted that Dell invested in usability on their Web site, and online sales "went from $ 1 million per day to $34 million per day. When Staples.com invested in usability to redesign their Web site, there was a 31% decrease in drop-offs after viewing only the homepage. Users were more likely to view product descriptions, and sales increased 491% in year-over-year growth (Roberts-Witt, 2001).
681
Mayhew and Mantei (1994) provided an excellent description of the usability process, the usability tasks that can be completed during the development process, and how to calculate the associated costs. To illustrate the points, we use an example on which we can build. Assume that there is a Web design project for an online sales company that sells "Everything '70's." They specialize in selling era-specific memorabilia. The company has been in business for 3 years. They started on a small scale and built a quick user interface to sell their products. After 3 years, the company has grown from five sales per month to 500 sales per month. The redesign project is attempting to present a better organization for displaying the products and an easier-to-use checkout system. It will require 3 months of human factors support. This work will include the human factors professional's time and a usability study with five participants. The participants will be paid $ 50 per hour for their work and the test will run for 3 hours. The study participants are all local and do not require room and board expenses; however, the test facility must be rented at a rate of $500 per day. It is possible to calculate the total costs for the project. Human factors support 65 days at a rate of $200 an hour 5 participants for 3 hours at $50 per hour Test facility for 4 days at $500 per day Total
$104,000 $750 $2,000 $106,750
BENEFITS COSTS Although these statistics are quite compelling, eventually one would need to provide his or her own analyses on how human factors impact the company. ROI calculations consist of two pieces of information: costs and benefits. We start with the costs. One factor in the ROI equation is the cost of providing human factors assistance to a project. A fundamental measure to include is the loaded salaries of the human factors professionals that will be working on the project. The "loaded" estimate typically takes into account salary, benefits, office, supplies, and so on. For example, a corporation's loaded salary may work out to be $200 per hour. If a project requires two human factors personnel for 28 full days of work, then the costs for the headcount would be $11,200. $200 x 2 x 28 = $11, 200 In addition to the headcount hours, the calculated costs might include the processes that will be used to help design the project. If the project requires a usability test, then there will be a need to include the testing facility costs, paying a recruiter to find users, potentially paying users for their time, and possibly flying in users and paying for their room and board.
The next ROI piece to determine is the benefits. Donahue (2001) said there are four steps to measuring usability ROI benefits: Select the usability criteria for measurement, determine the appropriate unit of measure, make reasonable assumptions of the benefit magnitude, and translate that measurement into a monetary value. The biggest challenge is to match the usability metrics to the corporate goals. These goals are usually basic and include increasing revenue and decreasing costs. There may be other related subgoals, such as increasing brand awareness or increasing customer retention. Matching the usability metrics to the corporate goals may be challenging. In many cases, the typical metrics that can be gathered by usability are only indirectly related to the corporate goals. Assumptions and logical connections must be made to relate the measures to the corporate goals. One example of a typical usability measurement is the ease of use of the product. The argument made is that easier-to-use products will be preferred by users. If users prefer this product's interface, then it should sell better. However, there is no direct causeand-effect relationship between increased ease of use and increased sales. At best, one can hope for some strong correlational data. There are other measures that seem to provide a more direct link between usability enhancements and corporate goals. For example, if a company is interested in increased sales, then
682
BERTUS AND BERTUS
before and after the usability improvements, measure the conversion rate and the rate of completion once the user has moved to the shopping cart. If the results show an increase in conversion rates and a decline in drop-offs (people who abandon the site before completing a purchase), then a greater sales volume can be assumed. Both of these metrics could be related back to the goal of increased sales. This relationship requires some communicative logic (if A then B; if B then C; therefore, if A then C), but it is a fairly straight line of logic to follow. Table 37.1 shows some human factors measures and the link to increasing revenues. Likewise, if a corporation's goals include reducing costs, then measures such as the number of calls generated to customer support or the amount of time to develop and provide training can be gathered and calculated. One report (Souza, 2001) showed that when a company improved descriptions and images of its merchandise, the calls and e-mails to support were reduced by 20%. In another example, a redesign of the user interface decreased training costs by $2.5 million (Chapanis, 1991). Another way to reduce costs is to reduce development costs. Earlier in the chapter we discussed how much time and budget are spent developing the user interface. Human factors methods can help decrease development costs by reducing the amount of time needed for quality testing or reducing the number of redesigns required because errors were eliminated early in design. There are numerous ways to pair usability metrics to the underlying business goals. Table 37.2 lists some of the possible pairings of usability metrics to the goal of reducing costs. There are some considerations to keep in mind while determining the cost-benefit measures for the ROI calculations. To make the calculations convincing, conservative assumptions should be used. Take, for example, a claim that a product helps decrease costs by providing better efficiency. It can be shown that the product saves 3 seconds of processing for a particular window, with the window appearing on average 20 times per day. If there are 300 employees who will use this window and their time costs $100 per hour, then the savings is $500 a day or $117,500 a year (based on the assumption of 235 working days in a year).
(20 x 3) x 300 = 18,000 sec or 5 hours per day 5 x $100 = $500 per day $500 x 235 = $117,500 per year The argument could be made that not every second that is saved on processing will be used for productive work. To be more conservative, estimate that 75% of the saved time will translate into other productive work and the other 25% is lost productivity. Therefore, the savings are really $375 per day or $88,125 per year. In the previous section, we started an example that calculated the costs of human factors for redesigning a Web site. The costs were calculated to be $106,750. Now let's calculate the benefits of our imaginary project. The project was aimed at two specific areas: reorganizing the product catalog and improving
TABLE 37.1. Usability Metrics That Help Meet the Corporate Goal of Increasing Revenue Corporate Goal
Measure to Meet Corporate Goal
Increase sales
Increase revenue
Increased customer retention
Usability Measure Increased conversion rate Increased repeat customer visits Decreased drop-off rates Increased customer satisfaction Decreased level of customer frustration Decreased time to complete a task Increased number of functions used
TABLE 37.2. Usability Metrics That Help Meet the Corporate Goal of Decreasing Costs Corporate Goal
Measure to Meet Corporate Goal
Decreased customer support calls Decrease cost
Reduced training costs
Reduced development costs
Usability Measure Increased task completion rates Increased ease of use Decreased error rate Increased ease of use Decreased error rate Increased task completion rate Increased learning rate Decreased test time Decreased number of errors due to design Decreased number of redesigns Decreased number of patches
the checkout process. Based on these goals of the product, the usability metrics collected were drop-off rates (the number of people who leave the site after viewing four or less pages) and customer survey ratings for ease of use for the checkout system. For the example, the ease-of-use rating was 6 on a 7-point scale; up from an earlier rating of 4.5. In addition, let's say that drop-off rates decreased to 45% from 60%. The number of sales increased from 500 per month to 575 per month. Monthly gross sales grew from $50,000 per month to $86,250. Based on these numbers, a compelling argument can be made that the human factors work had a positive effect on site sales. If the numbers are broken down, not only have the number of sales increased, but the average dollar amount per sale has increased from $100 to $150.
($50,000 - 500) = $100 vs. $150 = ($86,250 - 575)
37. Human Factors in Web Design The increased number of sales could be linked to the customer ease-of-use ratings for the improved checkout process and the improved product organization. Customers find the system overall easier to use, and thus are more likely to complete their purchases and to return for future purchases. In addition, customers are able to find a desired product and make a purchase versus dropping off in frustration without finding the product in 'which he or she was interested. These numbers demonstrate the following benefits. Sales increased $36,250 per month ($86,250 $50,000) or $435,000 for the year ($36,250 x 12). A simple calculation of ROI shows returns of $328,250. [$435,000 $106,750 =$328,250]
DEFINITION OF METRICS In addition, when providing ROI information for human factors input, the variables that are being measured should be specific and quantifiable. The definitions used should be precise and unambiguous about what is being measured and how it is measured. For example, to show an increase in conversion rates after the release of a usability-enhanced product, define that specifically as the comparison of the number of people who purchased on the site divided by the number of people who entered the site on date "X" before the usability improvements versus the number of people who purchased on the site divided by the number of people who entered the site on date "Y" after the usability improvements. This will allow anyone looking at the data to be able to understand what the measure was exactly, including when and how it was taken. This helps establish credibility with the reader. A recipient of this information can look and see if conversion rates at the slowest sales time and the highest time were compared, or any other possible confounding factors. Inherent in most of these examples of benefits is the idea that a metric is performing better after the usability enhancement to the Web site. The "better than" assertion can only be made if there is something to compare. The best comparison is to the Web site prior to the usability enhancements being incorporated. Take a baseline of the metrics so a direct comparison can be made of the numbers of the current functionality to the new Web site design. This is the most convincing comparison possible. It is possible to compare metrics when there are not before and after measurements; however, the most compelling and clearest comparison is a before and after look. Other possibilities for comparison might be comparing two different products, where one has human factors integrated into the process and the other does not. There are other creative possibilities; however, the pre-post comparison is the best. There is one more thing that is necessary to adequately present an ROI justification. The findings should be communicated in the language appropriate for the intended audience. In most situations, the information will be provided to people with a business or financial background. A simple measure of benefits minus costs is not a convincing argument for a financial person. Financial people are interested in how long it takes to make an ROI and whether investing in human factors was better than leaving the money in the bank. To develop a
683
convincing argument, one must learn to present the human factors worth in terms of business values, such as the net present value (NPV) or the internal rate of return (IRR). We have gone through an example that computes the dollar value of human factors worth through benefits minus costs. Now we introduce some financial concepts that will help communicate the worth of human factors in terms that a financial person would appreciate.
OVERVIEW OF ROI The primary goal of financial management is to maximize the value of the company. With that, management is continually faced with the question: How should the company, facing uncertainty over future market conditions, invest its capital? These strategic decisions are focused on investments in a variety of real assets, which include tangible assets, such as plant and machinery, and intangible assets, such as management contracts and patents. For the computing or the software industry in particular, these investments include networks, engineers, programmers, and human factors experts. Technology companies invest in these assets with the idea that the assets' expected marginal benefits exceed their expected marginal costs. Valuing these benefits and costs can be difficult because they occur today and in the future. In this section of the chapter, we detail the basic steps for understanding how assets are valued and their relationship to the human factors ROI.
INTRODUCTION TO THE TIME VALUE OF MONEY We stated earlier that financial managers must make decisions to invest the firm's money on real assets that produce cash flows over a fixed time horizon. It is management's hope that today's value of these future cash flows exceeds the amount invested in the assets. We will see that the value of any asset is a function its cash flows, and the challenge for managers is that the cash flows occur at different periods of time with varying degrees of uncertainty. To account for this problem, managers need to understand how asset prices are determined. They need a theory of value (Ross, Westerfield, & Jordan, 2001). The value of any asset (financial or real) is a function of its cash flows, risk, and time. The economic theory that analytically determines this relationship is called the time value of money. In the simplest terms, the time value of money can be defined by the adage "money today is worth more than money tomorrow." It is this precise notion that economically motivates managers to optimally invest their capital. For instance, recall the Web design project for an online sales company that sells "Everything '70's." If the managers decide to redesign their Web page, it will cost them $106,750. Managers commit their capital to this endeavor believing this redesign will increase revenues for the firm. If revenues do increase for the firm, does this mean the venture has been profitable? The answer is "not necessarily." This is because managers have alternative choices today. The Web redesign is only one possible investment that management could make with their capital. Managers are interested in knowing whether the
684
BERTUS AND BERTUS
investment in the Web design will earn them at least as much as their next best alternative investment. To illustrate the previous notion, we are saying that $ 106,750 today is not the same as $106,750 one year from now. This is true because investors have the ability to earn interest or returns. For example, if investors have $106,750 today, they may choose to forgo spending the $106,750 today and invest the money in a 1-year marketable security that pays 10% interest. At the end of 1 year, the investment returns the original principal as well as the interest earned on the principle. That is,
0 $106,750
$106,750 $10,670 $117,425
Because investors have the ability to earn interest on their money, the $106,750 today is not worth $106,750 one year from today, rather it is worth $117,425. We see that money today is worth more than money tomorrow. Moreover, in the Web design case, managers want to know that if they commit their capital, this investment will return the same cash flow 1 year later that is equal to or greater then the next best alternative. Analytically, the value of the 1-year investment can be expressed more formally. The payout at the end of the first time period is
$117,425 = $106,750 + $10,675 To generalize, let the payout at time period one be denned as FV. Furthermore, define the principal as PFand the interest rate as r. Substituting these expressions into the previous equation, we have FV=PV+PVr Factoring yields FV=PV(1+r} The future value of a 1-year investment is equal to the original principal, PV, times an appreciation factor, (1 + r). We have formally expressed the value of today's investment in terms of the next period's dollars. Suppose that managers want to invest their money earning 10% interest for 2 years instead of 1 year. In this case, after 1 year managers will realize a cash flow of $117,425, which they reinvest for 1 more year at 10%. This is a single period problem, and the value is
FV 2 =FV1(l+r) $129,167.5 = $117,425(1.1)
expression for FV1. This is
FV2 = FV1Q. + r) FV2 = PV(1 + r)(l + r) FV2 = PV(1 + r)2 For our example, we have
$129,167.50 = $106.750(1.1)2 The value of the investment at the end of the second year is composed of four parts, which are shown in the following example. 0
1
I
1
$106,750
1 $106,750 $10,675 $10,675 $1067.5 $129,167.5
The first part is the original principal of the investment. The second part is the interest earned on the principal over the first period. The third part is the interest earned on the principal over the second year, and the fourth part is the interest earned on the interest from the first period. The act of reinvesting interest and earning interest on interest is called compounding. In this example, compounding is captured by (1 + r)2. In general, if we want to find the value of an investment today at any future point in time, we may simply use the formula FV=PV(1 +r)t where t denotes a future date. We know how to determine the value of a lump sum of money today at a future point in time. A more interesting question is, "How can we determine what the current value of a promised payment in the future is?" For instance, we know the value of $106,750 one year from now is $117,425 given a 10% interest rate, but sometimes the managers would like to know what is today's value of a promised payment in the future. To answer this question, we may ask it another way. In the context of future values, we may ask, "How much money should we invest today at 10 percent interest so that we have exactly $117,425 one year from now?" Using the expression for future value, we have $117,425=FF(1.1) To solve for the present value, we simply divide the future value by the appreciation factor. This is
$117,425 (1.1) PV= $106,750
PV = If we want to express the future value at time period two in terms of today's investment, then we can substitute the
2
37. Human Factors in Web Design If a management wants to have $117,425 one year from now, then they must invest $106,750 today at 10%. Alternatively, we may say managers with the ability to invest their money at 10% would be willing to buy an asset that pays $117,425 one year from today for $106,750 today. The present value is just the reverse of future value. Formally, the present value is
685
Today, managers would be willing to spend $2,486.85 on an asset that pays out $1,000 a year for the next 3 years. The $2,486.85 is known as the fair market value or market value because this value describes what investors would need to invest in any asset today to receive $1,000 over the next 3 years.
OPPORTUNITY COST OF CAPITAL FV PV = (1 + ry To find the present value instead of compounding the money forward, we must discount it back to today. The expression (1+r)t is called the discount factor, and r is known as the discount rate. So far, we have determined how to price cash flow in the future by compounding it forward or pricing future cash flow today by discounting it back. Suppose we have a stream of cash flow payments across time, can we find the future or present value of this cash flows stream? The answer to this question is yes. Picture the following cash flow stream: 1
0
$1,000
$1,000
$1,000
If managers can earn a 10% rate of return, how much is the cash flow stream worth at the end of the third year? To solve this problem, we can use our knowledge about future values. In 1 year, managers will receive $ 1,000. If they invest this money for 2 years they will receive $1,210 = $1,000(1.1)2 at the end of the third year. In addition, if investors reinvest the second year cash flow they will receive $1,100 at the end of year three. Adding the cash flows together at time period three yields a future value of $3,310. Intuitively, the future value a cash flow stream is Future value = CF1Q(1 + r)n-1 + CF2(1 + r)n-2 + • • • + CFn_1(l+r)1+CFN(1+r)0 Conversely, how much would managers be willing to spend today in order to receive $1,000 each year for the next 3 years? The price that investors are willing to pay today is simply the sum of all the present values for the individual cash flows or Present value =
CF1 CF2 CF3 + + 2 + (1 + r) (1 + r)3 1+r) CFn
+
(1 + rr
The principles of the time value of money show that the market value of any asset is simply the present value of its expected future payoffs. An important element in calculating this value is the rate of return, r, which is referred to as the discount rate or the opportunity cost of capital. This rate of return is referred to as the opportunity cost of capital because it measures in percentage returns the implicit cost of the managers' best alternative, which they have forgone by investing their capital in a particular project. That is, managers are continually faced with multiple investment opportunities, and once they commit their capital to a particular venture, they forego the earnings they could have earned from their next best alternative. This implied cost is called an opportunity cost. Opportunity cost has significance important in terms of valuing investments. In business, opportunity costs for firms are measured by comparable investments. As such, a comparable investment is one that most closely matches the risk level and operations of a proposed investment. For instance, if a technology firm wants to compare an investment in a new call center to an investment in a treasury bond, the company may not properly evaluate the call center project. The T-bond is a much safer investment that will return lower yield. If the technology firm uses this lower rate of return to price the call center project, the firm is likely to overvalue the project and possibly make an ill-informed decision. It is important that managers properly assess their opportunities so they may optimally allocate their capital.
INTRODUCTION TO NET PRESENT VALUE Economic theory tells us that companies achieve their goals when managers maximize a firm's economic profit, which is defined as cash revenues minus cash costs minus the opportunity cost of using firm resources elsewhere. Note the objective of a manager is not to maximize profit, but economic profit. Because opportunity cost measures the value of the financial resources the owners devote to the firm, if the firm maximizes economic profit it earns revenues that cover both actual cash costs and implied costs. In essence, a company that earns a positive economic profit produces goods and services that society values more highly than all the resources consumed in production. Quantitatively, a firm's economic profit is determined by the NPV of its investments. NPV is the difference between an investment's market value and its costs. Formally, we define NPV
This is $2,486.85 =
$1,000 + (1.1)
$1,000 $1,000 —-2 + —— (l.l) (1.1)3
NPV = CF0 + E
CFI
I=1(1+R)I
686
BERTUS AND BERTUS
where CF0 is the initial cash outlay, CFi is the aftertax net marginal benefit at the end of period i, and r is the projects' appropriate discount rate. The expression for NPV can be viewed as an expression for economic profit where the CFi,s represent cash revenues minus cash costs and the discount factor represents the opportunity cost. In simple terms, NPV is a measure for how much value is created for the firm in today's dollars. Therefore, a company should adopt strategies that have a positive NPV, and should not adopt strategies with a negative NPV. Following the NPV criterion then leads firms to direct society's resources to their most productive use. Earlier, we discussed the relevant cash flows for valuing human factors in an information and technology firm such as salaries, equipment, supplies, and revenues. These expected benefits and costs are deemed relevant cash flows because they are the incremental value that the information technology (IT) firms witness by investing in human factors. Therefore, when we want to evaluate the economic significance of the human factors, we need only calculate the aftertax net marginal benefit attributed to human factors and then discount these cash flows by the appropriate opportunity cost of capital. If this value is positive, then human factors is said to add value to the firm. To illustrate, consider the Web design project. "Everything I 70's" invests $106,750 today to redesign the Web site. The project is projected to increase its online sales. The project will last for 3 months and is expected to increase revenues for the retailer by $435,000 for the year. Furthermore, maintenance and support costs for the Web page are expected to be $30,570 for 1 year. If the retailer's tax rate is 30% and has a 10% cost of capital, does the investment seem reasonable? After 1 year, "Everything '70's" plans to net $404,430 = $435,000 - $30,570 The aftertax proceeds for year 1 are
INTERNAL RATE OF RETURN NPV measures, in absolute dollar terms, the amount of value a project will earn for a firm. Using this criterion guides managers to efficiently allocate a firm's capital; however, at times, managers prefer to talk in terms of returns instead of absolute dollars. An alternative method to the NPV is the IRR. The IRR is a rate of return for a project's cash flows that makes the NPV equal to zero. Intuitively, it is a rate of return that balances a project's net marginal benefits to its net marginal costs. Formally, the solution for the IRR is obtained from
CFt
i=1(1+iRR)i
GPo + £
=o
Because the IRR is the implicit or promised return from a project's net cash flows, managers can calculate this promised yield and compare it with the opportunity cost of capital. If the IRR is greater than a project's opportunity cost, the project is said to earn excess returns, thereby adding value to the firm. As a general rule, managers should invest in any project whose IRR is greater than its opportunity cost. For instance, consider the previous example for the retailer looking to increase its online sales. The 1-year project generates aftertax net cash flow in the amount of $283,100 and has costs today of $106,750. Given the "Everything '70's" opportunity cost of capital is 10%, the NPV is $ 150,613. Alternatively, we may evaluate the project by computing its IRR and comparing this promised return with the 10% discount rate. The project's IRR is
$283,100 = $404,430(1 - .3) CFo +
The NPV for the project is
NPV = CF0 +
E
CFI
i=1(1+r)i
= -$106,750+
283,100
= -$106,750+ $257,363 = $150,613 The project has a positive expected NPV; therefore, the retailer should invest in a new Web design. In this example, "Everything '70's" would be willing to spend $257,363 today to receive $283,100 one year from now, yet the retailer only has to spend $106,750 today to receive $283,100 one year from now. "Everything '70's" is essentially purchasing an underpriced asset.
CFi
E E(1+IRR)i
.=0
$283,100 =0 (1 + IRR) „. , $283,100 $106,750 = ——
-$106,750 +
(1+IRK)
$283,100 $106,750 ($283,100) IRR = -1 -1 ($106,750) IRR = 2.6520-1 = 1.652
(1 + IRR) =
The IRR is equal to 165%. This implicit return is greater than the firm's opportunity cost so the retailer should consider investing in human factors to implement the new Web design.
37. Human Factors in Web Design
SUMMARY We described some examples of how to calculate the costs and the benefits for using human factors during designs. We discussed the process of estimating costs and the metrics to measure the benefits. It is important to remember the usability goals and the corporation's goals when calculating these numbers. Metrics should be gathered that relate to the corporate goal that needs to be taken into account (e.g., increase sales), and they should be linked back to corporate profitability. We have shown how to take the cost-benefit analysis and calculate metrics that will be effective with the financial people
687
in a company. This should enable presentation of the results in terms of NPV and IRR. As more and more samples of human factors ROI for products are gathered in the workplace, there will be a history of examples to draw on. As time goes on, it should become easier to predict where, when, and how much time human factors should be allotted to a design project. Eventually, it may be possible to review an upcoming product and estimate the ROI for particular types of usability tasks. This should allow justification of the need for more personnel or help with prioritizing which products should receive the most consideration for human factors intervention.
References Bias, R. G., & Mayhew, D. J. (Eds.). (1994). Cost-justifying usability. San Diego: Academic Press. Black, J. (2002, December 4). Usability is next to profitability. Business Week Online [On-line], Available: http://www.businessweek.com/ technology/content/dec2002/tc2002124_2181 .htm Boehm, B. W (1981). Software engineering economics. Englewood Cliffs, NJ: Prentice Hall. Bosert, J. L. (1991). Quality functional deployment: A practitioner's approach. New York: ASQC Quality Press. Chapanis, A. (1991). The business case for human factors in informatics. In B. Shackel & S. J. Richardson (Eds.), Human Factors for Informatics Usability (pp. 39-71). Cambridge, UK: Cambridge University Press. Donahue, G. M. (2001, January & February). Usability and the bottom line. IEEE Software, 31-37. Donahue, G. M., Weinscheck, S., & Nowicki, J. (1999). Usability is good business. Compuware Report, http://interface.free/fr/archives/ usability/is_good_business/ Fitzgerald, T. (2003). Online retail begins to look like offline: With sales jump, price-cutting, and poorer service. New Media. [On-line]. Available: http://www.medialifemagazine.com/news2003/jan03/jan06/ l_mon/newslmonday.html Karat, C. (1990). Cost-benefit analysis of usability engineering techniques. Proceedings of the Human Factors Society (pp. 839843). Karat, C. (1994). A business cost approach to usability cost justification. In R. G. Bias & D. J. Mayhew (Eds.), Cost-justifying usability (pp. 45-70). San Diego: Academic Press. Landauer, T. K. (1995). The trouble with computers: Usefulness, usability, and productivity. Cambridge, MA: MIT Press. LaPlante, A. (1992). Put to the test Computer World, 27, 75.
Lederer, A. L., & Prassad, J. (1992). Nine management guidelines for better cost estimating. Communications of the ACM, 35(2) 51-59. Leon, M. (2001). How to make sure the customer comes first. Infoworld [On-line]. Available: http://www2.infoworld.com/articles/su/xml/ 01/10/29/011029sustaple .xml?Template=/storypages/ctomover_ story.html Mayhew, D. J., & Mantei, M. (1994). A basic framework for cost-justifying usability engineering. In R. G. Bias & D. J. Mayhew (Eds.), Costjustifying usability (pp. 45-70). San Diego: Academic Press. Nielsen, J. (2003, January 7). Return on investments for usability. Jakob Nielsen's Altertbox. [Online]. Available: http://www.useit.com/ alertbox/20030107.html Pressman, R. S. (1992). Software engineering: A practitioner's approach. New York: McGraw-Hill. Rhodes, J. S. (2000). Usability can save your company. Webword. [Online] . Available: http://webword.com/moving/savecompany.html Roberts-Witts, S. L. (2001, September). A singular focus. PC Magazine. [On-line]. Available: http://www.pcmag.corn/article2/0,4l49,16651, OO.asp Ross, S., Westerfield, R., & Jordan, B. (2001). Essentials of corporate finance (4th ed). New York: McGraw-Hill. Souza, R. (2001, June). Get ROI from design. The Forrester Report. Cambridge, MA: Forrester Research, Inc. Available: http://www.uk.cgey. com/services/crm/docs/roi_design.pdf Wildstrom, S. (1998). A computer user's manifesto. Business Week [Online] . Available: http://www.businessweek.com/1998/39/b3597037. htm Wixon, D., & Jones, S. (1992). Usability for fun and profit: A case study of the design of DEC RALLY version 2. Internal Report, Digital Equipment Corporation.
38 STATE OF THE ART OF WEB USABILITY GUIDELINES Celine Manage ]ean Vanderdonckt Universite catholique de Louvain
Costin Pribeanu National Institute for Research and Development in Informatics
sets of design rules (O'Neill, 1997), and lists of principles (Marshall, Nelson, & Gardiner, 1987). More recently, some effort has been dedicated to introduce design patterns (Borchers, 2000), both for general UI and for the Web (van Welie, van der Veer, & Eliens, 2000). Using usability sources is not straightforward for developers and evaluators, primarily for some of the following reasons:
INTRODUCTION A considerable body of knowledge exists that is dedicated to the usability of human-computer interfaces for computer-based systems. This knowledge is typically referred to as usability guidelines, or guidelines for short. Guidelines can be found in many different formats with contents varying both in quality and level of detail, ranging from ill-structured common-sense statements to formalized rules ready for automatic checking. Guidelines are particularly appropriate for consolidating usability knowledge (Stephanidis & Akoumianakis, 1999) existing for various aspects, such as training workload and universal accessibility. The importance of guidelines was first revealed during the 1980s (Reed et al., 1999) when the use of the computer in the workplace dramatically increased: More computerbased systems were used by more users, who were not necessarily experts, for more interactive tasks that were possibly new or unusual. This situation progressively required that more attention be paid to usability. The advent of Web sites and Web-based applications led to an increase in the amount of sources containing guidelines for Web user interfaces (UIs). Among others, the following were published: seminal books such as those of Nielsen (2000, 2002), style guides produced by individuals such as Lynch and Horton (1999) or by organizations such as Sun (1999), standards such as those published the International Standards Organization (ISO, 1999),
1. Usability remains a quality factor of UIs that is still handled with some uncertainty. Applying guidelines is a necessary condition, but not a sufficient one: The adherence to guidelines certainly contributes to improved usability of a Web site, but a Web site that is compliant with all possible guidelines may still be experienced as unusable by some end users. 2. Identifying in the jungle of guidelines which ones need to be addressed for a particular Web site for a given target audience remains challenging. It is hard to select guidelines appropriate to a particular context of use because guidelines address many different issues: Some guidelines are related to writing code [e.g., how to write pretty and syntactically valid hypertext markup language (HTML)], whereas others encourage the support of the variety of existing Web browsers to ensure accessibility. In this philosophy, the Web should be usable through any browser by any user in the world, whatever the cognitive profile and/or prior experience with computer-based systems. 3. Little or no guidance exists to provide assistance to developers to locate, select, and gather guidelines relevant to their
688
38. Web Usability Guidelines
689
Web site. For instance, only two states of the art in Web usability guidelines are reported: Ohnemus (1997) and Wang (2001). 4. Once identified, guidelines are not usable by themselves. Some guidelines are not precise enough to apply unambiguously and to be assessed objectively once applied. This confusion may be misleading. For instance, many guidelines previously existing for traditional user interfaces, such as graphical user interface (GUI) guidelines, have been considered for Web sites without prior knowledge about applicability. To address these shortcomings, the goals of this chapter are (a) to clarify some fundamental differences existing between the various sources of usability knowledge for the Web to assess their validity, (b) to provide a state of the art summary of major sources containing usability guidelines for the Web with respect to traditional GUIs, (c) to provide more detailed consideration of some selected sources considered as representative in the domain. This chapter intends to inform and guide any person interested in the development of Web pages, to encourage usercentered design, and to provide an overview of ongoing efforts in Web usability guidelines.
TERMINOLOGY In this section, terms used throughout this chapter are denned. Usability: ISO (1999) defined usability as the degree to which specified users can achieve specified goals in a particular environment, with effectiveness, efficiency, and satisfaction, and in an acceptable way. The user can find an element of the interface problematic for various reasons: The system is difficult to learn, it slows down the performance of the tasks, it causes errors of use, or it is dreadful and unpleasant (Wang, 2001). Pearrow (2000) defined usability as a broad discipline based on the scientifically rigorous application of the observation, the measurement and principles of design useful for the creation and the maintenance of the Web sites in order to bring to the final use of the system the ease of use, the speed of training, a high level of utility, and a low level of discomfort. Usability problem: Cockton and Lavery (1999) considered a usability problem to be something that causes a certain difficulty for the end user when interacting with a system. Usability problems typically cover the cause, the location, and the explanation of any potential problem induced by an interface. The objective of a usability evaluation is to identify as much as possible potential problems that users could have with an application (Jeffries, 1994). Wang (2001) defined usability problems as UI aspects that are likely to affect the system usability from the user viewpoint. Head (1999) quoted the current problems of usability for the Web: • • • • •
Wording and the vocabulary are not clear Users must remember too many things Graphics are useless and overused Understanding the site design is approximate Correspondence between the site design and the users' needs remains vague • Navigation poses problems
FIGURE 38.1. Ergonomic knowledge in a style guide. • Site is conceived without a population target well denned • Design is not guided by the user's goals • Privacy and safety are insufficient Style guide: A style guide is used as a basis for the development of a Web site and consists of a collection of principles, rules, and conventions gathered to define a unified look and feel for products and services (Ohnemus, 1997). Usability guidelines: Guidelines span from high-level expressions that apply to a wide variety of cases to low-level statements that are limited to specific families of cases (Ohnemus, 1997; Stephanidis & Akoumianakis, 1999; Fig. 38.1): 1. Principles are aims and goals guiding design decisions that occur during the development life cycle. They reflect broad knowledge about behavior of human interacting with computers. Therefore, principles are generally stated in terms such as "using consistent images and metaphors with the external world in appearance and the behavior." They are objectives of the highlevel design. 2. Rules are based on principles specific to a given field of design. For instance, a rule can stipulate "use a consistent presentation and a visual language throughout the site." These rules are prone to more interpretation that can reflect the needs for a particular organization or a case for design. 3. Conventions or recommendations dictate specific conceptual decisions to follow and should reflect the needs and terminology of the organization. They are unambiguous statements relating to physical artifacts. The usability guidelines contained in a style guide can be principles, rules, or recommendations (Fig. 38.1). Certain rules are validated by experimental results provided by user testing, laboratory experiments, or other techniques, and others are not. There are thus different levels of validity. A Likert scale of five levels is commonly presented (Grose, Forsythe, & Ratner, 1998) to judge level of validity of the rule (Fig. 38.2). It goes from nonessential to essential to guarantee the usability. In this case, the intermediate level means important.
690
•
MARIAGE, VANDERDONCKT, PRIBEANU
1
2
Nonessential
3
4
Important
5
Essential
FIGURE 38.2. Levels of validity of usability guidelines.
TRADITIONAL GRAPHICAL USER INTERFACES VERSUS WEB USER INTERFACES Having clarified the terminology to be referred to, this section identifies differences and similarities between traditional GUIs, as one can find them in most applications, and Web UIs. Based on this identification, differences and similarities between usability guidelines addressing both families are discussed. Does the Web UI differ from a GUI and to what extent? Butler (1997) argued that strong similarities exist between the two types of interface (e.g., the same guidelines may be applicable to both families, various target populations can use them, the two types can be very attractive). However, Nielsen (1997) considered the Web to be a type of application that is under control of the user and not of the designer. Significant differences between the two types of UIs cannot be identified merely by examining individual properties, but by analyzing a series of elements (Scapin, Bastien, et al., 2000). Three elements significantly change for the two types of interface (Table 38.1): the various actors involved in the development of the Web site, the end users of the UI, and the responsibility for producing the contents. User Interface Design and Development Who designs and/or develops the application? The quality control of designed UIs is variable. Theoretically speaking, a Web page can be designed by almost anyone,
although only designers trained with usability actually produce quality pages. Web sites consequently range from personal sites developed by individuals to professional sites developed by a complete team. In contrast, the temptation to design UIs of limited interest, with useless sophistications, seldom exists for GUI, except perhaps for shareware. What is the goal of the UI? A GUI is typically aimed at an application, whereas a Web page can be of any type (Grose et al., 1998): • Informative (to disseminate information and to constitute a navigation point toward other pages), applicative (to manipulate and derive information from queries), educational (to teach contents to learners) • Commercial (to buy and sell products) • Any combination thereof. Wang (2001) distinguished four types of informative sites: informational site, research site, multipurpose portal, and leisure site. The goal of a Web page also depends on the site type: Internet site (from an organization to any external audience), intranet (internal to an organization only), extranet (from an organization to partners with whom information should be exchanged; Riggins & Rhee, 1998). Which technologies are used? Environments for developing GUIs are believed to be less volatile and evolving than those for Web sites. Some years ago, only HTML was needed to produce a Web page. In contrast, many additional scripting languages (e.g., CGI, Perl, ASP, JavaScript), programming languages (mainly Java), dedicated markup languages (e.g., XML languages, DHTML), and proprietary languages (e.g., Macromedia Flash, Shockwave, Director, ColdFusion) have flourished in recent years, thus increasing the complexity of Web development.
TABLE 38.1. Differences Between Interfaces Web and GUI Designer /Developer Who Nature Technology Disciplines Usability
Interactivity
Life cycle of application User
Content responsible
Web Professionals and nonprofessionals (almost anyone can design a Web page) Interface oriented toward navigation in contents Low risk in deployment, user testing Information architecture, human factors, graphics, marketing, etc. Depending on the profile of designers/developers Usability may be hard to control because Web navigators and user populations vary Ranging from almost noninteractive (contents viewing) to highly interactive (depending on technology used) Fast development, short life time, risk to disappear quickly No license needed and no installation; hence, sites are competitive and switching is frequent Content is updated regularly
GUI Professionals Interface oriented toward functionality and application domain Moderate risk in deployment, software testing Information technology and application domain specialists Depending on the development process followed Usability can be built in the software with no UI variation Generally highly interactive Potentially with immediate feedback and direct manipulation Moderately long development, long life time, stay stable for a while License and installation are required; software is moderately competitive New versions are produced from time to time
38. Web Usability Guidelines Which disciplines take part in the design and/or the development? The Web calls upon competences coming from disciplines that were less involved in traditional GUIs: graphic design, electronic publishing, marketing, branding image, and management. This situation explains why a larger variety of elements are studied in Web design than for GUIs: task support, usability, aesthetics, attractiveness, emotion, and attention (Wang, 2001). How is usability considered? Web UIs are rendered individually by a navigator on a given computing platform, thus releasing the control of some parameters expected by the designers (Scapin et al., 1999). For example, keyboard shortcuts are governed by some functionalities that are specific to each navigator, thus eliminating the need to address related guidelines. Font sizes set by a designer may be overwritten by values set up by a user in the navigator. In addition, fonts of identical families can be rendered differently, depending on the resolution of the user's screen. What are the UI levels of interactivity? Depending on the site type, its related level of interactivity can range from low (passively browsing pages) to high (with Java, Flash). This variation remains more or less similar for GUIs. How can the life cycle of the UI be characterized? Due to a longer history and a longer-standing experience, GUIs are potentially developed based on a more established body of knowledge than UIs for the Web.
User Interface Users Web site visitors do not own the sites (even if they pay to visit), whereas a GUI user often bought a copy of the software. Thus, visitors easily switch from one site to another when unpleased, confused, or unhappy with the UI. In contrast, users of software GUIs feel forced to stick to the same GUI, becoming accustomed to it by solving problems they encounter rather than switching to another software package (Scapin et al., 1999).
Content Responsibilities The content of a Web site is expected to be more regularly updated than the content of a traditional interactive application. Similarly, new versions are expected to appear more frequently.
potentially all types of UIs to those specific for the Web, we note the following trends: • The emergence of new topics covered by Web guidelines that were not previously covered. Accessibility guidelines present recommendations to transform a Web site so it can be used by the widest population possible of users, including users having disabilities and using limited computer resources. For example, adding an alternative text to each image allows screen readers to synthesize this text for describing the image. e-Commerce sites have to sell products in a highly competitive market so usability is a key issue for differentiating Web sites (Nielsen, 2001). Marketing issues play a more crucial role as Web sites can play the role of a window, contain advertisement, or convey an image of the organization. Marketing goals may meet or contradict usability principles and therefore need to be reconciled. For instance, the marketing goal inviting any visitor to browse several categories of information may be in conflict with the principle of reducing the workload of the user. Personalization has been studied in adaptable and adaptive GUIs, but techniques for adapting the contents, the interaction, and the UI with respect to user needs have been largely developed under the banner of the Web, as more shared resources are available. For example, collaborative filtering displays information personally interesting to the user according to previous interaction histories from the same user and users belonging to a similar cluster. Collaboration technologies also developed rapidly with the advent of the Web: forum, chat room, shared spaces, virtual spaces, and so on. • The type of guideline has evolved over the years (Fig. 38.3). Guidelines existing for traditional GUIs in the past may simply disappear (line 1 of Fig. 38.3) as they are no longer valid for the Web. New guidelines that did not previously exist appear, such as those in the new topics identified previously (line 2 of Fig. 38.3). Existing guidelines can be transferred "as is" if their expression remains understandable for the Web (line 3 of Fig. 38.3). If not, they are modified accordingly to produce a new guideline (line 4 of Fig. 38.3) based on restriction/extension, specialization/generalization, or deep modification. For instance, the generic principle "Ordering of list should be designed to assist readers' task an in a hypertext" (Hardman et al., 1990) is refined into "Use unordered list Traditional application GUI
Web site UI
WEB GUIDELINES VS GUI GUIDELINES Evolution of Guidelines The earliest UI guidelines appeared in the 1980s (e.g., Mitre Corporation Guidelines; Smith & Mosier, 1986), whereas Webspecific guidelines appeared only in the 1990s [e.g., Nielsen (1995), Lynch & Morton (1999)]. When observing the evolution of guidelines over the years, from guidelines addressing
691
FIGURE 38.3. Types of guideline evolution.
692
MARIAGE, VANDERDONCKT, PRIBEANU
TABLE 38.2. Example of Design Rules Every Web page needs An informative title The creator's identity (author or institution) A creation or revision date At least one link to a local homepage The homepage URL on the major menu pages in your site Note. From Web Style Guide, by P. J. Lynch and S. Morton, 1999, Yale University Available: http://info.med.yale.edu/caim/manual. Copyright 1999 by Reprinted with permission.
FIGURE 38.4. Types of guidelines and sources. Note. From "Accessibility guidelines and scope of formative HCI design input: Contrasting two perspectives," by C. Stephanidis and D. Akoumianakis, 1999, User Interfaces for All, Proceedings of the 5th ERCIM Workshop, p. . Copyright 1999 by Adapted with permission. to present items of equal status or values, i.e., •when the items do not contain an inherent sequence, order, or rank" (Detweiler & Omanson, 1996). Finally, one or many existing guidelines can be assembled or disassembled by composition/decomposition to produce a new series of guidelines for the Web (line 5 in Fig. 38.3).
Categorization of Web Guidelines Usability guidelines exist in many different forms and can be found in various sources of information (Scapin et al., 2000). To express the differences between these sources, each Web guideline is classified (Fig. 38.4) by type (ranging from the most general to the most specific: principles, guidelines, and recommendations) and source. Principles are general objectives guiding conceptual UI decisions. They reflect the knowledge around human perception, learning, and behavior, and are generally expressed in generic terms such as "Use images and metaphors consistent with real world" so they can be applied for a wide range of cases. Guidelines are based on principles specific to a particular design domain. For example, a Web design rule can stipulate to "use a consistent look and a visual language inside the site." Some guidelines have to be interpreted more and altered to reflect the needs of a particular organization or design case. Recommendations (also called conventions) univocally determine conceptual decisions specific to a particular domain of application, and should reflect the needs and the terminology of a given organization. They are unambiguous statements so no place for interpretation is left. Recommendations include design rules and ergonomic algorithms. Design rules consist of a set of functional and/or operational requirements specifying the deisign of a particular interface. They do not require any interpretation from designers or developers, thus reinforcing constant application throughout
different cases. Typically, they cover screen format, window templates, navigation bars, definition of frames, and location of contents with respect to navigation controls (Table 38.2). Ergonomic algorithms typically translate simple design rules into understandable and systematic procedures that can be applied more rapidly than a set of simple rules. In this way, they introduce some flexibility enabling designers to select appropriate values of parameters controlling design rules and preventing designers from forgetting any design rule. Such algorithms typically exist more like procedures in a software than instructions in a paper manual. As such, they are intended to be systematically executed in a consistent way. For example, the software Design Advisor (Faraday, 2000) contains an ergonomic algorithm that predicts the visual path between elements on the Web page, depending on their type, size, color, and location (Fig. 38.5). The algorithm assumes for instance that the animations will first attract the eye (according to some guidelines), then large images, then large text, and so on. In general, a style guide is defined as a set of guidelines and/or functional or nonfunctional requirements ensuring the consistency of a collection of different interfaces. This collection can be specific to an operating system (e.g., Windows, see Microsoft Corporation, 1999), to a computer manufacturer (e.g. , IBM CUA), to a software vendor, to an activity domain, or to an organization. This definition tends to change for Web sites because a Web style guide can simultaneously contain principles, guidelines, and recommendations. Web style guides are more homogeneous because a common look and feel is developed (Steward & Travis, 2002). Usually, design rules govern UI design very precisely, (e.g., with guidelines such as "Every web page should end with the author's email formatted in Arial 10 pt and associated with email link"). On the contrary, style guides can be considered more generic than design rules (Scapin, 1990b) because they enlarge the scope of design rules (e.g., with guidelines such as "Every web page should be terminated by a linkable author's email"). A Web style guide is applicable to large sets of Web guidelines, provided that these sets are structured to provide designers with assistance in designing usable Web sites. Web style guides differ from traditional style guides in a number of ways (Grose et al., 1998; Ratner et al., 1996): • They insist less on traditional aspects of interface usability. • They emphasize technical problems encountered by the authors with distributed hypermedia systems. • Usability is a priority of a hypertext environment, such as consistency (navigational, graphical, etc.) inside the Web pages.
38. Web Usability Guidelines
693
FIGURE 38.5. Design Advisor interface. • Only 20% of the usability guidelines applied to the Web were actually contained in established Web style guides. Limited consistency was found among twenty-one studied Web style guides, with 75% of the recommendations appearing in only one style guide. Grose et al. (1998) attributed some of these differences to the Web introducing a unique genre of UI, often promoted by a common look and feel found in Web style guides. In a Web environment, where the user navigates between interrelated sites, a common look and feel is more important than in stand-alone applications where differences are less apparent. The development of Web style guides appear to be less rigorous and contain less references than for traditional GUIs. Links between high-level principles to last-level recommendations are less obvious. Significant examples of Web style guides include Berners-Lee (1998), IBM (1999), National Cancer Institute (2002), Builder.com (2002), Webreview.com (1998), Sun (1999), and Nielsen (2000). Standards consist of a formal document containing functional and/or operational requirements standardizing the design of UIs (Steward & Travis, 2002). Standards are promulgated by national or international organizations of standardization. They can be military, governmental, civil, or industrial. According to Nielsen (1993), the three big types of standards are national or international, industrial, and in-house. International standards have more significance because of the importance of their organization (e.g., the European Union directive arguing that "software must be easy to use and software usability principles have to be applied"). Industrial standards specify look and feel of interfaces in detail. Unfortunately, these standards may appear mutually conflicting. In-house standards have to be understood and applied directly by developers within an organization. Today, there is no dedicated Web standard, but some standards are considered to be good candidates for adapting to the Web: ISO 9241 (1999) or HFES/ANSI 200 (1997).
Standards related to UIs have been extensively researched in more recent years at the international level. ISO started a working group to define a standard for the Web, but at the time of this writing, we do not have any information on it. Groups dedicated to interfaces were created inside standardization organizations. Nevertheless, the standards for a good design always need to be reconsidered as new questions arise and new problems appear (Scapin, 1990b). Significant examples of standards include ISO 9241 "Ergonomic Requirements for Office Work with Visual Display Terminals" (ISO, 1999), HFES/ANSI200 "Standard: Draft, Human Factors and Ergonomics Society (HFES/ANSI 200,1997), and ISO 9126 "Quality of Software Systems." Isolated guidelines prescribe a statement to be applied for a UI, sometimes with examples, with or without rationale or comments. Each prescription results generally from a human consensus from people involved by the guidelines (e.g., users, designers, developers). This process is obvious when the prescription is empirically tested and validated. As such, they are often published in conference proceedings or in journal papers, which are not always accessible to a nonaware audience. Each guideline can be dedicated to a particular aspect of usability (e.g., the user expectation of objects located on a Web page) or to a family of tasks and domains. For the Web, isolated guidelines can be applied to a particular aspect of usability, as accessibility (Vanderheiden et al., 1997). Some examples of isolated guidelines include Borges, Morales, & Rodriguez, (1998), Nielsen (1995), Comber (1995), W3C (1999), All Things Web Site (http://www.pantos.org/atw/), Leulier, Bastien, and Scapin (1998), Nielsen (2000), and Nielsen (2002). Ml patterns adopt a different approach in trying to condense the application of several isolated guidelines into a comprehensive design that is supposed to be usable by construction. Moreover, patterns present a global solution to a family of design problems that are similar across different computing platforms or across various styles of Web sites. Unlike guidelines that are often presented out of context, patterns have the advantage to
694
MARIAGE, VANDERDONCKT, PRIBEANU
Type A
Type B
ColumnLabel
ColumnLabel
D ItemName1
ItemName1
Function1
Function2
D ltemName2
ltemName2
Function1
Function2
ItemName3
Function1
Function2
D ItemNarne3 Function1
Function2
When an tern is added to thelist,the view on the list shows the added item by highlighting it, as feedback to the users that the operation has been performed correctly. If neccessary the list should "scroll" to the position of the new item in the fist. Why By showing the overview first the users always know what the current status is. Editing functionality is then seen as "operations" on the current list. More Examples This example from the Hotmail service shows a type A solution.
Related Patterns When lists get longer (more than 10-15 items), consider a Paging mechanisme for making it easier to find particular items.
FIGURE 38.6. An example of user interface pattern. Note. From www. welie.com/patterns/list-builder.html. be presented with a context of use that is assumed to be the one in which the pattern can be applied (Fig. 38.6). A rationale then explains why the particular design suggested in the pattern can be applied. Each pattern can be exemplified either by generic examples that are platform independent or by specific examples, such as for the Web or for a family of consistent Web sites. The previous classification of guidelines represent one attempt to uniformly structure them from the most general to the most specific, as supported in the MetroWeb tool (Mariage & Vanderdonckt, 2003—http://www.isys.ucl.ac.be/ bchi/research/metroweb.htm). Another classification of guidelines at multiple levels, as supported in the bull's-eye framework (Beier & Vaughan, 2003), also starts from the most general to the most specific: overarching features and principles, combination of page flows to create interaction pattern, page flows, page template, individual or combined components.
SHORTCOMINGS OF GUIDELINES USAGE Usability guidelines typically suffer from a series of shortcomings that reduce the impact of their usage and their scope within the development life cycle of UI development. Major shortcomings are as follows: 1. The expressiveness and the trust in the guideline validity heavily depend on the guideline source. Guidelines sources can be sorted along an axis denoting the interpretation required by their application: high-level principles require an abstract interpretation of their statement, which can lead to many different processes, whereas low-level design rules are
FIGURE 38.7. Guideline interpretation according to guideline source. written to no longer require any concrete interpretation. The more general a guideline, the wider its scope of application. For instance, the guideline, "Widgets should be selected according to the task type", is considered more general than the guideline, "A list box should be selected to input the customer's country." The first guideline requires that the designer understand the user's task, have some understanding of the available controls, and be able to match the appropriate controls to the user's task. 2. Nearly att guidelines require some interpretation. This activity may vary significantly from one person to another. The contents of a usability guideline may be transmitted in a format such that the setup conditions that served for the experimental validation of the guideline disappeared, thus preventing the reader to know when and how the guideline can be applied. The lack of such conditions may also invalidate the application of a guideline. Specific guidelines, such as design rules, do no
38. Web Usability Guidelines
•
695
FIGURE 38.8. Aspect coverage in guidelines sources. require such interpretation. However, the scope of their application can be so narrow that a lack of applicable guidelines can be identified. Figure. 38.7 shows that the guideline source determines the need for interpretation: General guidelines as found in standard generally require an extensive interpretation because their expression is assumed to be abstract to cover a wide range of possibilities. Conversely, design rules are written almost to no longer require any interpretation and to be applied in a straightforward manner. In this case, little or no room for interpretation is left. 3. The jargon used in the initial guideline may slow down designers. The vocabulary of the discipline used to experimentally validate the guideline may appear difficult to understand to people who do not belong to this discipline. Some extensive experience may be needed to avoid any misleading generalization or invalid specialization of results. 4. Guidelines are not equally distributed across all linguistic levels. Nielsen (1986) introduced a model that decomposes human-computer interaction into a series of seven subsequent layers: goal, pragmatic, semantic, syntactic, lexical, alphabetic, and physical. Usability guidelines located at lower linguistic levels are believed to be easier to interpret and apply than those located at higher levels. Guidelines located at syntactical and lexical levels are by far the most numerous, thus identifying a need to develop other guidelines for underexplored levels. 5. Applying and checking guidelines require varying workloads. The workload implied by the application of a guideline and the checking of a Web page against this guideline depends on the linguistic level, the quality of the guideline contents, and its scope. A given guideline can be applied for every Web page of a Web site, independently of any context of use, whereas other guidelines may be concerned only with widgets
or contents of a Web page in a given context of use. Unfortunately, many guidelines are delivered free of context of use, thus preventing people from knowing when and how to apply them.
CONCLUSION In this chapter, we highlight that a profusion of Web usability guidelines exists. However, there is little to facilitate the designer's task to select and apply appropriate guidelines. Usability guidelines dedicated to the Web should be clearly differentiated from guidelines for traditional GUI applications. Some sources, such as standards, profess to be sufficiently general to accommodate a wide variety of design issues. Although this assumption is probably true for traditional applications, it is unlikely to be the case for Web usability. Several factors suggest that Web
FIGURE 38.9. How to select appropriate guidelines.
696
MARIAGE, VANDERDONCKT, PRIBEANU
W3C Printable hypertext In an ideal world, paper might not be necessary. In a next to ideal world, one would have enough time to write a hypertext version of a document and also to write a completely separate paper version However, the real world, you will probably want to generate any printed documents and online documents from the same file Suppose the HTML files will be the master, and you will generate the printable from this, by making one long document, and possibly printing it via translation into TeX 01 some word processor format, for example. You might not initially, but you might want to one day. Try to avoid references in the text to online aspects. "See the section on device independence " is better than "For more on device independence, click here .". In fact we are talking about a form of device independence . Unfortunately, the recommended practices of signing each document and giving navigational links tend to mess up the printable copy, though one can of course develop ways of stripping them out if they follow a common format. For example, the most common comment about this document was that it is difficult to print. I therefore made a single page version of the whole thing with a few scripts, and put a pointer to it from the cover page. But then people still ask, not having read the cover page. (The scripts were just bits of "sed", which I am not supporting, I have put rules in at the top and bottom of each page and the scripts use these to chop off bits which are not needed in the printed copy.) (Up to: within each document; back to Using Style Sheets, on to readable text)
FIGURE 38.10. Recommendation for printing a hypertext document.
Option Buttons & Chech Boxes
Option (radio) buttons and check box examples, When to Use
* Use option buttons (also referred to as radio buttons) to allow the user to select a single choice from a set of mutually exclusive choices. • Use check boxes to allow the user to set values that have two clearly opposite states such as "on" and "off.
Design & implementation
9 Use two option buttons instead of a check box for any field that does not have opposite and unambiguous states. . One option button should always be selected by default • If a field would require a large number of option buttons (more than 4), consider using a drop-down list box instead. FIGURE 38.11. IBM guidelines for using option buttons and check boxes.
38. Web Usability Guidelines usability guidelines should be treated separately. These include differences discussed in UI design and development both for GUI and the Web, the emergence of new questions for the Web that were unprecedented, and the rapid evolution of guideline types. For example, GUI mainly focuses on guidelines for controls, whereas the Web divides guidelines for navigation (the privileged interaction mode) and for forms, although Java applications can be considered similarly to GUI applications. Many holes in usability knowledge still exist for the Web. To identify them, Basden (2003) used Dooyeweerd's (1955) notion of irreducible aspects, according to which there is a series of fifteen aspects, each having a set of laws that enable meaningful functioning in everyday living. Basden (2003) then compared the Yale Style Guide (Lynch & Horton, 1999) with other style guides for the Web to discover (Fig. 38.8) that some aspects are largely represented (e.g., spatial and kinematic in a Web site are well described), whereas others remain underrepresented (e.g., the juridical aspect that discusses legal issues for the Web is rarely found, as well as ethical issues). Figure. 38.8 shows potential areas where a need for guidelines exists that is not filled, thus providing an opportunity for researchers. To select appropriate guidelines, Fig. 38.9 classifies different types of guidelines according to two dimensions: the need of interpretation they require before being applied and the quantity of implementation details provided in the guideline definition. At the bottom right of the framework are located principles that require a lot of interpretation and offer little or
no implementation guidance. On the other extreme, at the top left, ergonomic algorithms are probably the most deterministic as they build a Web interface entirely by applying design rules, which are in turn less detailed for implementation than algorithms. Interestingly, UI patterns reduce the gap of interpretation, but leave the developer free to implement the pattern as he or she wants, provided that it is compliant with the pattern definition. A guideline in isolation keeps the same freedom, but requires more interpretation. A series of potential problems has been identified previously. Standards typically include principles and/or compilation of high-level guidelines because the desire for them to be general is predominant. Figure. 38.8 also reveals that the term "style guide" can be found for designating many different types of usability knowledge for the Web, which was not the case for traditional GUI applications. Finally, one may appreciate that several efforts exist today to express usability knowledge in a common format that is sharable among teams. For example, guideline definition language (GDL) is an attempt to relate each guideline with several interpretations, each of them being decomposed into sets of evaluation conditions imposed on UI elements (Beirekdar, Vanderdonckt, & Noirhomme-Fraiture, 2002). These sets can then be subject to optimization of evaluation. Equally important, pattern language markup language (PLML) defines a common format for UI patterns to create a distributed body of knowledge that is consistent across sources. In this way, patterns can also be compared and become incremental (http://www.hcipatterns.org).
1 KeepnavigationAidsConsistent Guideline: Use the same navigation aids (navigation scheme) on all pages. Comments: Create a common navigational look to ensure that users can use the Web site navigation effectively,
Strength of the evidence: 00000 •How to interpret "strength of evidence" scale
697
Source: Detweiler, M.C. and Omanson, R.C. (1996),
Ameritech Web Page User interface Standards and Design Guidelines fwww.ameritech.com)
Example:
FIGURE 38.12. Guideline promoted by the National Cancer Institute.
698
MARIAGE, VANDERDONCKT, PRIBEANU
Titled Borders for Panels The JFC enables you to specify a titled border for panels, which you can use as containers for components inside your application's windows. Figure 31 Spacing for a Panel With Titled Border
Since titled borders take up considerable space, do not use them to supply titles for components; use labels instead. Use a titled border in a panel to group two or more sets of related components, but do not draw titled borders around a single set of check Use titled borders sparingly: they are best when you must emphasize one group of components or separate one group of components from can be distracting and more confusing than simply grouping the elements with a design grid. Never
nest titled borders. It becomes difficult to see the organizational structure of the panel and too many lines cause distracting optical e
Insert 12 pixels between the edges of the panel and the titled border. Insert 12 pixels between the bottom of the title and the top of the first: component groups and between the bottom of the last component and the lower border. Allow for internationalized titles and labels in panels that use titled borders. A titled border can be created as follows: myPanel.setBorder (new TitledBorder (new LineBorder (HetalLookAndFeel.getControlShadow()), " > " " ) ) ;
FIGURE 38.13. Example of a guideline provided by Sun.
APPENDIX Here, some sources for Web usability guidelines are briefly presented: Berners-Lee's (1998; http://www.w3.org/Provider/Style) guide provides many directives for writing the contents of a Web page (Fig. 38.10). This guide is more oriented toward writing Web pages than toward the pure usability of a Web site. IBM Web Design Guidelines (1999; http://www-3.ibm.com) promote online various recommendations related to Web usability as an activity structured into phases corresponding to stages of the development life cycle: planning, development, maintenance, and evaluation. For this reason, it goes beyond a mere design guide. Figure. 38.11 shows a
guideline precising the role played by option buttons and check boxes. The National Cancer Institute's (2002; http://usability.gov/ guidelines) Research-Based Web Design and Usability Guidelines present guidelines in a systematic way that incorporates the strength of the evidence of each guideline (Fig. 38.12). This parameter can range from a score of 1 on 5 for a guideline having only limited cases where the guideline was successfully applied to a score of 5 on 5, where the guideline has been tested according to several experimental studies. This parameter is particularly useful to judge the validity of each guideline as reported in the guideline source. Builder.com's (2002; http://builder.cnet.com/webbuilding/ pages/Graphics/CTips2/index.html) Designing Pages and Sites presents recommendations for the Web exemplified with copious illustrations. The planning of the Web site is also covered.
38. Web Usability Guidelines Webreview.corn's (1998; http://www.webreview.com) Navigation and Usability Guide. Focuses more on issues raised by browsers and cascading style sheets. Although these aspects are technically important, they do not raise the true usability questions. No particular structure was found.
699
Sun's (1999; http://java.sun.com/products/jlf/edl/dg/index. htm) Java Look and Feel Design Guidelines provide the reader with online recommendations for Java-based applications and their integration within the Web environment. Figure. 38.13 provides an example of such guidelines.
References Basden, A. (2003). Levels of guidance. In C. Stephanidis (Ed.), Proceedings of 2nd International Conference on Universal Access in Human-Computer Interaction UAHCI'2003 (Vol. 4). Mahwah, NJ: Lawrence Erlbaum Associates. Beier, B., & Vaughan, M. W (2003). The bull's-eye: A framework for web application user interface design guidelines. Proceedings of CHI 2003 (pp.489-496). New York: ACM. Beirekdar, A., Vanderdonckt, J., & Noirhomme-Fraiture, M. (2002). A framework and a language for usability automatic evaluation of Web sites by static analysis of HTML source code. Proceedings of 4th International Conference on Computer-Aided Design of User Interfaces CADUI'2002 (pp. 337-348). Dordrecht: Kluwer Academic. Berners-Lee, T. (1998). Style guide for online hypertext [On-line]. Available : http://www. w3. org/Provider/Style Borchers, J. O. (2000). A pattern approach to interaction design. Pattern languages. Proceedings of DIS'OO: Designing Interactive Systems: Processes, Practices, Methods, & Techniques, 369-378. Borges, J. A., Morales, I., & Rodriguez, N. J. (1998). Page design guidelines developed through usability testing. Human Factors and Web Development, 137-152. Builder.com. (2002). More great tips from CNET designers, Part 2 [On-line]. Available: http://builder.cnet.com/webbuilding/pages/ Graphics/CTips2/index. html Butler, S. (1997). Why Web design is similar to GUI design? [On-line]. Available: http://www.useit.com/alertbox/scottbutler.html Cockton, G., & Lavery, D. (1999). A framework for usability problem extraction. In M.A. Sasse & C. Johnson (Eds.), Human-Computer Interaction—INTERACT '99: Proceedings of the Seventh IFIP Conference on Human-Computer Interaction (pp. 344-352). London: IOS Press. Comber, T. (1995). Building usable Web pages: An HCI perspective. Proceedings of Australian Conference on the Web AusWeb'95,119124. Detweiler, M. C., & Omanson, R. C. (1996). Ameritech Web page user interface standards and design guidelines [On-line]. Available: http://www.ameritech.com. Dooyeweerd, H. (1955). A new critique oftheoritical thought. Ontario: Paideia Press. Faraday, E (2000). Visually critiquing Web pages. Proceedings of HFWeb'OO. Grose, E., Forsythe, C., & Ratner, J. (1998). Using traditional style guides to design Web interfaces. Human Factors and Web Development, 121-136. Hardman, L., & Sharrat, B. (1990). User-centered hypertext design: The application of HCI design Web principles and guidelines. Hypertext State of Art, 252-258. Head, A. J. (1999). Design wise: A guide for evaluating the interface design of information resources. Information Today. HFES/ANSI 200. (1997). Standard: Draft, Human Factors and Ergonomics Society. IBM. (1999). Web design guidelines [On-line]. Available: http://www3.ibm.com.
International Standards Organization (ISO). (1999). Draft International Standard (DIS) 9241. Requirements for office work with visual display terminals. Jeffries, R. (1994). Usability problem reports: Helping evaluators communicate effectively with developpers. In J. Nielsen & R. L. Mack (Eds.), Usability inspection methods (pp. 273-294). New York: Wiley. Leulier, C., Bastien, J. M. C., & Scapin, D. L. (1998). Compilation of ergonomic guidelines for the design and evaluation of Web sites, Commerce and Interaction Report, INRIA, Rocquencourt, France. Lynch, E J., & Horton, S. (1999). Web style guide, Yale University [Online]. Available: http://info.med.yale.edu/caim/manual. Manage, C., & Vanderdonckt, J. (2003). MetroWeb: a tool to support guideline-based Web evaluation. In J. Jacko & C. Stephanidis (Eds.), Proceedings of the 10th International Conference on HumanComputer Interaction HCI International'2003 (Vol.1, pp. 158162). Mahwah, NJ: Lawrence Erlbaum Associates. Marshall, C., Nelson, C., & Gardiner, M. M. (1987). Design guidelines. In M. M. Gardiner & A. Christie (Eds.), Applying Cognitive Psychology to User-Interface Design. Microsoft Corporation. (1999). Microsoft Windows user experience: Official guidelines for user interface developers and designers. Microsoft Press. National Cancer Institute. (2002). Research-based Web design and usability guidelines [On-line]. Available: http://usability.gov/ guidelines. Nielsen, J. (1986). A virtual protocol model for computer-human interaction. International Journal of Man-Machine Studies, 24(3), 301-312. Nielsen, J. (1993). Usability engineering. Boston: Academic Press. Nielsen, J. (1995, December). Guidelines for multimedia on the Web. Alertbox [On-line]. Available: http://www.useit.com. Nielsen, J. (1997). The difference between Web designing and GUI design. Alertbox [On-line]. Available: http://www.useit.com/ alertbox/9705a.html Nielsen, J. (2000). Designing Web usability: The art of the simplicity. Indianapolis: New Riders Publishing. Nielsen, J. (2001, August 19). Did poor usability kill e-commerce? Alertbox [On-line]. Available: http://www.useit.com. Nielsen, J. (2002). Homepage usability. Indianapolis: New Riders Publishing. Ohnemus, K. R. (1997). Web style guides: Who, what, where. Proceedings of the 15th Annual ACM International Conference on Computer Documentation SIGDOC'97, 189-197. O'Neill, P. (1997). CUL Web design guidelines, Version 2.0 [On-line]. Available: http://www.library.cornell.edu/design/dsigndoc.html Pearrow, M. (2000). Web site usability handbook. Rockland, MA: Charles River Media. Ratner, J., Grose, E. M., & Forsythe, C. (1996). Characterization and assesment of HTML style guides. Proceedings of CHI'96, Interactive Posters, 115-116.
700
MARIAGE, VANDERDONCKT, PRIBEANU
Reed, P., Holdaway, K., Isensee, S., Buie, E., Fox, J., Williams, J., & Lund, A. (1999). User interface guidelines and standards: Progress, issues and prospects. Interacting with Computers, 12(2), 119-142. Riggins, F. J., & Rhee, H. S. (1998). Towards a unified view of electronic commerce. Communications of the ACM, 41(10), 8895 [On-line]. Available: http://sol.brunel.ac.uk/~jarvis/tbilisi/ilhec/ teachingresources/p88-riggins.pdf Scapin, D. L. (1990a). Decyphering human factors knowledge for the evaluation and design of interfaces. Ergonomics of hybrid Automated Systems, II. Scapin, D. L. (1990b, February 2). Guidelines for user interface design: Knowledge collection and organisation. ITHACA.INRIA89.D12.03, INRIA. Scapin, D. L. (1990c). Organizing human factor knowledge for the evaluation and design of interfaces. International Journal of Human-Computer Interaction(2), 203-229. Scapin, D. L., Bastien, C., Leulier, C, Vanderdonckt, J., Mariage, C, Farenc, C., Palanque, P., & Bastide, R. (2000). Transferring knowledge of user interfaces guidelines to the Web. In Proc. of Int. Workshop on Tools for Working with Guidelines TFWWG '2000 (pp. 293-304). London: Springer-Verlag. Scapin, D. L., Garrigues, S., Farenc, C., Vanderdonckt, J., Palanque, E, Bastide, R., Bastien, J. M. C., & Leulier, C. (1999, December). Conception ergonomique d'interfaces web: demarche et outil logiciel de guidage et de support, Projet EvalWeb, Rapport d'avancement. Scapin, D. L., Leulier, C., Vanderdonckt, J., Mariage, C., Bastien, C., Farenc, C., Palanque, P., & Bastide, R. (2000). Towards automated testing of Web usability guidelines. In P. Kortum & E. Kunzinger (Eds.), Proceedings of 6th International Conference on Human Factors and the Web HFWeb'2000 [On-line]. Avail-
able: http://www.isys.ucl.ac.be/bchi/publications/2000/ScapinHFWeb2000.htm Smith, S. L., & Mosier, J. N. (1986). Guidelines for designing user interface software. Report ESD-TR-86-278, Bedford, MA: The MITRE Corporation. Available: http://www.syd.dit.csiro.au/hci/ guidelines/sam/guidelines.html;ftp://ftp.cis.ohio-state.edu/pub/hci/ Guidelines/ Stephanidis, C., & Akoumianakis, D. (1999). Accessibility guidelines and scope of formative HCI design input: Contrasting two perspectives, User Interfaces for All, Proceedings of the 5th ERCIM Workshop, 119-130. Steward, T, & Travis, D. (2002). Guidelines, standards and style guides. In J. A. Jacko & A. Sears (Eds.), The human-computer interaction handbook (pp. 991-1005). Mahwah, NJ: Lawrence Erlbaum Associates. Sun. (1999). Java look and feel design guidelines. Palo Alto, CA: Sun Microsystems, Inc. Available: http://java.sun.com/products/ jlf/edl/dg/index.htm van Welie, M., van der Veer, G. C., & Eliens, A. (2000). Patterns as tools for user interface design. International Workshop on Toolsfor Working with Guidelines, 313-324. Vanderheiden, G. C., Chisholm, W. A., Ewers, N., & Dunphy, S. M. (1997). Unified Web site accessibility guidelines [On-line]. Available: http://trace.wisc.edu/redirects/htmlgide/htmlgide.htm W3C. (1999). Web content accessibility guidelines 1.0 [On-line]. Available: http://www.w3.org/TR/WAI-WEBCONTENT Wang, C. H. (2001). A survey of design guidelines for usability of Web sites. Proceedings of HCI International, 1, 183-187. Webreview.com. (1998). The navigation and usability guide [On-line]. Available: http://www.webreview.com
AUTHOR INDEX AAAI, 40
Abascal, J., 240, 645 Abbott, S. G., 151 Aber, A., 519 Abeysekera, J. D. A., 293, 299 Abowd, G. D., 24, 30, 32, 34, 46, 149, 261, 535, 544 Achille, L. B., 95 Achorn, B., 230 Ackerman, E, 158 Ackerman, P. L., 428, 430 ActivMedia Research, 514 Adamopoulos, J., 577 Adams, A., 136 Adams, W. P., 126 Adar, E., 199 Adaval, R., 605 Adelson, M., 8 Adler,A., 330 Agrawal, V, 519, 521, 523 Ahumada, A.J., 122, 124 Ainsworth, L. K., 73-74, 303, 318, 392, 404 Ajzen, I., 575-577, 579-580, 582, 594 Akoumianakis, D., 241-243, 245-246, 252-254, 257, 537, 542, 544, 547, 688-689, 692 AlbaJ., 605 Albaladejo, S. A., 436 AlbersJ. W, 55 Albert, W, 118 AldortJ., 519 Alessi, A. A., 146 Alexander, D., 71 Alexander, M. W., 309 Alfaro, L, 121 Al-Gahtani, S. S., 332, 617 Allan, R., 327 Allen, B., 199 Allen, V G., 671 Alluisi, E. A., 5, 7 Altinsoy, M. E., 151 AmbiteJ. L, 214 Ambler, S., 386 Amelink, M., 412 American Heritage Editors, 515
American National Standard Institute (ANSI), 63, 65 American Psychological Association, 335, 485 Ames, A., 128 Anastasi, D., 460 Andersen, E., 458 Andersen, H., 329 Anderson, A. M., 94 Anderson, J. R., 78, 81, 92, 425, 428-430, 433, 670 Anderson, M. J., 10 Anderson, R., 552 Anderson, T, 199, 467 Andersson, B. J. G., 65 Andre, E., 225, 229-231, 233-234 Andre, A. D., 59, 200, 203 Andrews, E. J., 168 Anell, R., 59 AngiolilloJ., 205 Anick, P. G., 511 Annett, J., 390 Anonymous, 97 Anschau, K., 580-581 AntilJ. H., 599 Appelt, W., 553 Apple Computer, 41, 125, 128, 451 Archer, W, 467 archive.org, 40 Ardissono, L., 253 Arditi, A., 124 Argote, L., 530 Argyris, C, 529 Arial, M., 648 Arjona, L. D., 519, 521, 523 Armitage, C. J., 594 Armstrong, C., 392, 394 Armstrong, T. J., 55, 58 Arnault, L. Y, 59-60 Arnold, A., 241 Arnott,;. L, 436 Arnould, E., 608 Aroyo, L., 170 AshleyJ., 499 Assimacopoulos, A., 146, 152 AstleyJ. A., 404 Atwood, M. E., 327, 403 Aubel, A., 147 Austrian, G., 15 701
Ausubel, D. R, 158 Avant, D., 218 B
Borner, K., 223 Baber, C., 672 Babrys, G., 171 Bachtiger, M. T., 485 Bach-Y-Rita, P., 653 Back, K., 91 Badler, N. I., 229, 230 Baecker, R., 128, 554, 652 Bagozzi, R. P., 594, 597 Bailey, R. W., 118, 327, 332, 640 Baines, A., 544 Baird, L., 530 Baker, E., 161 Baker, G., 169 BakerJ. R., 112 Baker, P., 651,653 Baker, R., 112, 114, 127 Bakker, R., 205 Balabanovic, M., 253 Balakrishnan, H., 641 Balakrishnan, R., 652-653 Baldes, S., 225, 230-231 BalestraJ., 169 Ball, G., 228-230 BallardJ., 486 Bandilla, W., 488 BandosJ., 200,203 Banks, D. L., 286 Barab, S., 431 Barbacci, M., 245, 248 Barbesi, A., 108 Barfield,W., 651,653 Barish, G., 214 Barker, A., 392, 394 Barlow, C. B., 331 BaronJ., 309,478-488 Baroody, A.;., 169 Barret, R., 431 Barry, C. L, 194-195 Barry, D. T., 309 Bartels, B. H., 169 Bartlett, F. C., 4 BartlettJ. E., II, 309
702
AUTHOR INDEX
Bascones, J., 165 Basden, A., 697 Bashshur, R., 662 Bass, B., 519 Bass, L., 245, 248 Bastide, R., 690-692 Bastien, C, 690, 692 Bastien, J. M. C., 691, 693 Batagelj, Z., 311 BatemanJ., 198-199 Bates, M.J., 196 Batinic, B., 485, 488 BatteyJ., 514 Bauer, D., 121 Bauerschmidt, D. K., 10 Baumann, S., 231 Baumgartner, H., 596, 598, 601 Baxter, I., 332 Beale, R., 24, 30, 32, 34, 46, 86,149, 535, 544 Beard, C. M., 55 Beaton, R.J., 60 Beatty, S., 599-600 Beauregard, G. L, 151 Bechhofer, S., 179 Becket, T., 230 Beebe, S. A., 91 Beer, M., 588 Beier, B., 502, 508, 694 Beirekdar, A., 697 Belk, R. W., 601 Belkin, N.J., 195-196 Bell, B., 94-95, 97, 99 Bell, G., 312-313 Bellotti, V, 554 Bellovin, S., 622 Benbasat, I., 119 Benenfeld, A., 194 Beneventano, D., 218-219 Benford, S., 225 Bennett, D., 651 Bennett, J. L, 324 BenseJ., 169 Bensmaya, S. J., 145 Bentler, E M., 575-576 Bentley, R., 553 Benton Foundation, 664 BenwayJ. E, 118, 516 Benyon, D. R., 252 Benysh, D. V, 75, 80-81, 83, 85 Berbaum, K. S., 654 Bergamaschi, S., 212, 218-219 Bergamasco, M., 146 Berge, Z. L., 463 Bergel, M., 510 Bergenti, E, 223 Bergqvist, U. O., 52 Berker, T., 487 Berkun, S., 203, 205 Bernard, M., 108, 110-112, 114-117, 121,126-127, 205, 516 Berners-Lee, X, 157, 693, 698 Bernstein, D. A., 311
Bernstein, T., 613 Berris, M., 146 Berry, D. G., 161 Bersen, N. O., 511 Berthold, M. R., 86 Bertram, C., 218 Besser, H., 467 Bessiere, K., 128 Best, S.J., 312 Bevan, N., 241, 252, 257 Beyer, H., 242, 303, 313, 318, 329, 331, 343, 357-358, 371, 670 Bhatnagar, A., 596, 601-602 Bhatti, N., 108 Bhimini, A., 613 Bias, R. G., 324, 356, 679-680 Bichsel, M., 648-649 Bickmore, T., 230, 234 BidarraJ., 165, 169 Biggs, S.J., 652 Billinghurst, M., 647, 652 Billsus, D., 430 Bini, A., 254 Biocca, F., 598 Birebent, G., 652-653 Birmingham, H. E, 5 Birmingham, W, 202 Birnbaum, M. H., 309, 471-476, 478-479, 483-484, 486-489 Birtley L., 330 Bishu, R. R., 126 BizRate, 514 Black, J., 681 Blackwell, R. D., 604, 608-609 Blade, R. A., 647 Blake, D. T., 144, 146 Blattner, M., 142 BlauertJ., 136, 138-139, 148, 151-152, 652 Bleeker, M. Q., 55 Blendon, R., 667 Blickensderfer, E. L., 91 Bloch, F. H., 599-600 Blocher, A., 224, 231 BlombergJ., 670 Bluedorn, A. C., 286 Blumberg, B., 228, 230 Blythe, M., 45, 462 Bodden, M., 136, 141, 148 Boehm, B. W, 245, 680 Boehm-Davis, D. A., 402-403 Boettcher, J. V., 446, 451 Bohan, M., 118 Bohlman, M., 55 Bohme, F. G., 15 Boies, M. R., 125 Boies, S.J., 125 Bolanowski, S. J. Jr., 145 Bolas, M., 651 Bond, N. A., Jr., 7 Boose, J., 72-73 Booske, B. C., 658-659, 661-664, 666-667, 672-673
Booth, P., 386 BorchersJ. O., 688 Borella, M. S., 535 Borg, I., 421 BorgesJ. A., 693 BosertJ. L, 680 Bosnian, D., 61 Bosnjak, M., 311, 471, 485, 488 Bothell, D., 425 Bouch, A., 108 Bovair, S., 78 Bowdish, B., 431 Bower, G. H., 81 Bowers, C. A., 91,96, 161 Bowers, J., 311 Bowker, D. K., 311-312, 475 Bowman, D., 649, 651-652 Boyarski, D., 121 Boyd, C., 309 Boyle, C. E, 433 BoyntonJ. L, 121-122 Broder, A., 485 Bradford, J. H., 325, 333, 672 Brannon, N., 81 Bray, C. W., 4 Breaux, R., 64, 650-653 BreeseJ., 228-229, 252 BrennanJ. D., 511 Brent, E., 436 Breuel, T., 199 Brewer, E., 201, 203 BrewerJ., 241,252, 257 Brewster, S. A., 135-136, 142-143, 146, 151-152 Bricklin, D., 110 Briggs, G., 165 Brigham, T., 169 Brin, S., 201-202 Brinck, T., 341-343, 346, 349, 353, 391 Brink, T., 321, 325-326 Broadbase., 523 Broadbent, D. E., 56 Brodbeck, E C., 329 Brogmis, G., 59 Bronfenbrenner, U., 632 Brookes, B., 195 Brooks, F. P., Jr., 61,651 Brown, B. A. T., 330 Brown, C. E., 166-168 Brown, C. H., 121 Brown, E. L., 9 Brown, S., 280 Bruce, H. W., 195 Bruns, T., 200 Brusilovsky, P. 252-253, 434, 436 Bruza, P., 199-200 Bryant, E., 119 Buchanan, G., 643-644 Buchanan, T., 488 Buck, L., 59, 75, 80-81, 83, 85 Buckle, E, 52, 59-60 Buckley, C., 195 Buie, E., 688
AUTHOR INDEX
BuUder.com, 693, 698 Bull, S., 434 BuUinger, H.J., 544 Bumatay, M., 25 Burastero, S., 58 Burdea, G. C., 146-147, 652-653 Burgess-Whitman, N., 330 BurkellJ., 199 Burley Gant, D., 537, 542, 544, 547 Burmeister, O. K., 296 Burns, C.M., 93 Burrell,J.,648 Burrell, M., 670 Burton, A. M., 168, 171 Burton, M., 331 Burzagli, L., 243 Busbach, U., 553 Bush, B., 659 Bush, V, 118 BusinessWeek, 71 Bussemakers, M. P., 136, 140, 143 Butler, S., 690 Buxton, H. W, 15 Buxton, W, 58, 60, 654 Buzan, B., 158-159, 166
Buzan, T., 158-159, 166 Byrd, D., 198, 202 Byrne, M. D., 323, 425, 428, 670
Cakir, A., 55, 58 Canas, A.J., 161, 164-165, 167-169 CacioppoJ., 599-600 Cadotte, E. R., 609 Cai, D. C., 292, 299 Gail, E, 55 Cain, M. M., 660, 662-663 Calantone, R., 582-583 Calderwood, R., 74-75 Caldwell, B. S., 289, 299 Caldwell, D. G., 146 Callaway, C., 228, 232 CameronJ. A., 332-333 Campbell, A. J., 121 Campbell, C., 431 Campbell, D. T., 318 Campbell, G., 168 Campbell, J. I., 146 Campbell, M. C., 603 Campbell, R., 327 Campbell-Kelly, M., 15 Campus Computing Project, 441, 446, 451 Candland, K. M., 509 Cangelosi, V E., 529, 543 Cannon-BowersJ. A., 91, 93-94, 96-97 Capin, T., 147 Carayon, P., 51, 330, 334 Card, S. K., 59, 74, 78-79, 81, 83, 213, 323, 389, 393, 396-398, 402-403, 426, 664, 670
Caret, W. L., 7 Carello, C., 92-93 Carey, M. S., 404 Carey, T. T., 333 Carff, R., 165, 168-169 Carhart, R. R., 7 Carion, S., 147 Carlson, P., 531, 534, 544, 546 Carmel, E., 197 Carnot, M.;., 161, 165, 168-169, 171 Caroflglio, V, 228, 230 Carpenter, B. E., 15 Carpenter, R A., 642 Carr, C. L, 581, 583 Carrithers, C., 433 Carroll,;. M., 33, 324, 358, 386, 433 Carron, A. V, 579 Carson, S., 581, 583 Carter,;. P, 651 Carter, L., 119-120 Carter, R. L., 123 Carton, L., 117 Case,;., 663-664, 671 Cass, T., 199 Cassell, EJ.,667 CassellJ., 230, 234 Castano, S., 218-219 Castillo,;. C., 333, 394-395, 510 Catledge, L., 39, 197 Caudell, T., 150 Cavalier, G. M., 666-667 Cavalli-Sforza, V, 171 Cavonius, C. R., 121 Ceaparu, I., 128 Celsi, R. L., 599-600 Centers for Disease Control and Prevention, 667 Ceruzzi, P. E., 15 Chafnn, D. B., 55 Chaiken, S., 594 Chakrabarti, A., 332 Chakraborty, A., 641 Chambel, T., 165, 169 Chamis, Y, 196 Chandrasekar, R., 199, 205 Chang, C-H., 669 Chang, H., 641 Chang, W-T., 169 Chapanis, A., 5, 8, 333, 382 Chaparro, B. S., 116, 516, 517 Chapelle, B. D. L., 651 Charlton, C., 553 Charron, C., 519 Chatterjee, P., 607 Chau, P., 580-581, 583 Cheak, A., 161 Checkosky, C. M., 145 Chen, A., 583-584 Chen, B., 205 Chen, C., 40, 223 Chen, H., 127, 197, 200 ChenJ. L., 650, 651 Chen,;., 290, 299, 544
703
Chen, L.-D., 581 Chen, M. T., 292, 299 Chen, Q., 607 Cheng, B., 146 Chernev, A., 608 Cherry, E. C., 135, 138 Cherry,;., 658 Cheswick, B., 622 Chi, C. E, 292, 299 Chi, D., 229 Chia, Y. T. B., 294, 300 Chiang, K., 596-597 Chignell, M., 202 Childers, T. L., 581, 583 ChinJ. R, 332 Chin, W W, 580, 583-585 Chinn, C. A., 169 Chion, M., 136 Chipman, S. E, 74-75, 389, 396 Chisholm, W, 269 Chiu, C.-H., 169 Cho, B., 597 Choi, I., 284 Chong, R. S., 425 Choong, Y. Y, 285, 294-296, 299-300 Christ, R. E., 123 ChuangJ. H., 292, 299 Chung, G. K. W. K., 169 Chung, G., 161 Chung, L., 245, 248 Cianchette, C., 110-111, 510 Cimino,;.;., 669, 671 Citera, M., 166-168 Civit, 645 Clancey, WJ., 92-93 Clark, A., 15 Clark, H., 39 Clark,;., 35 Clarke, L., 324 Clavijo, I. E., 165 Cleary, A. G., 61 Clements, P., 245, 248 Clementson, G. C., 7 Clippingdale, S., 276 Clore, G. L., 228 CNN.com, 26 Cockburn, A., 431 Cockton, G., 324, 672, 689 Code, S., 91 Coen, M., 641 CoffeyJ. W., 161, 164-165,168-169, 171 Cofino, T., 496 Cohen, B. G. E, 52, 55 Cohen,;., 514 Cohen, M., 652 Cohen, N., 641 Cohen, W.;., 58, 63 Coiffet, P., 146, 652-653 Cole, L. L., 58 Coleman, D., 544 CollJ., 119-120 Coll, R., 119-120 Collier, G., 228
704
AUTHOR INDEX
Collins, A., 228 Collins, H. M., 539-540, 544 Colwell, C, 146 Comber, T., 693 Comer, D., 14 Commonwealth of Australia, 523 Conner, C. E., 144 Johnson, K. O., 144 Compeau, D., 585 Competitor analysis, 456 Conati, C., 433 Conner, M., 594 Conner-Sax, K., 14 Connor, C. E., 144 Connor, M. B., 59 ConradtJ., 311 Constantine, L. L., 357-358, 462 Consumer Reports, 518 Contreras, P, 214 Converse,;., 310, 312 Converse, S. A., 91 Conway, E, 58 Conway, M. J., 654 Cook, P. R., 151-152 Cooke, A. D., 607 Cooke, N. J., 74, 91, 93-97, 99, 396 Cool, C., 195 Cooley, R., 199 Cooper, A., 315,322-323 CooperJ., 392, 394 Cooper, M. B., 141 Cooper, M. D., 393 Cooper, W, 194 Cooperstein, D. M., 519, 524 Corbett, M., 332 Cortell, E., 138 Costa, M., 229 Cothey, V, 199 Coughlin, E. C., 464 Couper, M. P., 309-312 Courtney, A. J., 294-295, 300 CoveJ., 197 Cowan, S., 75 Cox, J., 305 Cox, R, 614 Coyne, K., 505 Coyne, R. D., 86 Crabtree, A., 32 Craig, S. D., 225 Craik, K.J.W., 4 Crandall, B., 73, 168 Crawford, S.D., 309-311 Crawford, S., 197 Crawford, W., 36 Crease, M. G., 136 Cristea, A., 165 Crockett, R. O., 389 Croft, W. B., 198, 202 Cross, R., 530 Crow, D. C., 323 Crutchfield,;., 99 Csikszentmihalyi, M., 462 CTA Resource Web site, 385
Cunningham,;., 305 Curran, P. S., 622 Curry, A., 542, 544, 546 Curry, R. E., 672 Cushman, W H., 121 CyberAtlas, 70 Cyert, R. M., 529 Cytryn, K. N., 659 Czerwinski, M., 39-40, 108, 303
Diirrer, B., 139-140 Daft, R. L., 97, 529 Dai, P., 292, 299 Daily, L. Z., 425 Dainoff, M.;., 52, 55 Dalai, R. S., 309, 471 Dale, R., 647 Damos, D. L., 83 Damsgaard,;., 538, 540, 542-545 Danielsen, D., 603 Danis, E., 651 Dansereau, D., 169 Danzico, L., 203 Darby,;., 449 Dardailler, D., 241, 280 Darian-Smith, L, 145 Darken, R. P., 650-651 Dasan, V, 536 DATech, 358 Daughterly, T., 598 Davenport, L., 35 Davidson, D., 195 Davies, R. C., 649-650 Davies, S. P., 92, 97 Davis, E D., 575, 577-578, 580, 583, 594 Davis, H. C., 648 DavisJ. H., 92 Davis, K., 667 Davis, M. D., 15 Davis, M., 169 Dayton, T., 331 De Antonellis, V, 219 De Capitani Di Vimercati, S., 219 De Carolis, B., 228, 230 deHaan, A., 136, 140, 143 De Haan, G., 79, 80 de;uan, M. D., 289, 294-295, 299-300 De La RosaJ. L., 212 de Rosis, E, 228, 230 Deakin University Equity Access Research and Development Group, 252 DeAngelo, T., 110, 112, 116, 118, 126-127, 341, 349, 353 Deatherage, B. H., 136 Decker,;.;., 121 Degen, H., 509, 617 DeGroot, A., 289 Deisigner,;., 653-654 Deitel, H. M., 14
del Galdo, E. M., 392-396 Delbecq, A. L., 331 Delemos, K. A., 145 Delhagen, K., 519 Delia Mea, V, 661 Dennerlein,;. T., 146, 152 Dennis, A., 551, 554 Dennis, S., 199-200 Department of the Taoiseach, 523 Deravi, E, 243 Derbyshire,;., 201 Derjani-BayehA., 63 Dern, D. P., 14 DeRouvray, C., 311-312 Derry, S.;., 433 Designing Web-based training, 461 Desurvire, H. W, 327 Detweiler, M. C., 692 DeyA. K., 57,261 Dhamija, R., 511 Dholakia, R. R., 596-597 Diaper, D., 33, 74, 404 Dicheva, D., 170 Dickinson, T. L., 91 Dickson, M. W, 552 Diehl, V A., 332 Dieterich, H., 252 DiFranco, D. E., 151 Dill, W. R., 529, 543 Dillman, D., 310-312, 475 Dionisio,;., 146 Distance education at a glance, Guide #4, Evaluation for distance educators, 467 Dittman, P E., 5 Divett, M., 580-581, 583, 590 Dix, A., 24, 30, 32, 34, 38, 40, 43, 45-46, 149, 535, 544, 552-554 DiZio, P., 653-654 DMOZ, 39 Doan, K., 200 Dobbie, A. E., 311-312 Dobson, M. W, 436 Dodge, B., 442 Dodson, D., 168 Dold, C., 658 Donahue, G. M., 380-381 Donahue, S., 467 Dong,;., 293-294, 299 Donoghue, K., 496 Dooyeweerd, H., 697 Dorsey, D. W, 168 Dougherty, D.;., 9 Doughty,;. M., 7 Douglas, S. A., 61 Douglas, S. D., 58 Dourish, P., 554 Douville, B., 230 Dowling, G. R., 602-603 Downes, S., 451 Downey, L., 502 Doyle, D.;., 93 Drake, D., 110, 117, 121, 667
AUTHOR INDEX
Drake, P. J. W., 487 Draper, S., 460, 670 Dray, S., 318, 330 DrennanJ., 582-583 Dreyfus, H. L., 416 Driscoll, M., 456 Drott, M. C., 333 Drucker, P. F., 531 Drucker, S. M., 234 Duane,A., 542, 543-546 Dubach, E. M., 535 DubbeyJ. M., 15 DuchnickyJ. L., 121 Duchnicky, R. L., 121 Duffy, V G., 537, 542, 544-545, 547 Dumais, S., 127 Dumas, J. S., 324-325, 345 Duncan, J. C., 167 Duncan, K. D., 390 Dunn, B., 169 Dunne, P. E., 14 Dunphy S., 269 Dupper, T., 533-534, 539-540, 541,
544-545 Durlach, N. L, 650 Durso, F. T., 99 DyckJ. L, 332 Dye, C, 121 DyerJ., 168, Dyson, M. C., 112, 121 Dziadosz, S., 199, 205
E.Grandjean, 51 Eagly, A., 594 Eastlick, M. A., 580, 582-583, 585 Eastman, C., 607 Eberle, D., 169 Eberman, B., 146 Eberts, R. E., 72-73, 76, 79, 86, 650 Eckman-Orderberg, G., 54 EconoNet, 252 Eden, C., 158, 166 Edmonds, A., 199 Edwards, A. D. N., 142-143 Edwards, P., 212 Efthimiadis, E., 195 Eggemeier, F. T., 83 eHealth Ethics Initiative, 668 Eichhorn, K., 313 EichstaedtJ., 479 Eisenberg, M., 194-195 Eisenstat, R., 588 Ekhaml, L., 463 ElamJ. L., 577 Elfstrom, G., 65 Eliens, A., 688 Elias, R., 55 Ellen, P. S., 594 Elliot, K., 504 Elliott, C., 228 Ellis, D. G, 90-92
EllisJ., 92-93 Ellis, S., 653-654 Ellison, E., 653-654 Ellison,;., 537, 542, 544, 547 Ellison, M., 109, 127 Ellison, R., 248 Elmagarmid, A. , 641 ElmanJ. L, 86 El-Nasr, M. S., 228, 230 El-Shinnawy, M., 97 Elson, S. A., 643 EltingeJ., 310 eMarketer, 514, 518-519 Embrey, D. E., 672 Emery, L., 642 Emery, V K., 537, 542, 544, 547 Emiliani, P. L., 240, 241, 243, 252, 254, 257, 263 Endsley, M. R., 99 EngelJ.E, 604, 608-609 English,;. D., 59, 126 Englmeier, K-H., 652, 654 Enthoven, A. C., 662 ePaynews, 514 Ericsson, K. A., 72, 94-95, 168 ErnoJ., 121 Esiobu, G, 166 Esposito, C., 651 Essa, L, 436 eTranslate, 524 Etzioni, O., 202 European Union, 523 Evanoff, B., 55 Evans, A. W., Ill, 161 Everitt, B. S., 94 Evers, V, 296 Evett, M., 187 Ewers, N., 269 Eysenbach G., 661
Fabrizio, R., 121 Fain, W. B., 524 Fairhurst, M., 243 Fallman, D., 413 Fang, N., 602-603 Fang, X., 193, 195-196, 198-200, 300, 507, 617 Faraday, P., 118, Farenc, C., 690-691 Farina, N., 648 FarleyJ., 36 Farnham, S. D., 234 Farquhar, A., Farrant, S., 643-644 Farrell, S., 127 Fasti, H., 137, 141 Faulring, A., 199 Fausett, L., 85 Fazio, R. H., 576, 577 FBI, 613 Federico, R-A., 431, 433
705
Feigenbaum, B. A., 524 Feiner, S., 645 Feldman, E., 330 Feliciano, G., 119 Fellenstein, C., 574 Feltovich, P., 161, 165 Fenton, R., 392, 394 Ferandez, M., 205 Ferguson, D. C., 121 Ferguson, T., 663-664 Fernandes, T., 287-288 Fernandez, A., 601-602, 604 Fernandez, M., 112, 121, 127 Ferrell, W. R., 389 Festinger, L., 91 Fichman, M., 88 Fidel, R., 196 Fiegel, T., 327 Filipic, B., 86 Filippou, T., 141 Finin, P., 641 Finin, T., 231 FinkJ., 254, 430, 436 Finkelstein, L., 199, 204 Finlay,;., 24, 30, 32, 34, 45-46, 86, 149, 535, 544 Finn, R., 121 Finnegan, P., 542-546 Fischer, G, 408, 413-415, 424, 430 Fishbein, M., 575-577, 579-580, 594 Fisher, A. B., 90-92 Fisher, A., 40 Fisher, C., 97 Fisher, D., 511 Fisher, K. M., 158 Fiske, D. W, 318 Fitts, P. M., 5, 7, 8, 31, 385, 393-394 Fitzgerald, L., 582-583 FitzGerald, P.;., 228, 232 Fitzgerald, S., 667 Fitzgerald, T., 681 Fitzgerald, W., 62 Fjeld, M., 648-649 Fliickiger, B., 140, 142-143 FlachJ. M., 76, 412, 509, 617 Flanagan, D., 36 Flanagan,;. C., 4-5, 392-394 Fleischman, S., 121 Fleming,;., 322, 517 Flexman, R. E., 4 Flieschman, S., 510 Flood, M. M., 7 Florentine, M., 141 Floyd, R., 641 Flynn, R. A., 511 Fogg, B.;., 602-603 Fogleman, G, 653-654 Fogleman, M., 59 Folds, D.;., 524 FoleyJ., 252 Foltz, P. W., 91,99, 396 Fontaine, G., 653 Ford, B., 199
706
•
AUTHOR INDEX
Ford, K. M., 164-168, 171 Forlizzi, J., 121, 308, 313, 315 Forsythe, C., 689-690, 692-693 Foskett, D., 194 Foster, L, 168 Fowler, F. J., Jr., 309-310, 312 Fowler, S. L, 122 Fox, J., 688 Fox, R.J., 596, 599 Fox, S., 664 Foxall, G., 594 Francis, G., 479 Francisco, J. S., 166 Franek, L., 488 Frank, X, 121 Franklin, S., 435 Franzblau, A., 55 Franzke, M., 325, 672 Freeman, L. C., 511 Frei, R. L., 332 Freitag,D., 195 Freivalds, A., 74 French, T., 287 Frese, M., Peters, H., 329 Frick, A., 485 Fried, B. M., 666-667 Friedkin, N. E., 91 Friedman, B., 602 Fritsch, D., 648 Fritzson, R., 231 Frommer, W. D., 123 Fu, L., 293 Fu, W. T., 428 FucceUaJ., 515 Fuchs, F., 86 Fukuoka, W., 287, 296 Fulk, J., 97 Furner, S., 146 Furness, T. A., 62
GabbardJ.L., 651-654 Gabrilovich, E., 199, 204 Gagne, R., 461 Gaines, B., 165, 171 Galetsas, A., 241, 252, 257 Galinsky, T. L., 58 Galletta, D. E, 582-584, 589 Gallimore,;., 653-654 GalushaJ. M., 464, 467 Garber, A. M., 662 Garcia, E. V, 436 Gardiner, M. M., 688 Gardner, H., 464 Garg, C., 294-296, 299-300 Garman, R. A., 7 Garner,W. R., 5 Garrigues, S., 691 Garrison, D. R., 467 Garzon, M.H., 225 Gaver, W.W., 140-141 Gay, G. K., 648
Gebhard, P., 230 Gefen, D., 602-603 Gemperle, E, 56 General Services Administration, 523 Gentry, L., 582-583 GeorgeJ. E, 93, 243 Georgia Tech GVU Center, 193, 201, 203 Gerard, M.J., 58 Gergle, D., 321, 325-326, 341-343, 346, 349, 353, 391 Gershoff, A. D., 607 Gertner, A., 433 Gescheider, G. A., 145 Geusebroek, J. M., 432 Ghani,;., 119 Ghiringhelli, L., 55 Gholson, B., 225 Gibbs, W. W., 462 Gibson, F. P, 88 Gibson, J. J., 93, 408 Gierere, R., 58 Gilbert, S., 467 Gilbreth, F. B., 385, 387 GUI, R. T., 164, 166 Gill, Z., 537, 542, 544, 547 Gillenson, M. L., 581 Gillespie,;., 660 Gillett, R., 92 Gilmore, D.J.,92, 97 Giorgianni, S. J., 660, 663-664 Girard, T., 601, 602, 607-608 Gitlin, R. D., 641 Gladwell, M., 455 Glass, J. R., 62 Glasser, R C., 60 Global Reach, 524 Goble, A. K., 145 GobleJ. C, 651 Golbeck, J., 178, 184-186 Goldberg, A., 229 Gomes, M. E., 167 Gomex, D., 652-653 Gonzales, E, 436 Gonzalez,;. S., 533, 535, 540-541, 544-545 Good, R., 158 Goodrum, A., 200 Goodstein, L. P., 411, 413 Goodstein, R. C., 603 Goodwin, A. W., 146 Goonetilleke, R. S., 292-294, 299-300 Goransson, B., 388 Gordon, T. L., 166, 169 Gordon, S. E., 164, 166 Gordon, S., 73 Goritz, A. 480-481,488 Gorman,;., 99 Gosney, C., 146 Gott, S. P., 387, 396, 509, 617 Gottlieb,;., 602-603 Gould, E. M., 287-288, 295, 300 Gould, E. W., 287 GouldJ. D., 121, 125
Gowin, D. B., 158, 162, 165-166 Goy, A., 253 Grace, G., 119 Graeber, D. A., 651-654 Graesser, A. C., 164, 435 Graetz, K. A., 331 Graf.W., 121 Graham, D., 536 Gralla, P, 14 GranaJ., 660, 663-664 Grandjean, E., 52, 55, 58, 63-64 Graphics, Visualization, and Usability Center, 78, 515 GratchJ., 228 Gray, M., 201 Gray, P., 136, 151 Gray, W. D., 402-403 Green, M., 651 Green, S., 647 Green, T. R. G., 78, 92, 97 Greenberg, R., 142 Greenberg, S., 39-40, 197, 554 Greene, D. P., 86 Greene, S., 125, 200 Greenes, R. A., 671 Greenhalgh, C., 225 GreenoJ. G., 93 Greenstein,;. S., 59-60 Greer, X, 535-536, 544-545 Gregory, M., 121 Gregov, A., 392 Greif, L, 332 Grewal, D., 602-603 Grice, M., 231 Grier, R. A., 121 Griffin, M.;., 148 Grimm,;., 517 Grimm, W., 517 Grinter, R. E., Groenen, RT. E, 421 Grose, E., 689-690, 692-693 Grossberg, S., 86 Group,E,394-395 Grove, M., 184-186 Groves, R., 310 GrudinJ., 553,671 Grunert, K., 604 Guastello, D. D., 93 Guastello, S.;., 93 Guengerich, S., 536 Guernsey, L., 518 Guest, G., 670 Guest, R., 243 Guimaraes, N., 165, 169 Gujar, A., 330 Gunderson, A., 518, 523 GundersonJ., 280 Gunnarsson, E., 55 Gupta,;. N. D., 546 Guski, R., 136, 148 GutermanJ., 202, 206 Gutwin, C., 554 Guzzo, R., 552
AUTHOR INDEX
Hager-Ross, C., 146 Hagg, G. M., 146 Hiinting,W, 55, 63 Ha, H. Y, 601,606-607 Haala, N., 648 Haas, M., 81, 86 Haase, R. E, 286 Haataja, S., 252, 257 Haataya, S., 241 HaberJ., 35 Hackler, T., 121, 127 Hackman, G., 498-499 Hackos, J. T., 303-304, 313, 315, 343, 357-358 Hackworth, C. A., 99 Haddadin, I., 200 Haddow, G., 461 Hagberg, M., 51, 59 Hagen, E R., 518 Hahn, H., 14 Hajdukiewicz, J. R., 93 Hajnal, C., 330 Hall, A., 305 Hall, B., 461 Hall, E. T., 284-286, 295-296, 300 Hall, M., 284-286 Hall, R. H., 166, 169 Halliday, P., 392 Hamblin, C., 110-111 Hameed, A., 167 Hamlin, L. S., 451 Hammond, B., 218 Hammond, J., 76 Hammond, N., 169 Hammond, W. E., 669 Han, X., 292, 299 Hanes, L. E, 168 Hanesian, H., 158 Hanley, G. L., 332 Hannaford, B., 653-654 Hansell, S., 514 Hansen, E., 280 Hanssen, D. R., 511 Hapeshi, K., 652 Happ,A.J., 52, 58 Harada, A., 288, 296 Harbison-Briggs, K., 73 Hardin, 475 Hardman, L., 432, 691 Hardwick, A., 146 Hare, A. P., 91 Harel, D., 287, 296 Hari, R., 151 Harmon, J., 97 Harper, A., 658 Harper, M. E., 161 Harper, R. H. R., 94, 97 Harper, R., 330, 334 Harpoid, T., 643 Harris Interactive, 516 Harris, D. H., 10
Harris,;., 145,442,663 Harris, R. L., 120 Harris, R. W, 332 Harrison, B. L., 330 Harrison, T., 607 Harrsch, M., 451 Hart, D.J., 55 Harter, S., 194
Hartson, H. R., 333, 345, 393-394, 510 HartwickJ., 594 Harvey, C. E, 434 Harvey, C. M., 75-76, 80-81, 83, 85 Harvey, M. G., 534, 542-545 Haselgrove, M., 121 Hashiya, Y, 108 Haslam, R., 52, 59-60 Hasler, S. G., 9 Hassenzahl, M., 462 Hastings, S., 52, 59-60 Hauben, M., 14 Hauben, R., 14 Haubner, M., 652, 654 Haumer, P., 242 Haupt, B., 121 Hausenblas, H. A., 579 Hawkes, L. W, 433 Hayes, P. J., 160, 165-166 Hazel, H., 544 Head, A. J., 689 Hearst, M. A., 126, 511 Heckerman, D., 252 Hedberg, B., 529 Hedge, A., 59 Hedstrom, L. D., 52 Heerwegh, D., 312 HeflinJ., 177 Heidingsfelder, M., 485 Helal,J.,64l Helander, M. G., 51 HelmE., 91,97,99 Kelson, H., 578 Hempel, T., 136, 139, 141 Henderson, R., 580-581, 583-584, 587 Hendler, J., 178-179, 184-187, 435 Hendrick, H., 51 Henze, N., 253 Herbert, L. B., 327 Herl, H. E., 169 Herman, L., 296 Herrmann, D., 509, 617 Hersh, H., 121, 122 HersheyJ. C, 309 Hertz-Lazorowitz, R., 169 Hess, R., 124 Hettinger, L., 653-654 Hewson,C.M., 311-312 HFES/ANSI, 200, 693 Hicinbothom, J. H., 75 Hieb, M. R., 435 Higgins, C. A., 585 Hilbert, D. M., 643 Hill, A., 121, 122,520-521 Hill, R., 60
•
707
Hill, G., 169 Hilliard, M. R., 86 Hills, M., 530, 532, 535-536 Hinckley, K., 651 Hincley, K., 57, 59 Hinrichs, E., 553 Hinsz, V B., 92 Hinton, G. E., 85 Hirose, M., 653 Hirsch, L., 523 Hirschman, E. C., 596, 598-600 Hishiyama, R., 232 Hix, D., 333, 345, 394, 651-654 Hobbes, N., 392 Hoch, S.J., 605-606 Hodgins, H. W., 449 Hoeft, R. M., 161 Hoege, H., 141 Hofacker, C. E, 605 Hoffman, D. L., 609 Hoffman, L. R., 97 Hoffman, R. R. 161, 164-165, 168, 171 Hoffman, R., 480-481 Hofstede, G., 288, 295-296 Hoft, N. L. 338, 341, 349, 353 Hoglund, D. H., 663 Holbrook, M. B., 596, 598-600 Holdaway, K., 688 Holguin, C., 152 Hollan,;. D., 57 Hollier, M. R, 150 Hollingshead, A. B., 91 Hollins, M., 145 Hollnagel, E., 74, 422, 672 Holmes, E., 147 Holtzblatt, K., 242, 303, 313, 318, 329, 331, 343, 357, 358, 670 Honeycutt, J. C., 14 Hong, J.-K., 320 Hoogstraten,J., 594 Hope, B., 539 Hopkins, C. O., 10 Horbach, U., 138 Horn, R. E., 463 Horney, M., 431 Hornof, A.J.,427 Horrocks, L, 179 Horton, M. S., 97 Horton, S., 34, 118, 124-126, 512, 688, 691-692, 697 Horton, W., 293-294, 299, 338, 341, 349, 353 Horvitz, E., 205, 252 Housel, B. C., 641 Hovel, D., 252 Howard,;. A., 609 Howe, D. C., 602 Howell, W. C., 6, 74 Howerton, B., 573 Howes, A., 46, 428 Hsiao, S. S., 144, 146 Hsieh-YeeJ., 196
708
AUTHOR INDEX
http://www.library.cornell.edu/design/ dsigndoc.html, 688 Hu, P., 580-581, 583 Hu, X., 225 Huang, C.-C, 169 Hubbard, C, 312 Huber, G. P., 529-531, 540, 543, 547 Huey, B. ML, 385 Huff, S., 585 Huggins, W. H., 7 Hughes, J. A., 32. 308, 312, 317, 329 Hull, S., 110, 117, 121, 127 Hunkins, F. P., 466 Hunting, W, 58 Hupprich, L., 25 Hurson, A., 200 Hussam, A., 199 Hutchins, E., 57, 92-93, 412, 414 Hutchinson, W., 605 Hutton, R.J. G., 75 Hyde.J., 199 Hynes, C, 112
lacucci, G., 242 lakovidis, I., 241, 252, 257 Ifukube, T., 62 Igbaria, M., 332, 580, 583 Iglseder, H., 141 Ignacio, A., 338, 341, 349, 353 Ihde, S., 61 IHETS, 457 Ikar, D., 123 Ilioudis, C., G. 615 Ingram, D., 459 Ingwersen, P., 196 Ino, S., 62 Insko, C. A., 599 Institute of Medicine, 658-659, 661, 664, 672 Intelligent enterprise, 573 International Business Machines (IBM), 124, 252, 303, 349, 398, 353, 693 International Standards Organization (ISO), 245, 303, 523, 688-689, 693 InternetWeek, 523 Inxight Software, 40 loerger, T. R, 228, 230 Irani, T., 467 Irby, C., 388 IrelandJ., 330 Isaacs, E. A., 331-332 Isbister, K., 234 Isensee, S., 358, 688 Ishida, T., 225, 227, 232, 234 Ishii, H., 652 Ishizuka, M., 229-230 ISO 13407, 135 ISO 14915-3, 135, 149-150 ISO 9241-11, 135 ISO/IEC 13407, 358, 360 ISO/IEC 18529, 358
ISO/IEC 9241-10, 360 ISO/IEC 9241-11, 360, 362 Isono, H., 276 ITU-T Telecommunication Standardization Sector of ITU, 293 Iwata, H., 62 Izumi, T, 62 I
Jegou, G., 39 Jacko, J. A., 32, 51, 241, 252, 535, 537, 542, 544, 546-547, 658-659, 661-664, 666-667, 670-673 Jacob, R., 651 Jacobs, I., 280 Jacobsen, S., 653 Jacques, P. E, 412 Jacques, R., 333 Jacquette, D., 15 Jadad, A. R., 660 Jaffe, M., 667 James, C. A., 58, 63 James, W., 411 Jameson, A., 432 Jamieson, B., 193, 199 Janes, J., 194-195 Jansen, B. J., 126, 198-199, 200, 203 Jansson, G., 147 Janzen, K., 121, 127 Jarke, M., 242 Jarvenpaa, S. L., 582-583, 602-603 Jeffries, R., 94, 324-325, 327, 672, 689 Jekosch, U., 139, 148 Jellinek, H. D., 59 Jenkins, P., 241, 252, 257 Jenkins, R. L., 609 Jenkins W., 59 Jentsch, E, 161 Jessup, L. M., 93 Ji, Y. G., 537, 542, 544, 547 Jin,W., 292, 299 Jing,J.64l Jinks, T S., 169 Joachims, T, 195 Johannsen, G., 86, 146 Johansson, L., 146 Johansson, R. S., John, B. E., 323, 397, 402-403, 670 Johnsgard, T.J., 125-126 Johnson, B. E., 5 Johnson, E.J., 403, 605 Johnson, J., 71, 341, 346, 349, 353 Johnson, K. O., 144, 146 Johnson, R., 195 Johnson, S., 455 Johnson, T. J., 309, 311-312 Johnson, W. L., 225, 227, 233 Jokela, T, 358-359 Jonassen, D., 169 Jones, A. M., 544 Jones, C. S., 55
Jones, D., 652 Jones, G., 165 Jones, J., 167 Jones, L., 144, 146 Jones, M., 643-644 Jones, P. C., 659 Jones, P. M., 537, 542, 544, 547 Jones, R. E., 385, 393-394 Jones, S. K., 58 Jones, S., 431, 511, 514, 680 Jordan, B., 683 Jose, J., 205 Joseph, K. M., 121 Jouse, 2, 61 Jousmaki, V, 151 Jung, B., 231 Just, M. A., 642 K
Kiihrne, T, 252 Kiihne, A., 121 Kaarstad, M., 672 Kaczmarek, K. A., 653 Kahn, P. H., 602 Kaiser, J., 294-295, 300 Kalawsky, R. S., 651-654 Kalita, M., 573-574 Kalliath, T.J., 286 Kalyanpur, A., 184-186 Kamali, N., 598 Kamba, T, 643 KamlerJ., 510 Kanselaar, G., 169 Kantor, P. B., 195-196 Kapadia, R., 514 Kaplan, I., 121 Kaplan, R. S., 510 Kaplan-Leiserson, E., 454 Karagiannidis, C., 254 Karampelas, P., 243 Karamustafaoglu, A., 138 Karat, C. M., 62, 327, 680-681 Karat, J., 324 Karlqvist, L., 59 Karsh, B., 58, 63, 537, 542, 544, 547 Karshmer, A. L, 241, 252, 257, 537, 542, 544,547 Karypis, G., 206 Kashyap, V, 216 Kassell, N. E, 651 Kassirer, J. P., 664 Kato, H., 647, 652 Katz, A. S., 123 Katz, D., 144-145 Kaufman, D. R., 671 Kaufman-Scarborough, C., 286 Kaur, K., 650, 652 Kawasoe, T, 232 Kaye, B. K., 309,311-312 Kazman, R., 245, 248 Kearsley, G., 462 Keikel, P. A., 396
AUTHOR INDEX Keim, D. A., 511 Keir, P.J., 52 K e l l a ,J.J.,5 2 Keller J. M., 462, 464 Keller, T. A., 642 Kelley, C. R., 7, 10 KelsoJ. A. S., 93 KelsoJ., 333, 510 Kempf, E, 544 Kendall, J., 306 Kendall, K., 306 Kendon, A., 227 Keniston, R. C., 55 Kennedy, P., 58 Kennedy, R. S., 654 Kennedy, R., 653-654 Kerka, S., 467 Kerr, D., 553 Kerr, N. L., 92 Kiekel, E A., 91, 94-95, 97-99 Kiel, G., 582-583 Kieras, D. E., 74, 78-81, 388-389, 397-398, 400-404, 425, 427, 670 Kiesler, S., 97, 459, 567 KigerJ. I., 107 Kilbom, A., 146 Kim, D. J., 597 Kim, J., 601-603 Kim, K., 199 Kim, S., 608 Kimball, R., 388 King, M., 332, 617 King, V, 329 Kinkade, R. G., 386 Kipp, M., 230 Kipping, G. J., 112, 121 Kirakowski, J., 332 Kirkpatrick, D. L., 466 Kirlik, A., 86-87 Kirwan, B., 73-74, 303, 318 Kitajima, M., 426 Kitamura,Y, 232 Kjellberg, A., 146 Klauer, K. C., 485 Klein, G. A., 73-75, 168, 171 Klein, M., 245, 248 Kleinke, J. D., 662, 665, 667, 670 Klesen, M., 225, 230 Klinec, D., 648 Klobas, J., 461 Klusch, M., 211-212 Knapp, E, 485 Knave, B. G., 52 Knieling, J., 286 Knoblich, G., 92 Knoblock, C. A., 214 Knott, B. A., 121 Knox, S. T., 60 Knutson, S. J., 52 Kobayashi, M., 202 Kobb, R., 658 Kobsa, A., 252-254, 430, 436 Kocharekar, R., 574 Kochen, M., 195
Kochut, K., 218 Kodimer, M. L, 125 Koehler J. W., 533-534, 539-541, 544-545 Koh, T. K., 643 Kohlrausch, A., 134, 149 Kojima, Y, 287, 296 Kokubo, T., 232 Kolers, P. A., 121 Kommers, P., 169 Kondziela J. M., 327 Kontarinis, D. A., 62 Koonce, J. M., 4 Kopala, C. J., 123 Kopp, S., 231 Korgaonkar, P., 601-602, 607-608 Korn, P, 241, 252, 257 Kornbrot, D., 146 Koubek, R. J., 71-72, 75-76, 78, 80-81, 83, 85-86 Koufaris, M., 581, 583-584 Koyani, S., 118 Kromer, H., 509 Kramer, A. E, 83 Kramer, J., 331,496 Krantz J. H., 309, 471, 486 Krapichler, C., 652, 654 Krasner, G., 43 Kremer, R., 165, 171 Krenn, B., 231 Kress, G., 287 Krishnan, K., 295-296, 300 Krizman, V, 86 Kroemer, K.H. E., 51, 58, 61 Kroemker, H., 617 Krol, E., 14 Krueger, B., 312 Krueger, H., 121, 648-649 Krugman, H. E., 599-600 Kruk, R. S., 121 KruskalJ. B., 421 Kubo, K., 648 Kuchinsky, A., 108 Kugel, P., 194 Kuhn, W., 388 Kuhns, J., 194 Kulikowich J., 431 Kumar, V, 206 Kunreuther, H., 309 Kuo, C., 596 Kurland, L. T., 55 Kurlander, D., 230 Kurniawan, S. H., 294, 299 Kuruso, M., 287 Kushniruk, A. W., 671 Kuutti, K., 242 Kuwano, S., 141 Kwak, H., 596, 599 L
Losch, A., 652, 654 Laubli, T., 55, 63
709
La Porta, T. E, 641 Laakso, K.-R, 313, 317 LaBore C., 233 Lackner, J., 653-654 Lafrehiere, D., 331 Laham, D., 99 Lai, V S., 544-546 LairdJ. E., 78, 81, 230-231 Lake, D., 514 Lam, A., 530-532, 547 Lamers, 516 Lamias, M. J., 310-311 LaMotte, R. H., 146 Lamping, J., 511 Landa, A. S., 667 Landauer, T. K., 51, 99, 680 LandayJ. A., 309 Landgrave, T., 573 Landis, D., 577 Lane, D. M., 118, 516 Lane, N. E., 654 Langan-Fox, J., 91 Langneld-Smith, K., 91 Langrana, N., 652-653 LaPlante, A., 680 Lapp, W., 394-395 Laraki, O., 602-603 Large, A., 431 Larsen, L., 115 Larson, K., 39, 108 Lashina, T., 140 Laskowski, S., 502 Lastovicka, J. L., 599-600 Lattanze, A., 248 Lau, P. W C., 292, 294, 299-300 Lau, T., 205 Lauche, K., 648-649 Laufmann, S., 211 Laurent, D., 311-312 Laurillard, D., 458 Lavery, D., 324, 672, 689 Lavidge, R. J., 600 Lawless, K., 431 Lawson, B., 653-654 Lazar, J., 128, 511-512, 537-538, 542, 544, 547 Lazarus, R. S., 228 Lea, M., 97 Leake, D. B., 168 LeBaronJ., 461, 465, 467 Lebiere, C., 81, 425, 428, 670 Lebmann, R, 231 Lechner, J., 496 Lederer, A. L., 680 Lederman, S. J., 145-146, 151, 194 LeeJ., 601-603 Lee, D. Y, 286 Lee, E. J., 234 Lee, E., 147 Lee, F. J., 428, 430 Lee, H-C., 672 Lee, K. P., 288, 296 Lee, M., 601 LeGrow, G., 663
710
•
AUTHOR INDEX
Lehnert, H., Lehto, M., 72-73 Leigh, B., 488 Lemke, C., 464 Lemmens, P. M. C., 136 Lemmens, R., 519, 521, 523 Lengel, R. H., 97 Lenzner, R., 455 Leon, M., 680 Lergier, R., 126-127 LeRouge, C., 632 Lesgold, A. M., 171 Lesh, N., 231 Lester, J. C, 228, 232 Lester, K. J., 666-667 Leulier, C., 690-693 Leventis, A., 254 Levin, K., 519 Levine, L. M., 451 Levitt, B., 529 Levy, E., 120 Lewis, C., 324-325 Lewis, J. R., 58, 332, 425 U, C,524 Li, H., 596, 598 Liang, J., 651 Liang, S. M., 286 Liang, S. F. M., 285-286, 294-296, 300 Liao, C., 121 Liaw, S. S., 582-585 Licklider, J. C. R., 7 Lida, B., 121, 127 Lieberman, H., 195 Lien, M. C., 617, 620-621 Liepins, G. E., 86 Light, A., 45, 392, 394 Lightner, N. J., 509, 596, 607, 617 lilienthal, M. G., 654 Lim, V. K. G., 332 Lin, C. J., 467 Lin, H., 81 Lin, M. L., 61 Lin, M., 671 Lind, M., 388 Lind, S., 167 Lindquist, D. B., 641 Lindquist, J. D., 286 Ling, D., 230 Ling, J., 110, 117 Lintern, G., 9, 10, 168 Littell, C. L., 663, 667 Little, J., 553 Little, R., 310 Liu Sheng, O., 580-581, 583 Liu, I. M., 284 Liu, K., 206 Liu, L., 289, 294, 299-300 Liu, X., 286 Liu, Y, 73 livinglnternet, 23 Ljungstrand, P., 57 Lloyd, C.J. C, 121 Lockner, M., 648 Lockwood, L. D., 357-358, 462
Logan, G. D., 604 Lohr, C., 40 Lohse, G. L., 79, 403, 598-599, 606 Loker, S., 598 Longest, B. B., 659 Longstaff, T., 245, 248 Lonsdale, R., 392, 394 Loomis, J. M., 194 Loosveldt, G., 312 Lopez, B., 212 Lotem, A., 200 Lotz, S. L, 580, 582-583, 585 Lovett, J., 662 Lovett, M. V, 425 Lowengart, O., 602-603, 608 Lu, 2., 225 Lubin, K., 509, 617 Lucas, H., 119 Lucas, W., 199-200 Luck, M., 211 Luczak, H., 61-62, 404 Lumpkin, J. R., 669 Luna, D., 289, 294-295, 299-300 Lund, A., 688 Lund, P., 434 Lundstrom, R., Luximon, A., 294, 300 Lynch, K., 651 Lynch, P. J., 34, 118, 124-126, 512, 688, 691-692, 697 Lyngbaek, U., 332 M
Mueller, K., 544 Muller, S., 648 Muller, J., 229, 233-234 MacDonald, J., 480-481 MacDonald, K., 663-664, 671 Macgregor, D., 74-75 Mack, D. E., 579 Mack, J. D., 146 Mack, R. L., 74, 345 MacKenzie, I. S., 31,651 Mackinlay, J. D., 213 Macuga, K. L., 425 Maczewski, M., 109 Madden, T.J., 594 Maeda, K., 63 Maes, P., 213, 430, 437 Maglio, R, 431 Magnenat-Thalman, N., 147 Magnusson, K., 36 Magyar, R. L., 58 Mahapatra, R. K., 544-545 Maida, J., 653, 654 Majchrzak, A., 76 Maldonado, C., 127 Malhotra, Y, 540, 582-584, 589 Malinowski, U., 252 Malt, L. G., 58 Maker, A. J., 598 Mancini, R., 40
Mandel, N., 605 Manfreda, K. L., 311 Mankoff, J., 58 Mann, S., 56 Manning, C. A., 99 Manning, H., 514, 518 Mantei, M., 681 Maou, N., 254 Marable, L., 603 MarchJ. G., 529 Marchetti, F. M., 121 Marchionini, G., 197, 200 Marcus A., 123, 194, 241, 252, 257, 287-288, 295, 300 Marcus, R., 194 Maren, A. j., 86 Mariage, C., 690, 692, 694 Markham, K., 165 Markin, R. J.. Jr., 595 Marklin, R. W, 58 Marks, W., 225 Markus, M. L., 97 Marmostein, H., 602-603 Maron, M., 194 Marsella, S., 228 Marsh, M., 242 Marshall, A., 233 Marshall, C., 688 Marshall, E., 392, 404 Marshall, J., 602-603 Martin, B. L., 165 Martin, B., 431 Martin, F. T., 544 Martin, G. D., 286 Martin, M., 99, 523 Martinez, K., 647 Martinez, M., 461 Martucci, J., 663 Maseda, J. M., 436 Masterson, J. T., 91 Mastrianni, S., 641 Matessa, M., 670 Mathieson, K., 580, 583-585 Matias, Y, 199, 204 Matsunobe, T., 108 Mauldin, M., 222 Mavor, A. S., 650 Mawarimichi, Y, 232 Maybury, M. T., 213 Mayer, R., 169 Mayes, T., 135 Mayhew, D. J., 303, 305, 318, 338-339, 343, 345-346, 348, 356-358,386,389, 679-681 McArthur, R., 199-200 McBurney, P., 211 McCabe, S., 309 McCalla, G., 434 McCarthy, J. C., 514, 517 McClelland, J. L., 85-86 McCormack, J., 665 McCormack, M., 308, 313, 315 McCormick, E. J., 52, 59, 652 McDermott, L, 606, 608
AUTHOR INDEX
McDonald, C.J., 661 McDonald, S., 169, 536 McEntire, R., 231 McFarland, A., 331 McFarlane, D. C., 651 McFarren, M. R., 167 McFedries, P., 201 McGee, M. R., 151 McGrath, J. J., 10 McGraw, K. L., 73 McGraw, K. O., 309, 479 McGuinness, D., 179, 183 McKay, D., 231 McKendrick, H., 61 McKinney, V, 582 McKownJ., 121 McNeely, W. A., 653 McNeese, M. D., 166-168 McNeil, J. M., 523 McNulty, M., 510 McTyreJ. H., 123 Mead, A., 653-654 Mead, S., 193, 199 Meadow, C., 195 Meadows, K. D., 55 Meehan, M. M., 507 Meek, C., 202 Meek, D., 92 Meertens, R., 594 Meister, D., 387-389, 392, 394 Mellers, B. A., 485 Meluson, A., 125 Meluson, M., 125 Meng, W., 206 Menon, S., 605 Merayyan, A., 199 Mers, L., 121 Metta, S., 39 Metzger, J., 663-664, 671 Metzker, E., 358 Mewhort, D. J. K., 121 Meyer, D. E., 78, 81, 389, 425, 427, 670 Meyer, J., 660 Meyer, M., 658 Meyers-Levy, J., 289 Mezias, S.J., 529 Michalak, E. E., 489 Microsoft Corporation, 125, 128, 252, 692 Microsoft Task Gallery, 214 Miles, D., 168 Milgram, P., 93, 651, 654 Milheim, W., 431 Militello, L. G., 73, 75 Millard, D. E., 648 Millard, E., 519 Millen, D., 312-313 Miller, C. S., 108 Miller, D. P., 107 Miller, D., 460-461, 465, 467 Miller, G. A., 8, 39 Miller, J. R., 324, 327 Miller,J., 230 Miller, J. A., 608-609
Miller, M., 536 Miller, R. B., 386-387, 393, 516 Miller, T., 652 Millican, P. J. R., 15 Mills, C. B., 121, 122 Mills, J. A., 289, 299 Mills, M., 121 Milosavljevic, M., 647 Mine, M. R., 651 Miner, A. S., 150,529 Miniard, P. W., 608 Minocha, S., 287 Minor, M., 599-600, 609 Minsky, M., 149 Minton, S., 214 Mintzes, J. J., 157,165-166 Minuto, A., 121 Misra, S., 596, 601-602 Mitchell J., 661 Mitchell, T., 195 MITECS, 71 Mithal, A. K., 61 Mitrovich, T., 169 Mittman, R., 660, 662-663 Miyagawa, S., 457 Miyazaki, A. D., 601-602, 604 Mizzaro, S., 194 Mladenic, D. S.J., 212 Moe, W. W., 596-598, 608 Moffat, D., 228 Molet, T., 147 Molich, R., 127, 324 Mollaghasemi, M., 651-654 Mollica, R. L., 660 MoltgenJ., 388 Monaghan, M., 200, 203 Money, K. E., 654 Monk, A. E, 35, 39, 45, 324, 462 Monkman, G. J., 146 MonroeJ. E, 58 Monroe, K. B., 605 Montaner, M., 212 Mon-Williams, M., 653-654 Moon,J., 601-603 Mooradian, M., 519-520 Moore, A., 514 MooreJ. L, 93 Moore, M., 58 Morales, I., 693 Moran, D. W. Moran, T. P., 74, 78-81, 83, 243, 323, 389, 396-398, 402-403, 426, 552, 670 Moreau, L., 648 Moreland, R., 55 Moreman, D., 168, Morgan, C. T., 5 Morgan, D. L., 331-332 Morgan, G., 446 Morgan,;., 58 Morimoto, C., 61 Morkes,;., 118, 295, 300 Moroney, W. E, 332-333 Morris, M. G., 581, 583-584, 586 Morris, S., 553
711
Morrison,;. B., 393 Morrison,;., 199 Morville.L, 341,343 Morville, P., 447, 448 Moschovitis, C. J. P., 13-14, 22, 24, 118 Mosier, J. N., 691 Mosier, J., 501 Moskel, S., 121 Moss, C., 463 Moss, E A., 385 Moston, S., 574-575, 577-578 Mothe, J., 214 Motwani, R., 202 Mowen, J. C., 599-600, 609 Mtei, L., 107 Mudumbai, S., 216 Muesch, H., 141 Mukherjee, A., 607 Mukhopadhyay, A., 607 Mulder, M., 412 Muldoon, J., 169 Mullen, M. P.Jr., 651 Muller, M. J., 303, 331, 670 Munch, J., 86 Munsterberg, H., 385 Murakami, Y, 232 Murphy, H., 241, 252, 257 Murphy, J., 605 Murray, I. R., 436 Murray, W. D., 333 Murtagh, E, 214 Musch, J., 472, 485, 487, 489 Muslea, M., 214 Muter, P., 121 Myers, B. A., 511 Myers, L. D., 55 Mylopoulos, J, 245, 248 Mynatt, E. D., 58, 142 N
Nachemson, A., 65 Nagaseko, M., 58 Nagata, S., 62 Nahamoo, D., 62 Najjar, L. J., 509, 514-515, 523, 617 Nakanishi, H., 234 Nail,;., 118 iNamba,S., 141 Nambisan, S., 544 Nardi, B. A., 92, 242 NASA, 166 Nass, C., 234 Nathan, P. A., 55 National Academy of Sciences (NAS), 55 National Assistive Technology Research Institute (NATRI), 393 National Cancer Institute, 124, 349, 353, 693, 698 National Learning Infrastructure Initiative (NLII), 445, 449 National Research Council, 663, 665, 667-668
712
AUTHOR INDEX
Navathe, S., 200 Nawrocki, I., 119 Neal, L, 455, 457, 459-461, 464 Neale, W C., 333, 510 Neath, I., 479 Negroponte, N., 220 Neilson, L, 553 Neisser, U., 92 Nelson, C., 688 Nelson, T., 39 Netguide, 599-600 Neuhaus, C., 478, 485 Neuwirth, C., 121 New Zealand E-government, 523 Newell, A., 72, 78-79, 81, 83, 92, 323, 389, 396-398, 402-403, 425-426, 428, 670 Newhouse, J. P., 667 Newlin, M. H., 452 Newman, D., 653-654 Newman, M., 309
Newt, J ., 9, 14, 35, 37, 74, 94,109-110, 112-113,116-119, 124, 126-128, 136, 295-296, 300, 306, 321-326, 332-333, 335, 338, 341, 345-346, 348-349, 353, 386, 443, 509, 516, 518-519, 523-524, 528, 538-539, 544-546, 681, 688, 690-691, 693 Nikolic, D., 99 Nilan, M., 194 Nill, A., 254 Nilsson, B., 52 Nilsson, R., 169 Niranjan, S., 165-166 Nisbett, R. E., 284 Nishiyama, K., 63 Nishiyama, Y, 108 Nitzkin, J. L., 667 Nixon, B., 245, 248 Noirhomme-Fraiture, M., 697 Noland, S., 71, 72, 78 Nonaka, L, 531 Norenzayan, A., 284 Norman, 639, 643 Norman, D. A., 57, 71-73, 92-93, 415, 462, 464, 670 Norman, K. L., 332 Norman, K., 110 Noronha, S., 496 Nortel Networks, 305 Northwestern Michigan Web Services, 110 Norton, D. P., 510 Noser, H., 147 Novak, J. D., 157-158, 165, 161-162, 165-166 Novak, T. R, 609 Nowak, A., 93 Nowicki, J., 380
o O'Brien, J., 32 O'Connell, M., 658 O'Connor, C., 519 O'Connor, J., 195 O'Donnell, A. M., 169 O'Donnell, P. J., 332 O'Fallon, W. M., 55 O'Hara, K. P., 330 O'Modhrain, 134 O'Neil, H. F. Jr., 169 0'Neil, K. M., 311,312 Oakley, I., 136 Obendorf, H., 119 Odagawa, T., 62 Oehme, O., 61, 62 OhJ., 214 Ohlsson, K., 293, 299 Ohnemus, K. R., 389 Okamoto, T., 165 Oke, L. E., 145 OLIVE (Online Library of Information Visualization Environments), 214 Oliver, R. L., 609 Olsen, D., 239 Olshavsky, R. W., 608-609 Olson, J. C., 595, 599-600, 602, 671 Olson, J. P., 529 Olson, J. S., 243, 551, 554, 671 Olson, W A., 59 Omanson, R. C., 692 Online Computer Library Center, 70 Open Software Foundation, 36 Oppermann, C., 241 Oppermann, R., 253 Orderberg, G., 54 Ordorica, L., 536 Orlikowski,W.J., 553 Ormerod T. C., 392 Ornstein, A. C., 466 Orr, A., 516 Orr, B., 55 Ortengreen, R., 65 Ortony, A., 28 Osborn, R., 667 Osgood, C., 229 Osipovich, A., 602-603 Ostberg, O., 55 Ota, N., 56 Ottensmeyer, M. P., 146 Overbeeke, K., 45 Ozawa, S., 86 Ozok, A., 113 Oztekin, B., 206
P Padgett, M. L., 647 Page, C. R, 313, 317 Page, L., 201-202 Page, S. R., 125-126 Pai, D. K., 151
Pain, H., 437 Palanque, P., 34, 690-692 Palen, L., 632, 639 Palloff, R. M., 445 Palmer, J., 534, 542-545 Palmer, M., 86 Pan, C. S., 58 Pandzic, L, 147 Pane, J. E, 511 Pangalos, G., 615 Pantel, C., 502 Paparoulis, G., 254 Papazoglou, M. P., 211 Pape, L. J., 331 Paramythis, A., 241, 254, 257, 264 Parasuraman, R., 96, 672 Parchman, M., 311-312 Pardo, B., 202 Parducci, A., 475 Pare, G., 577 Parente, S. T., 663, 665, 669 Paris, C., 647 Park, S., 86 Park, T. K., 194-195 Parker, B., 121-122 Parker, L., 195 Parker, R. A., 309-310, 312 Parkinson, S., 107 Parsia, B., 178, 184-186 Parsons, H. M., 4 Parsuraman, K., 216 Partnership for Solutions, 660 Parush, A., 127 Parush, D. K., 127 Pascarelli, E. E, 52 Pasman, G., 419, 421 Passmore, C., 311-312 Pastoor, S., 123 Pastore, M., 516 Patel, S., 218 Patel, V L., 659, 670-671 Patel-Schneider, R, 179 Paterno, E, 34, 223 Patrick, D. L., 412 Patrick, J., 392 Patterson, M. E., 169 Paul, J., 602-603 Paul, Y, 518 Pausch, R., 651, 654 Pavlou, P. A., 582-583, 601 Paxon, R,533-534,539-541,544-545 Payne, P., 76 Payne, SJ., 26, 78 Payne, T. A., 9 Payne, T., 178 Pazzani, M., 430, 643-644 Peacock, B., 537, 542, 544, 547 Peacock, E., 580, 583, 584-585 Pearrow, M., 324-325, 327-328, 637, 644, 689 Pearsall, N. R., 165 Pearson, G., 61 Peck, J., 581, 583 Peio, K. J., 167
AUTHOR INDEX Pejtersen, A. M., 76, 408, 411, 413, 415-416 Pelachaud, C., 228-230 Pellegrini, R. S., 134, 136, 138 Peng, K., 284 Pengelly, M., 436 Pennington, E., 330 Penrod, S. D., 311-312 Pentland, A., 436 Peracchio, L. A., 289, 294-295, 299-300 Perkins, R., 345, 351, 354, 510 Perlin, K., 229 Perreault, L. E., 659, 661, 671 Persichitte, K., 463 Peruse, M., 51 Peter, J. P., 595, 599-600 Peters, S., 641 Peterson, M., 121 Petrie, H., 146 Petrucci, L., 146, 152 Petta, P., 212 Pettersson, E., 388 Petty, R., 599-600 Pew, R. W., 6 Pfaff, P, 35 Philips, J. R., 144, 146 Phillips, B., 109, 641 Phillips, C., 81 Phoenix Health Systems, 667 Piamonte, D. P. T., 293, 299 Piantanida, T., 653-654 Picard, R., 436 Piccoli, B., 64 Pidermann, M., 63 Pieper, M., 243 Pieper, S., 653 Pirhonen, A., 152 Pirker, H., 231 Pirolli, P., 199, 393, 428, 511 Pitkow, J., 19,39, 199 Piwek, P., 231 Pizzolato, J., 515 Pliicker, R., 141 Plaisant, C., 128, 200 Plaut, D. C., 88 Plocher, T. A., 285-286, 288-289, 294-296, 299-300 Plummer, B., 199 Poggi, A., 223 Poggi, I., 228-230 Pohl, K., 242 Pohl, W., 252 Pokorny, R. A., 387, 396 Pollard, D., 141 Poison, P. G., 78, 325, 426 Pooch, U., 126, 200, 203 Poole, H., 13-14,22,24, 118 Pope, S., 43 Posner, L, 554 Posner, M. L, 31,652 Postmes, T., 97 Potenza, M., 330 Potosnak, K. M., 58
Potter, M. A., 659 Poulton, E. C., 4, 121 Poupyrev, I., 652 Powers, R., 119 Prabbu, P. V, 51 Prabhu, G., 287, 296 Prager,;., 15 PrassadJ., 680 Pratt, K. 445 Preece, A., 167 PreeceJ., 108 Preist, C., 211 Prelec, D., 608 Prendinger, H., 229-230 Press, L., 551, 554 Presser, S., 310, 312 Pressman, R. S., 680 Prevost, S., 230 Price, L., 608 Prinz, W., 92 Priyantha, N. B., 641 Proctor, R. W., 10, 205, 232, 332, 335, 509,617,620-621,670 Proulx, N., 331 Prumper,;., 329 Pryor, H., 62 Pugh, D., 230 Pugh, H. L, 73 Pun, T., 146, 152 Pycock, J., 225 Pysarchik, D. T., 608 R
Rose, K., 289, 294, 299-300 Rabb, M., 120 Raben, M., 8 Rademacher, R. A., 544 Radl, G.W, 121 Radwin, R. G., 61 Ran, N., 294, 300 Ragus, D., 516, 522 Raiello, E, 327 Rainie, L., 664 Raman, T. V., 252 Ramey, J., 305, 313 Ramsay, J., 108 Ramsey, M., 200 Ramstein, C., 146 Randel, J. M., 73 Rangarajan, G., 86 Rangnekar, A., 602-603 Rao, H. R., 596-597, 601-602 Rao, R., 511 Rasammy, M., 125 Rasinski, K., 310 Rasmussen, J., 76, 84, 93, 411-413, 670 Rath, G., 194 Ratner, J., 349, 353, 689, 690, 692-693 Ratschiller, T., 128 Rau, P-L. P., 286, 295-296, 300 Rauterberg, M., 648-649 Ravaglia, R., 254
713
Ray, M., 599 Raymond, B., 658 Rea, L. M., 309-310, 312 Reader, W., 169 Readiness assessment, 456 Reber, A. S., 428 Reder, L. M., 425 Redish, J. C., 303-304, 313, 315, 343, 345, 357-358 Reed, D., 462 Reed, P., 688 Reed, S. K., 73 Rees, A., 194, 195 Reeves, L, 509, 617, 651-654 Regan, D. X, 576 Regazzi, J., 195 Regli, S. H., 121 Rehman, A., 519-520 Reibstein, DJ., 305, 596-598, 607 Reich, S., 648 Reichheld, F. E, 517, 519 Reichherzer, T., 165-166, 168-169 Reid, D., 146 Reilly, W. S., 228 Reinhold, N., 488 Reips, L. J., 310 Reips, U. D., 309, 311-312, 471, 474-475, 478, 485-489 Reiser, J. E, 433 Reisner, P., 78 Reitberger, E, 533-534, 539-541, 544-545 Reiterer, H., 358 Reithinger, N., 224, 231 Rella, L., 254 Remington, R. W., 108 Rempel, D., 55, 58 Renaud, K., 391-392 Resnick, A., 194 Resnick, M. L., 126-127, 200, 203, 537, 542, 544, 547 Resnick, P., 40 Reynolds, M., 75, 80-81, 83, 85 Reynolds, S., 169 Rhee, H. S., 534, 539, 544, 690 Rhodes, J. S., 521,681 Riccio, G. E., 654 Rich, A., 500 Rich, C., 231 Richard, P., 652-653 Rickel, J.,225, 227 Rickwood, D., 587 Rickwood, R., 581, 583, 584 Ridgeway, N. M., 599-600 Rieber, L. P., 442 Riedel, O., 653-654 Rich, S., 199 RiemanJ., 325, 330 Rigden, C., 124 Riggins, F. J., 690 Riggins, E, 534, 539, 544 Righi, C., 358 Rihs, S., 150 Riley, S., 114, 121, 127
714
•
AUTHOR INDEX
Riley, V, 96, 289, 299, 672 Rimell, A. N., 150 Rind, D., 659 Ringel, S., 119 Risden, K., 329-330 Rissman, A. K., 92 Rist, T., 225, 229-231 Ritchie, M. L, 8 Rivlin, E., 199, 204 Roa, R., 663 Roache, R., 305 Roberts, P., 580-581, 583-585, 587 Roberts-Witts, S. L., 680-681 Robinson, J., 128 Roblimo, 523 Robtyer, M., 463 Rodden, T., 32, 225, 308, 312, 317, 329 Rodgers, S., 596, 598 Rodriguez, N. J., 693 Roetting, M., 61, 62 Rogers, B. L., 517 Rogers, E. M., 455, 594 Rogers, P. S., 97 Rogers, W., 193, 199 Rogers, Y., 92-93 RohnJ., 503 Rojas, R., 15 Rollermouse., 61 Romero, S., 516 Rommelse, K., 252 Ros, C, 39 Rosa, J. A., 598 Roscoe, S. N., 4-6, 9, 11 Rose, A., 128 Rose, J., 545-546 Rose, M. J., 58 Rosen, A., 535-536, 538-539 Rosen, E., 662 Rosenbaum, S., 313 Rosenberg, D., 503-504 Rosenbloom, E S., 78, 81, 230-231 Rosenfeld, L, 341, 343, 447-448 Rosenholtz, R., 199 Ross, S., 683 Rosson, M. B., 324, 358 Roth, P., 146, 152 Rothrock, L., 86-87 Rouet, J. E, 39 Rounceneld, M., 32, 308, 312, 317 Rourke, L., 467 Rouse, W. B., 83, 86 Rousseau, G., 193, 199 Rowe, R. D., 62 Rowley, J., 606 Rowley, E, 60 Roy, E. J., 311 Rubin, A., 622 Rubinstein, 121 Rudorfer, A., 509, 617 Rumelhart, D. E., 85-86, 92 Ruppin, E., 199, 204 Rushton, S., 654 Russell, M. G., 596 Ruthven, I., 205
Ryan, P., 658 Ryan, W. T., 451 Ryder, J. M., 75 Ryu, D., 97
S Sabnani, K. K., 641 Sacharow, A., 519-520 Sadowski, W., 652-653 Safran, C., 659, 661, 671 Sahrhage. J., 152 Sainfort, E, 531, 537, 542, 544, 547, 658-659, 661-664, 666-667, 672-673 Saito, S., 64 Saito, Y., 652 Salas, E., 91, 93-94, 96-97, 99 Sales, G., 121 Salgeback, S., 54 Salisbury, J. K., 146 Salisbury, K., 653 Salton, G., 195 Salvendy, G., 51, 72-73, 76, 86, 113, 193, 200, 241, 252, 257, 293-294, 296, 299-300, 323, 334-335, 509, 537, 542, 544-545, 547, 617, 620-621, 653-654 Salvucci, D. D., 425 Salzman, M., 632, 639 Samaras, G., 641 Sandblad, B., 388 Sander, R., 652 Sanders, K., 58 Sanders, M. S., 52, 59, 652 Sanderson, P. M., 97 Sannier, G., 147 Sano, D., 341, 346, 349, 353 Santos, J. M., 127 Saracevic, T., 194-196 Sarasohn-Kahn, J., 663 Sartori, C., 219 Sato, M., 62 Sauter, S. L., 52, 58 Savage, T., 194 Savastano, H., 333 Savidis, A., 241, 243, 245-246, 253-254, 256-257, 264 Sawaragi, T., 86 Scaff, M. D., 533-534, 539-541, 544-545 Scanlon,T., 110, 112, 116, 118, 126-127, 341, 349, 353 Scapin, D. L., 324, 690-693 Scargle.J., 165 Scarlett, D., 118 Scerbo, M. W, 430 Schar, S. G., 648-649 Schachter, J., 169 Schachter, S., 91 Schad, E. J., 168 Schamber, L., 194-195 Schaninger, C. M., 601 Scharff, L. V, 122, 124
Scharff, L., 121-122 Schatz, B., 200 Schatz, S., 449 Schauble.C. J. C, 24 Scheepers, R., 538, 540, 542-546, 604, 609 Schefter, P., 517, 519 ScherJ., 486 Schiano, D., 120 Schick, A., 141 Schiffman, S. J., 580, 583 Schilit, B. N., 643 Schleifer, L. M., 52 Schmidt, R. C., 92-93 Schmidt, W. C., 471, 473, 478-481, 488 Schmidt-Nielsen, A., 95 Schmierer, K. A., 164, 166 Schmitt, E., 524 Schmitt, M., 229 Schmitz, J., 97 Schneer, J. A.,97 Schneider, G. E., 228, 230 Schneider, W, 92-93 Schneider-Hufschnidt, M., 252 Schneier, B., 619 Schoen, C., 667 Schoenfeld-Tacher, R., 463 Scholtz, J., 63, 502 Schon, D. A., 411,415, 529 Schopler, J., 599 Schroder, M., 231 Schraagen, J. M., 74-75, 389, 396 Schrage, M., 411 Schreiber, G, 432 Schroeder, R., 119 Schroeder, W., 110, 112, 116, 118, 126-127, 341, 349, 353 Schuhmann, D., 652, 654 Schultz, E. E., 613, 617-618, 620-621 Schulze, K. G., 95 Schumann, D., 599-600 Schutze, H., 199 Schuyler, T., 13, 14, 22, 24, 118 Schvaneveldt, R. W., 94, 99 Schwalbe, P., 388 Schwartz, J., 523 Schwarz, E., 434 Schwarz, S., 479 Scobie, G., 332 Searle.S.J., 23 Sears, A., 32, 51, 535, 537, 542, 544-547, 670-672 Sebanz, N., 92 Seeger, C. M., 7 Seemann, M., 652, 654 Selfe, C. L, 97 Selingo, J., 517 Selker, T., 431 Sellen,A.J., 330, 334 Sellis.T. K., 211 SeMdge, E, 108 Seminerio, M., 514 Senay, H., 333 Senft, T., 13-14, 22, 24, 118
AUTHOR INDEX
Senge, P. ML, 529, 534 Senning, J. R., 17 Sensus Internet Browser, 252 Sequin, C., 651 Serflek, C., 252 Sewell, S., 660, 663-664 Sfyrakis, M., 241, 254 Shackel, B., 58 Shadbolt, N., 168, 171, 331 Shah, K., 216 Shalin, V L, 74-75, 389, 396 Shamash, A., 645 Shamma, D., 165 Shanis, J., 59 Shao, Y, 583-584 Shapiro, M., 641 Shardanand, U., 430 Sharit, J., 72-73 Sharrard, J., 517 Sharrat, B., 691 Shaw, E., 233 Shaw, M. E., 91-93 Shaw, M., 171 Sheeley, G. A., 126 Sheldon, K. M., 596, 598 Sheldon, T., 614 Shen, M., 292, 299 Shepherd, A., 387, 390-392, 670 Sheppard, B. H., 594 Sheridan, T. B., 86, 389 Sheridan, T., 653 Sherif, C., 599 Sherif, M., 599 Sherrell, D. L., 581, 599-600 Sherry, J. E, 601 Sheth, A., 213, 216, 218 Shi, L., 659, 661 Shieh, K. K., 292, 299 Shiffrin, R. M., 92-93 Shifrin, J., 202 Shih, H. M., 292-294, 299 Shim, S., 580, 582-583, 585 Shimizu, S., 62 Shirk, H. N., 293 Shklar, L., 216 Shneiderman, B., 57, 110, 121, 128,198, 200, 202, 213, 220, 241, 414, 437, 462,511,650-651 Shoham, Y, 253 Shon, J., 602-603 Shope, S. M., 91, 99, 396 Shortliffe, E. H., 671 Shubin, H., 507 Sibert, J. L, 651 Sibert, L. E., 651 Sidner, C. L., 231 Sidorsky, R. C., 123 Siegel, C., 613 Siegel, D., 330 Siegfried, S., 196 Siepmann, M., 478, 488 Siewiorek, D., 56 Sigel, D., 318 Signorello, D., 648-649
Sills, J. A., 61 Silverblatt, R., 601-602, 607-608 Silverstein, B., 51 Sime,J.-A.,436 Simon, H. A., 72, 92, 94-95, 168, 412 Simoneau, G. G., 58 Sinclair, P. A. S., 647 Singer, M., 652-654 Singer, S.J., 662 Singh, D. A., 659, 661 Singh, M., 504 Singh, R., 178 Sinha, R., 126 Sirin, E., 178 Sisson, N., 107 Sit, R., 193, 199 Sivier, J. E., 9 Skaggs, L., 169 SkeenJ. R., 9 Skelly, T., 230 Skipper, J., 165 Slater, M., 653 Slaughter, L., 200 Sleeman, D., 167 Small, L, 652 Smeulders, A. W. M., 432 Smith, A., 287, 312 Smith, C. C., 7 Smith, D. C., 388 Smith, H. T., 293 Smith, J. B., 92 Smith, J. E, 5 Smith, L. A., 58 Smith, M. A., 234, 488 Smith, M. J., 51, 52, 55, 58, 63-64, 537, 542, 544, 547 Smith, R, 434 Smith, S. E, 86 Smith, S. L., 691 Smith, S., 501, 599-600 Smither, J. A., 332 Snowberry, K., 107 Snyder, C., 127, 341, 345, 347, 349, 351, 353 Snyder, D. E., 167 Snyder, H. L., 121 Solan, Z., 199, 204 Solomon, D.J., 311 Solomon, K., 524 Soman, D., 605 Sommerich, C. M., 58 Sommerville, I., 32, 42 Sonderegger, P., 356 Sonnenreich, W., 201 Soohoo, C., 603 Sorce, J. E, 327 Sorensen, S. J., 546,670 Sorkin, R. D., 136 Sotoyama, M., 64 Souza, R. K., 514, 681, 682 SowaJ. E, 158 Soyibo, K., 166 Spain, K., 119 Spears, R., 97
715
Specht, M., 253 Speckart, G., 575-576 Spector B., 588 Spector, L., 187 Spector, P. E., 91 Speier, C. 534, 542-545 Spence, R., 213, 325, 511 Sperling, D., 14 Spiller, P., 598-599, 606 Spink, A., 195-196, 198-200 Spink, S., 392, 394 Spoerri, A., 511 Spohrer, J., 649 Spool, J., 110, 112, 116, 118, 126-127, 341, 349, 353 Sproull, L, 97, 567 Spyridakis, J. H., 287, 296 Srinavasan, M. A., 146, 151, 652 Srull, T. K., 311 Stabler, E. R, 333 Staelin, R., 602-603 Stafford, J., 248 Stahl, H., 168 Stammerjohn, L. W., 52, 55 Stammers, R., B., 404 Stamper, T., 643 Stancich, L, 542, 544, 546 Stanford, J., 603 Stanford, X., 166 Stanford-Poynter Project, 118 Stankevich, D. G., 609 Stankosky, A., 230 Stanley, J., 517 Stanney, K. M., 509, 617, 650-654 Stanton, N. A., 672 Stanton, N., 33, 169, 392, 404 Stanwick, V R., 122 Stoppers, P. J., 419, 421 Starr, P., 667 Starr-Schneidkraut, N. S., 393 Stary, C, 241, 243, 252, 257 Steed, A., 649 Steedman, M., 230 Stein, L., 179 Steiner, G. A., 600 Steiner, I. D., 91, 92 Steinfield, C., 596-597, 603 Stephanidis, C., 240-243, 246, 251-254, 257, 263, 509, 537, 542, 544, 547, 617, 671, 688-689, 692 Stergiou, A., 254 Stevens, J. C., 55, 146 Stevens, S. S., 13, 145-146 Stevenson, R. J., 169 Steward, T., 692-693 Stewart, T. F. M., 55 Stewart, T., 501 Stiegn, S., 483, 485 Stigler, G., 599-600 Stoakley, R., 654 Stoffregen, T. A., 653-654 Stone, M., 230 Storey, M. A., 109 Storrer, K., 121
716
AUTHOR INDEX
Stout, R. J., 91,93-94, 96 Stoyanova, N., 169 Strathern, P., 15 Strauss, H., 138, 152 Strieker, D., 648 Stroetmann, K., 243 Strohecker, J., 660 Strongin, R. J., 666-667 Strube, M. J., 286 Strybel, T. Z., 332, 509, 617 Stuart-Hamilton, I., 92 Sturman, D. J., 653 Su, L, 194 Suchman, L., 32, 92 Sudweeks, E, 86 Sugita, S., 320 Sugiura, A., 202 Sujan, H., 607 Sujan, M., 607 Sukaviriya, P. N., 252, 643 Sullivan, D., 109, 201, 203, Sullivan, R, 332 Sullivan, T., 109, 516 Sumikawa, D., 142 Sun Microsystems, 252 Sun, S., 55 Sun, X., 286 Sun, 524, 688, 693, 699 Suresh, S., 544 Suri, J., 242 Suri, R, 165, 168-169 Surprenant, A. M., 479 Swani, P., 602-603 Swanson, D., 194 Swanson, N. G., 58 Swearington, K., 126 Sweeney, J. C, 394-395 Sweitzer, G., 64 Sycara, K., 178, 211 Synder, C, 110, 112, 116, 118, 126-127 Szabo, A., 489 T
Taatgen, N. A., 428-430 Tachi, S., 653 Tahir, M., 116-117, 518-519 Tait, C., 641 Takagi, A., 652 Takahashi, M., 62 Takahasi, K., 108 Takeda, K., 202 Takeuchi, H., 531 Tamura, H., 241 Tan, H. Z., 146 Tan, K. C., 294, 299 Tan, T. H. R., 294, 300 Tang, R C., 661, 670 Taniguchi, N., 652 Tanin, E., 200 Tannenbaum, P., 229 Tannenbaum, S. L, 91, 94, 97 Taskaya, T, 214
Tauber, E. M., 599-600, 603 Tauber, M.J., 78, 80 Tauscher, L., 39-40, 197 Taveira, A. D., 63 Taylor H., 663 Taylor, E V, 5 Taylor, E W., 385 Taylor, E, 29 Taylor, L, 338, 341, 349, 353 Taylor, P. M., 146 Taylor, R. S., 195 Taylor, R., 169 Taylor, S., 584-585 Taylor, M. M., 145 Teal, G., 121 Teasley, S., 551, 554 Tecuci, G., 435 Tedeschi, B., 514, 519, 521 Teichner, W. H., 123 Telleen, S. L., 541 ten Hagen, P., 35 Teo, T. S. H., 332 Tew, M. D., 309, 479 Thacker, S., 218 Thakur, G., 119-120 Thalman, D., 147 Tham, M. P., 294, 299 Thatcher, J., 523-524 Thiel, D., 230 Thimbleby, H., 643-644 Thomas, N. P., 195 Thomas, R. E., 58 Thomas, R., 392, 394 Thomasson, R., 116 Thompson, C., 194 Thompson, G. A., 436 Thompson, J., 29 Thompson, K., 497 Thompson, M. J., 514 Thulborn, K. R., 642 Tian, Q. H., 290, 299 Tinker, M. A., 121 Tipton, A., 55 Tittiranonda, R, 58 Tkach, D., 539-540 Todd, P. A., 584-585 Tognazzini, B., 242, 450 Tolcott, M. A., 7 Tong, J., 524 Topi, H., 199-200 Torenvliet, G. L., 93 Tortora, R., 311 Tourangeau, R., 310 Towns, S. G., 228, 232 Trace Research & Development Center, College of Engineering, University, 524 Tractinsky, N., 123, 582-583, 602-603, 608 Tracy, B., 517 Trautmann, M., 166 Travis, D., 341, 348-349, 353, 501, 692-693 Treasury Board of Canada, 523
Treier, C., 151 Treinen, M., 602-603 Treviranus, J., 252 Trevor, J., 553, 643 Trewin, S., 437 Triandis, H. C., 577 Trivison, D., 196 Trollip, S. R., 121 Tromp, J. G., 649 Truell, A. D., 309 Truitt, T. R., 99 Tsang, P. S., 388 Tsang, P., 83 Tscheligi, M., 241 Tsoukalas, L. H., 85 Tsunoda, T, 108 Tuckman, B. W., 93 Tufte, E., 120, 463 Tufts College and U.S. Naval Training Devices Center, 7-8 TulUs, T. S., 107, 110-111, 113-114, 116, 119-121, 123, 125, 510 Turisco, E, 663 Turnbull, D., 199 Turne, C. W., 168 Turnitin., 451 Turvey, M. T, 92-93 Tuten, T. L., 488 Tversky, B., 120 Tweedie, L., 169 Twigg, C. A., 468 Tygar, J. D., 617 Tysinger, J., 311-312
U U.S. Department of Education, 441 U.S. Census Bureau, 514, 523 U.S. Department of Commerce, 70, 193 U.S. Department of Health & Human Services, 660, 667 U.S. National Library of Medicine, 664 Ueda, H., 241 Uhrig, R. E., 85 Ullmer, B., 652 UNCTAD (United Nations Conference on Trade and Development), 514 United Nations, 523 Urbancic, T, 86 Urquhart, C., 392, 394 Uyeda, K. M., 324, 327 V
Vaccaro,J., 658 Vakali, A., 615 Vallacher, R. R., 93 Van Boxtel, C., 169 Van Cott, H. R, 385-386 van Dantzich, M., 230 van de Par, S., 134, 149 van de Wetering, H., 511
AUTHOR INDEX
vandenDoel, K., 151 van den Putte, B., 579, 585, 594 Van den Yen, A. H., 331 Van Der Linden, J., 169 Van der Ryn, S., 75 van der Veer, G. C., 79-80, 688 vanDijck, P., 126 van Dyk, T., 391-392 Van Ham, E, 511 van Harmelen, E, 179, 183 Van Leeuwen, T., 287 van Mulken, S., 225, 233-234 van Paassen, M. M., 412 van Schaik, P., 110, 117 VanVlietJ. C., 79 vanWelie, M., 688 vanWijk, J. J., 511 Vanderdonckt, J., 690-695, 697 Vanderheiden, G. C., 61, 241, 252, 257, 269, 281 VanLear, C. A., 93 VanLehn, K., 433 Vanniamparampil, A., 128 Varian, H., 40 Varma, C., 602-603 Vartabedian, A. G, 121 Vaughan, M. W., 323, 502, 508-509, 617, 694 Veerasamy, A., 200 Vehovar, V., 311 Vekiri, I., 160 Venema, S., 654 Venetsanopoulos, I., 252 Venkatesh, V, 581, 583-584, 586 Vera, A. H., 92 Vergo, J., 62, 496 Verillo, R. T., 145 Verplank, B., 388 vfridge limited, 46 Viant, 519 Vicente, K. J., 73-74, 76, 93, 412-413, 415 Vicino, E, 119 Vickers, B., 524 Vigoroso, M. W, 514, 523-524 Viirre, E., 62 Vilhjalmsson, H., 230 Villanuela, M. B. G, 64 Viller, S., 308, 312, 317 Vincini, M., 218-219 VirtualHuman Project, 234 Virzi, R. A., 327 Vitale, M., 582-583, 603 Vividence, 126 Voerman, J. L., 228, 232 Vogel, C. M., 311-312 Vogels, I. M. L. C., 151 Volpe, C. E., 91, 94, 97 Volz, S., 648 von der Heyde, M., 146 VonHippel, E., 307 vonUexkiill, J., 409 Voorbij, H. J., 199 Voorhorst, E, 648-649
Vosburgh. A., 125 Voss, M., 52 Vredenburg, K., 358 Vu, K-P. L., 10, 323, 332, 335, 509, 617, 620, 670 W
W3C, 524, 693 W3C—WAI, 252 Wachsmuth, L, 231 Wachter, R., 546 Wahlster, W., 213, 224, 231 Waite, K., 607 Wakcher, S. V, 478 Walker W., 241 Wallace, J., 166 Wallendorf, M., 601 Walsh, B., 197 Walther, J.B., 97-98 Wandersee, J. H., 157-158, 165 Wang, A. Y, 452 Wang, C. H., 389, 691 Wang, C. S., 294, 299 Wang, C., 601 Wang, H., 205, 509, 583-584, 601, 617 Wang, M., 109 Wang, Q., 288 Wang, S., 169 Wang, T., 58 Wang, Y, 544 Wann, J. P., 653-654 Ware, C., 652-653 Warke, Y, 218 Warren, D. H., Warrington, P., 580, 582-583, 585 Warshawsky, N., 641 Watchfire Corporation, 42 Wathen, C., 199 Watt, J. H., 93 Wax, T., 230 Wayne, J. C, 663 Weaver, B., 575 Weaver, J., 514 Webb, G, 430 Webb, N., 169 Weber, G, 241, 252, 257, 432 Weber, H., 509 Wedell, B., 650, 651 Wegner, D. M., 91-92 Wehrle, N. S., 323 Weick, K. E., 529 Weidenhaupt, K., 242 Weidenhausen, J., 648 Weiman, N., 60 Weiner, A. W, 171 Weinman, L., 123 Weinreich, G, 666-667 Weinreich, H., 119 Weinscheck, S., 380 Weinstock, C., 245, 248 Weiser, M., 57, 61 Weisman, L., 641
717
Weiss, G, 211 Weiss, M. J., 305, 596-598 Weisz, A., 8 Weitz, B. A., 607 Welch, N., 486 Welch, R., 653-654 Weldon, L. J., 121-122 Wellner, P., 647 Wells, R. P., 51-52 Wells, W. D., 607 Wenemark, M., 59 Wenzel, M.,652 Werge, I., 467 Werner, R. A., 55 Wernerfelt, B., 608 Westerfield, R., 683 Wharton, C., 324-325, 327, 672 Wheat, H. E., 146 Wheeler, B., 551,554 Wheildon, C., 121 Whitaker, R., 167-168 White, J. V, 122 White, R., 205 White, S., 15 Whitten, A., 617 Whitten, P., 596-597, 603 Wibom, R. L, 52 Wichansky, A. M., 509 Wickens, C. D., 73, 94, 311, 651, 653 Wieckowski, T. J., 580, 583 Wiemer-Hastings, P., 225 Wiener, E. L., 672 Wiggins, R., 13 Wikipedia, 19 Wiklund, M. E., 502, 505 Wilbur, S., 653 Wilde, D. N., 196 Wildstrom, S., 681 Wiley, D. A., 449-450 Williams, A. C. Jr., 4-5, 7-10 Williams, D. A., 194 Williams, J. E., 309, 479 Williams, J., 523, 688 Williams, M. R., 15 Williams, S., 121, 647 Williges, B. H., 392-396 Wilson, D. C., 168 Wilson, G. E, 83, 388 Wilson, J., 653-654 Wilson, P., 194 Wilson, R. D., 125-126 Wilson, S. C., 663 Wilson, S. R., 393 Winograd, T., 202 Winroth, U., 648, 651 Wish, M., 421 Witmer, B., 652-654 Wixon, D. R., 305, 313, 329, 392-396, 514,680 Woetzel, G, 553 Wogalter, M. S., 332 Wolf, S. R, 73 Wolfman, G, 199, 204 Wolgast, E., 52
718
AUTHOR INDEX
Wood, L. E., 331, 377 Wood, L, 313 Wood, R., 574 Wood, S. D., 321, 325-326, 341-343, 346, 349, 353, 391 Wood, S. L., 604, 606, 608 Wood, W, 248 Woodruff, A., 199 Woodruff, R. B., 609 Woods, D. D., 74, 414, 422 Woods, V, 52, 59-60 Woodson, W, 5 Woolrych, A., 324, 672 World Health Organization, 658 WrayJ., 7 Wright, P. C, 142-143, 324 Wright, P., 35, 45 Wucherer, K., 152 Wulfeck, J. W, 8
x Xiao, Y, 671 Xu, Y, 286 Y
Yahoo!, 39 Yale University, 349, 353 Yaltaghian, B., 202 Yamada, S., 320
Yamada, T, 232 Yamamoto, T., 232 Yamaoka, T., 108 Yamashita, K., 108 Yamazaki, S., 652 YanTam, K., 580-581, 583 Yang, D., 669 Yang, M. C., 146, 152 Yang, Y, 509, 617 Ye, N., 86 Yee, J., 511 Yee, P., 126 Yen, J., 228, 230 Yeo, A., 296 Yeoman, A., 392, 394 Yeung, B. Yoon, K., 582 Yoon, S., 228, 230 Yoshikawa, H., 670 Yoshitake, R., 64 You, M. L., 292, 299 Young K., 544 Young, D., 511 Young, M., 392 Young, R. M., 425 Youngman, M., 121 Yu, C, 206
z Zuhlke, D., 289, 294, 299-300 Zachary, W W, 75
ZacksJ., 120 Zaff, B. S., 167-168 Zahedi, E, 582 Zajicek, M., 252 Zakon, R. H., 21 Zaletel, M., 311 Zapf, D., 329 Zaphiris, E, 107, 110 Zarikas, V, 257, 264 Zehel, D., 58 Zeleznik, R., 652 Zeltzer, D., 651, 653 Zeno, R. J., 125-126 Zettelmeyer, E, 608 Zhai, S., 61,651, 654 Zhan, P, 126 Zhang, K., 286, 289, 299 Zhang, R., 544 Zhao, C., 286, 288, 295, 299-300 Zhao, L, 229 Zhou, R., 286 Zhu, B., 200 Zhu, W, 330, 509, 617 Zhu, Z., 292, 299 Ziegler, J., 241, 252, 257, 537, 542, 544, 547 Zinkhan, G. M., 596, 599, 608 Zona Research, 108 Zuboff, S., 545-547 Zue, V W, 62 Zurich, 488 Zwicker, E., 137
SUBJECT INDEX A
Abbreviation, 127-128, 289, 524 Abductive approach, 410 Abilities requirements analysis, 404 "Above the fold", 112, 516, 644 Abstract relations, 412 Abstraction, 94, 261, 412-414, 650, 670 Abstraction hierarchy, 414 Academic Advanced Distributed Learning (ADL), 441 Acceptability, 99, 667 Access control, 615-624, 666, 668 Accessibility, 25, 42, 125,177, 239-249, 252, 267-281, 380, 416, 441, 456, 463, 498, 504, 514, 523-524, 539, 545-547, 557, 573, 590-592, 599, 613, 644-645,659, 664, 667-668, 670, 672-673, 679, 688-699, 691, 693 Accessibility guidelines, 267, 270-271 Accessibility standards and regulations, 267, 269 Acoustic, 140-141, 377, 511, 652 Action, 142, 473, 508, 604, 642 Action constraints, 410 Action control, 203 Action interface, 643 Action oriented, 530 Action planning, 591 Action request, 231 ActiveX, 614, 622-623 Adaptability, 243, 245-246, 253-264, 357, 432-433, 434, 553 Adaptable systems, 430 Adaptation, 247, 252-264, 281, 358-359, 430-431, 437, 529, 543 Adaptation-Level Theory, 578 Adapting to Physical Limitations, 437 Adaptive automation, 432 Adaptive Control of Thought (ACT), 81, 670 Adaptive Control of Thought-Rational (ACT-R), 72, 78, 81, 425, 428 Adaptive interface and design, 10, 86, 251-264, 382, 430 Adaptive elements, 432 Adaptive GUI, 691 Adaptive Hypermedia Framework, 252
Adaptive Hypermedia Systems (AHS), 252 Adaptive information agent, 212 Adaptive navigational support, 434 Adaptive system, 411, 430, 432 Adobe GoLive, 472 Advanced Distributed Associate Memory (ADAM), 86 Advanced Research Projects Agency, 13 Advanced search, 126, 200, 203, 518 Advertisement, 113, 118, 600, 603, 605, 691 Aesthetics, 114, 123, 362-363, 419, 443, 638 Affinity diagrams, 220, 313 Affordance, 93, 408-410, 415-416, 420 AFPATRIC, 5 Aftereffects, 654 Agents, 177-178, 201, 281, 434-436, 529, 671 Aging, 659-660, 671 Air-navigation and traffic-control, 7, 96 Algorithm, 40, 83, 85-87, 94,127, 138, 195, 201-202, 206, 230, 276, 281, 433, 639 ALOHANET, 21, 629 ALT attribute, 271 ALT tag, 125, 443, 450, 523 ALT text, 272-273,691 Alta Vista, 200 Alternate spellings, 518 Alternate viewing tools, 323, 325 America Online (AOL), 24-25 American Association for Higher Education, 446 American Institutes of Research, 505 American Medical Association, 663 American National Standards Institute (ANSI), 63, 642, 693 American Psychological Association, 335 American Standard Code For Information Interchange (ASCII), 23 Analogies, 294 Animation, 37, 135, 229-231, 251, 273, 279, 297, 461, 463, 651-652 Animated icons, 128 Animated text, 276 Annotation, 184 Anthropological methods, 97 719
Anthropology, 31, 94 Antispoofing, 640 Apache, 618 Apple Computers, 630 Apple Learning Interchange, 443 Applets, 39 Application programming interfaces (APIs), 446, 635 Applications), 10, 33, 35-36, 43, 100, 127-128, 143, 169, 177, 252, 306, 339, 356, 386, 432, 436, 495-512, 641, 647-648, 651, 662, 688 Application context, 220 Application domain, 233-253 Application-independent languages, 230 Applied cognitive models, 80 Applied experimental psychology, 5 ARCHIE, 629 Architectural components, 247 Architectural space, 252, 649 Architecture of system, 35-36, 43, 85, 179, 202, 206, 213, 231, 245, 252, 256-257, 284-285, 318, 428, 496, 509, 553, 641, 649 Archiving function, 486 Arm-mounted (BOOM) display, 652 Army-Navy Instrumentation Program (ANIP), 7 ARPA, 20-21 ARPAnet, 21-22, 629 Arrangement of options, 474-475 ARTEMIS, 219
Artifact, 29, 42, 45, 95, 242, 246, 251, 255, 305-306, 308, 313, 315, 317, 442, 459, 468, 554, 638, 647, 649 Artificial intelligence (AI), 211, 642 Artificial intelligent agent, 408 Artificial neural networks (ANN), 85-86 Askjeeves, 203 ASP, 690 Assess and diagnose team performance, 99 Assessment, 95, 158, 165, 307, 462 Assessment tools, 223 Assistive technology (AT), 280-281, 645 Associations, 85-86, 158, 294-295 Associative graphs, 158 Asynchronous, 551-552
720
SUBJECT INDEX
Atlantic Packet Satellite Experiment (SATnet), 21 Attachments, 489 Attackers, 621 Attention, 123, 654 Attitudes of users, 310, 315, 660 Attributes, 179, 182-183, 214, 248-249, 252-253, 285, 295, 422 Audible Input/Output, 645 Audience identification, 361, 366-367, 369-371, 373, 380, 382 Audio,135, 178, 251, 273, 279, 454, 461, 551, 653 Audio files, 177 Audio interaction, 652 Audio lag and lead, 149 Audio presentations, 464 Audioconferencing, 459 Audio-visual media, 152, 222-223, 225, 233, 271 Auditability, 431, 613-624 Audition, 134, 273-375 Auditory channel, 135-136, 140, 149 Auditory cues, 142, 151 Auditory display, 56, 65, 140 Auditory icons, 140-143 Auditory information system, 79 Auditory modality, 81, 149, 272, 652 Auditory quality, 136 Auditory signs, 140-141, 143 Auditory symbols, 139 Auditory threshold, 136-138 Auditory-haptic interaction, 149 Auditory-visual interaction, 149 Augmented communications, 531 Augmentation, 642, 645 Augmented reality systems, 57, 647-655 Augmented Virtual City, 257, 260 Augmented virtual facilitation, 258 Authentication, 41, 488, 614, 617, 620-524, 668 Automation, 497-498, 503, 507, 672 Automated data collection, 329 Automated sessions, 332-333 Automatic checking, 688 Automatic relevance feedback, 195 Automatic system interventions, 96 Automatic Teller Machine (ATM), 643 Autonomous learning, 435 AVANTI, 241, 252-257, 264 Avatars, 225-228, 459 A-weight, 137 B
Back function, 37, 39-40, 45, 108-109, 171, 179, 504 Background, 122-124, 274-275, 348, 350 Background colors, 120 Background contrast, 110 Background information, 308 Background noise, 138-140, 143, 391
Backpropagation networks, 86 Back-tracking strategy, 197 Bandwidth, 108, 124, 135, 142, 259, 631-645, 652, 669 Banner advertisements, 117, 463, 488, 516 Banner blindness, 118 BASIC, 24 Batch processing, 17-18 Bayesian models, 433 Behavior Behavioral intention, 576-593 Behavioral control, 212, 573-593 Behavioral measures of user performance, 431 Behavioral research and data collection, 471-490 BellSouth Wireless Data Network, 630 "Below the fold", 112, 118 Benefits of technology, 662 Best practices, 324 Beta version, 330 Between-subjects design, 325 Binary device, 15 Binaural displays, 138 Biometric devices, 243, 249, 455 Blindness, 152, 254, 264, 267-268, 280-281, 523, 671 Blogs, 451 (see also, Web logs) Bluetooth, 635 Bookhouse interface, 408, 415-418, 420-422 Bookmarks, 37-38, 41, 109, 636 Boolean Boolean logic, 410, 511 Boolean operations, 203 Boolean operators, 199 Boolean queries, 126 Borders, 118, 120 Boxes, 350 Braille, 146, 148, 252, 269, 272 Brainstorming, 166 Breadcrumbs, 34, 39, 347, 517 Breadth, 108, 306, 530, 600, 605 Brightness, 123 Broadband, 630-645 Browser 10, 23, 32-33, 34-36, 38, 40, 44-45, 108-109, 113, 115, 119, 123, 152, 197, 252-264, 267, 269, 274, 280-281, 312, 325, 338, 342, 345, 348, 355, 451, 473-474, 476, 478, 480-481, 483, 486, 495, 497, 507, 515, 534-536, 540, 553-555, 560, 563, 596-597, 614, 616-618, 622, 624, 631-632, 636, 639, 649, 688, 699 Browsing, 59, 171, 196-197, 199, 201, 213, 218, 286, 295, 323, 341, 417, 422, 431, 459, 503, 524, 517, 519, 559, 561, 588, 598, 600, 631, 635, 644, 691 Buffer overflow attacks, 616-618 Bugs, 16, 42 Bulletin boards, 100, 125
Business and management science, 31, 165 Business environment, 509 Business intelligence (BI), 497, 499, 504, 511 Business to business (B2B) applications, 360 Business transactions, 503 Business-to-business, 495, 539, 574, 620,663 Business-to-consumer, 495 Button(s), 118, 142, 148, 286, 295, 501, 507, 515, 696, 698 C
C program language, 16, 18, 187, 635 Cable connection, 18 Cache, 124, 487, 641 California Learning Resource Network (CLRN), 443, 452 Card sorting, 73, 286, 342-345, 359, 361-362, 366-367 Carpal tunnel syndrome, 54, 58, 61 Carryover effects, 326 Cascading menus, 110 Cascading style sheets, 699 Case studies, 76, 84-85, 168, 186, 252, 340, 345, 347-348, 352, 354-355, 415, 555 Categorization classification patterns, 330 Category, 181, 187-188 Category of symbols, 142 Category rating methods, 195 Cathode ray tube (CRT), 61-62, 122, 643 CD-ROM, 39, 454, 458 Cellular Cellular digital packet data (CDPD), 633-634 Cellular networks, 629-645 Cellular phone, 52-53, 251, 633, 638-639 Centers for Disease Control and Prevention (CDC), 665 Central mainframe, 45 Central processing unit (CPU), 631 Centripetal structures, 158 CERN (Centre European pour la Recherche Nucleaire), 23, 630 Charts, 119, 124, 164, 171, 214, Chat, 149, 459, 497 Chat logs, 459 Chat rooms, 313 Cheating, 455 Checkboxes, 125-126, 349, 473, 504, 696, 698 Checking-out, 515, 519-523, 524, 682-683 Chronological order, 375, 418 Chunks, 39 Civil Aeronautics Administration (CAA), 6 Claims analysis, 324
SUBJECT INDEX
Classes, 177, 179-184 Class names, 185 Class of errors, 33 Classes of users, 35 Classification, 179 Classification scheme, 416-419, 649 Client-server, 22 Client-server architecture, 551 Client-server environment, 496 Client-server interactions, 614-624 Cluster analysis, 94 Clustering, 195, 219 Clustering multiuser interactions, 86 Clustering tasks, 85 Clutter, 128, 502 COBOL (common, business-oriented language), 16 Cocktail party effect, 138 Code dividion multiple access, 634 Coding schemes, 666 Cognition, 92 Cognitive activities, 650 Cognitive agent, 408 Cognitive architecture, 75, 81, 230, 425-437, 670 Cognitive demands, 33, 391 Cognitive disability, 268 Cognitive dissonance, 609 Cognitive engineering, 93, 97 Cognitive ergonomics, 71-72, 87 Cognitive impairments, 671 Cognitive maps, 166 Cognitive model, 81, 83-84, 424-437 Cognitive modeling, 85, 670 Cognitive needs and capabilities, 57, 670-671 Cognitive performance, 425 Cognitive processes, 85, 91, 408, 411, Cognitive profile, 688 Cognitive skills, 530 Cognitive strategies, 437 Cognitive structures, 90, 196-197 Cognitive system, 79, 427 Cognitive systems engineering (CSE), 408, 411-415, 420, 422 Cognitive task analysis (CTA), 74-75, 385, 387, 389, 396-405, 460 Cognitive theories, 324 Cognitive walkthrough, 323-325, 327, 426 ColdFusion, 690 Collaboration, 157, 161, 165-166, 169, 172, 212-213, 231, 451, 457, 460-461, 465, 467, 498, 531-32, 534, 536-537, 540, 543-546, 551-569, 629, 648-649, 661, 664, 670-673, 691 Collaboration suites, 497 Collaborative activities, 445 Collaborative applications, 551-569 Collaborative concept mapping, 167 Collaborative design process, 167 Collaborative environments, 648 Collaborative filtering, 691
Collaborative or cliche-based findings, 430 Collaborative requirements analysis,
354 Collaborative system, 551-569 Collaborative technologies, 459 Collaborative thinking, 90 Collaborative work, 671 Collaborative writing, 97-98 Collective and holistic perspectives, 92-93 Collective symbolic approach, 98 Color, 36, 123-124, 128, 203, 274, 279, 285-287, 293-295, 318, 349-350, 515, 524, 642, 644 Colorblindness, 123-124, 274, 463 Color contrast, 122 Color displays, 123 Color spectrum 123 Command language grammar (CLG), 78, 80 Commercial advertising, 489 Committee on Aviation Psychology, 4 Common Gateway Interface (CGI), 39, 473, 690 CGI scripts, 483-485, 618-619 CGI files, 481 Communication, 14, 25, 39, 52, 62, 80, 94-96, 99-100, 128, 136, 138, 150, 167, 171, 176, 212, 216, 224, 226, 228, 246, 256, 258-259, 284-287, 293, 295-296, 308, 313, 318, 323, 341, 374, 396, 412-413, 416, 434-437, 441-442, 445-447, 450-451, 458-462, 464-465, 467, 472, 478, 487, 489, 497, 499, 502-503, 514, 523, 528-548, 551-569, 574, 598-601, 603, 605, 614, 619, 632, 638, 641, 647, 660-662, 664-665, 671, 679-680, 683 Communication device, 13 Communication environment, 98 Communication infrastructure, 251, 531-548 Communication media, 444 Communication services, 566 Communicative agents, 229-231 Communicative behaviors, 213 Communicative functions, 230 Communicative intent, 229 Company logo, 117-118 Comparison, 196 Compatibility, 72, 149, 240, 456, 535, 538,539,581,662 Compatible languages, 479 Compiler, 16-17 Complexity Complex databases, 408-422, 502 Complex environments, 92 Components of the Internet, 14 Comprehension, 110, 112, 114 Computable functions, 218 Computational approaches, 228
721
Computational architectures, 29 Computational cognitive models, 78, 81 Computer, 11, 13-26, 30, 179 Computer development, 15 Computer display, 147, 163, 276, 647 Computer environment, 147, 149, 151 Computer Fraud and Abuse Act, 26 Computer interface, 52, 424 Computer language, 536 Computer network, 21, 169 Computer parts, 18-19 Computer science, 31, 160, 613 Computer software usability, 321 Computer software, 323 Computer supported collaborative work(CSCW), 32, 90, 94, 551-569, 671 Computer system, 33, 80, 87, 630 Computer terminal, 17-18 Computer-base applications, 253 Computer-based learning, 661 Computer-based systems, 689 Computer-based training (CBT), 455-469 Computer-mediated communication (CMC), 97, 100, 490, 671 Concepts, 157-159, 162-163, 184, 529 Concept map, 162, 157-166, 169, 172 Concept mapping, 168, 171 Concept sorting, 509 Concept testing, 307 Concept testing focus group, 307 Conceptual abstractions, 415 Conceptual design, 505, 692 Conceptual framework, 547, 608 Conceptual graphs, 158 Conceptual knowledge, 417 Conceptual model design, 80, 342, 345-354 Conceptual modeling, 505 Conceptual parsing, 422 Conceptual Structure, 94, 171 Conciseness, 644 Concreteness, 650 Conferencing, 100, 451, 459-460, 468, 551-569 Confidence level of the results, 326 Confidentiality, 489, 621, 660, 668 Connection speed, 108, 515 Connectivity, 632-645 Consistency, 80, 113, 123, 128, 219, 229, 272, 279, 346, 349, 350, 353, 396, 398, 400-401, 435, 443, 450, 456, 461, 465, 501, 503, 508, 514, 552, 558, 560-561, 563, 568, 576, 590, 602-604, 609, 669, 689, 692, 694 Constraints, 74, 151, 184, 294, 310, 322, 352, 387, 391, 409, 413-415, 421, 424, 427, 435, 457-458, 460, 468, 503, 506, 510-511, 584, 614, 632, 642, 659, 670-671 Constrained conditions, 267-281, 463 Constrained predictions, 427 Constructs, 530
722
SUBJECT INDEX
Consumer behavior, 595-609, 661 Consumer decision process (CDP), 604, 609 Consumer expectations, 661 Consumer experiences, 598 Consumer health, 663 Consumer needs, 660 Consumer's perception, 604 Consumer-to-consumer interactions, 495 Content, 38, 41, 126, 176, 178, 180, 185, 241, 246, 267-281, 310, 348-349, 352, 355, 360, 362, 377-378, 380-381, 396, 416, 419, 434, 448, 450, 461, 465-467, 469, 480, 495, 517, 533, 538, 540, 546, 552, 554-555, 561, 564, 567-568, 603, 615, 618, 631, 634, 636, 644-645, 649, 653, 661, 668-699 Content analysis, 331-332, 394 Content architecture, 361, 380 Content delivery and engagement, 441-452 Content display, 495 Content management, 564 Content organization, 442 Content profile, 379 Content transactions, 447 Context, 39, 53, 135, 141, 150, 159-160, 168, 211, 216, 240, 242, 248-249, 251-252, 254-264, 279, 286, 288-289, 295-296, 308, 316, 339, 348-349, 351, 359, 374, 387, 393, 409, 413, 427, 454, 457, 460, 465, 475, 495-497, 499-501, 507-509, 511, 535, 537-539, 545, 555, 558, 560, 562, 567, 575-576, 579, 581, 584, 588-589, 595, 601, 607-608, 621, 647-649, 663, 670-671, 688, 693, 695 Context analysis, 360, 365, 368, 371, 377, 379 Context information, 347 Context menu, 563 Context scenario, 375-377 Context-based search system, 204 Context-dependent, 95 Context-related parameters, 262 Context-sensitive, 128, 246 Context-specific information, 253, 433 Contextual constraints, 93 Contextual design, 313, 329, 670 Contextual information, 239, 289, 530, 640 Contextual inquiry, 329, 348, 371, 509, 515 Contextual navigation, 507 Contextual observations, 303, 306, 311-313 Contextual search, 199 Contextual theories, 97 Contrast, 123-124, 275, 515 Control of layout, 36
Control order, 9 Control/display, 4, 393-394 Control-display (C-D) gain, 59-60 Control-display principles, 9 Controls, 107, 352, 396, 410, 507, 546, 560, 645, 648 Conversational agents, 220 Conversional interfaces, 62 Conversions, 291 Conveying meaning, 158 Cookies, 41, 45, 614-624, 636-637 Cooperative evaluation, 324 Coordination, 463, 468, 546, 555, 661-662 Core user attitude, 575 Core user tasks, 323, 360-361, 368 Correlation, 100, 474, 576, 579-581, 642 Cost effective, 389, 471, 544 Cost justification, 679-687 Costs and benefits, 543-544, 605 Cost-benefit analysis, 680-687 Cost-benefit ratio, 326 Course environment, 465 Course management systems (CMSs), 441, 446, 450, 452 Credibility, 199, 539, 602-603, 683 Criterion, 426 Critical decision method, 74, 168 Critical incidents, 330, 333, 387, 393-394 Critical incident reports, 509 Critical incident technique (CIT), 390, 392-396, 404 Critical path method (CPM), 403 Critical path method GOMS (CPM-GOMS), 397, 401-403 Critical tasks, 405 Cross-cultural design, 41, 234, 284-297, 311 Cross-links, 159, 162-163, 166 Crossmodal effects, 149 Cross-platform, 100 Cryptography, 637 Cultural aspects, 312, 524, 553, 666-667, 669, 671, 673 Cultural background, 253 Cultural characteristics, 284-297 Cultural differences, 457 Cultural environment, 318 Cultural knowledge, 165 Cultural preferences, 296 Cumulative trauma musculoskeletal injury, 54 Customer Customer experience, 595-609 Customer frustration, 682-683 Customer preference, 596-609 Customer profiles, 360 Customer relationship management (CRM), 497-499, 502-503 Customer satisfaction, 682 Customer service, 523 Customers' perceptions, 578
Customizable, 499, 635, 642, 663 Customization, 556-557, 560, 562-565, 567, 662 Cybernetics, 26 D
DAML + OIL, 179, 183-184 DARPA agent markup language (DAML), 179 Data, 187, 196 Data accessibility, 613-624 Data accuracy, 501 Data analysis, 15, 466, 568 Data availability, 621 Data collection, 93, 466, 510 Data confidentiality, 621 Data entry, 147 Data integrity, 613-624, 668 Data management, 499, 665 Data mining, 211 Data sharing, 660 Data storage, 187 Data structure, 620, 629 Data types, 163 Database search engines, 518 Database, 43-45, 176-177, 186-188, 193, 200, 202, 206, 211, 213, 304-305, 339, 352-353, 408, 417, 442-443, 447, 496-497, 499-501, 504, 506-508, 515, 524, 551-552, 555, 616, 662-664 Date restrictions, 203 Deaf, 268 Deaf-blind, 269 Debriefing, 309 Debugging, 16 Decision, 84, 574 Decision aiding, 165, 168 Decision making, 75, 79, 86, 94-95, 100, 166, 170, 252, 256, 262, 403, 499, 533, 540, 591, 597, 600-601, 609, 664, 670 Decision process, 595, 601, 604 Decision schemes, 92 Declarative knowledge, 81, 84, 94 Declarative memory, 425, 428-429 Decomposition, 412-413 Deduction search, 168 Deep structure, 39, 414 Defense advanced research projects agency (DARPA), 159 Denning parameters, 80 Definitions), 166, 529, 573, 661, 663, 683, 689 Delays in system, 108, 151 Delivery systems, 659 Demographics, 261, 305, 309, 315, 430, 457, 463, 468, 476, 519, 595-597, 660 Denial of service, 613-624 Department of Defense, 6
SUBJECT INDEX
Department of Health and Human Services (DHHS), 660 Depth versus breadth, 107-108, 341, 600,605 Descriptive methods, 94 Design, 33-34, 96,135, 167, 249, 324, 647 Design and Consulting Services, 7 Design artifacts, 241 Design attributes, 120 Design concepts, 168 Design constraints, 29, 253 Design domain, 692 Design environments, 241 Design-evaluation feedback, 245 Design goals, 29, 255 Design guidelines, 34, 63, 65-66,118-119, 120,140, 142-144, 162, 285, 296, 324, 391, 609, 644, Design implications, 93 Design interface, 139 Design interventions, 29 Design issues, 52, 141, 695 Design life cycle, 245, 252, 673 Design of sounds, 140 Design principles, 10, 34, 40, 48, 206, 501, 585, 651 Design process, 32, 77, 87,140, 245, 323, 351, 353, 357-382, 386, 389, 514, 553 Design products, 42 Design recommendation, 100,135, 206, 243, 311, 314, 393, 597-599, 601-602, 604-606 Design requirements, 248, 252 Design specification, 362, 692 Design standards, 340, 446 Desktop computer, 630-631 Desktop environment, 128 Desktop interface, 540 Development life cycle, 305, 323, 648, 680, 694 Development process, 681 Diagnosis, 95 Diagnostic findings, 100 Diagram, 157, 169, 158, 160-161, 166 Dialog box, 345, 353, 412, 506-507, 622-623 Dialogue (syntactic), 35 Dial-up connection, 630 Diary studies, 330-332 Dichotomous judgment methods, 195 Dictionary, 204, 278 Digital Digital cameras, 451 Digital communication, 629-645 Digital content, 442 Digital environment, 452, 604 Digital imaging, 287 Digital library, 442 Digital resources, 449 Digital subscriber line (DSL), 630 Dimensions, 418
Directories, 193, 203, 295, 536 Disability, 240, 246, 251-252, 257, 264, 267-281, 268, 276-279, 281, 393, 456, 463-464, 515, 523-524, 644-645, 668, 691 Discomfort, 55, 63 Discount usability engineering, 325 Discriminability, 142, 145, 151 Display(s), 61-63, 113, 187, 285,410, 510, 643,652, 654 Display augmentation, 9 Display capabilities, 338 Display content, 451 Display design, 86 Display information, 648 Display integration, 9 Display of content, 304 Display of results, 215 Display resolution, 515 Display screen, 290 Display sizes, 515 Distance collaboration, 170-171 Distance education, 454, 461 Distance learning, 157,161,169, 171-172, 454-469, 465 Distributed learning, 538 Document management, 566 Document sharing, 551 Document skeleton, 180 Documentation, 36,169, 177-179 Domain, 161, 167, 177, 239, 249, 411, 415, 419-420, 497, 502, 511, 692 Domain application, 98 Domain expert, 160, 164,168, 360, 362, 421 Domain knowledge, 160,169, 171, 196, 433, 435, 509 Domain name, 21-22 Domain ontologies, 170 Domain tab, 182 Domain-specific search agents, 232 Work domain, 413, 422,670 DOS, 24 Downloads, 37,112, 118,124-125,142, 149, 169, 187, 442-443, 449, 458, 473, 480, 483, 487, 489, 514, 614-615, 622, 624, 636 Download speed, 322 Download time, 125, 329, 515-516, 519 Drawn images, 124 Driving simulator, 147 Drop-down menus, 110, 125, 475, 636 Drop-off rates, 682-683 DURESS, 413-415 Dyads, 92 Dynamic, 135, 142, 160, 504 Dynamic content, 636,641 Dynamic presentation, 652 Dynamic queries, 200
723
E
Earcons, 140-143 Ease of use, 448, 578-581, 583-584, 586, 681,683 EBCDIC, Extended Binary Coded Decimal Interchange Code, 23 E-Business, 251, 357-382 Ecological approach, 409-422 Ecological aspect, 419 Ecological context, 422 Ecological interface design, 75-76, 408-422 Ecological perspective, 93, 408-422 Ecological system, 529 Ecology, 409, 411,422 E-commerce user interfaces, 514-524 E-commerce Web sites, 21, 29, 33, 41-42, 44, 71, 74, 86-87, 126-127, 176, 206, 321, 324, 341, 343, 347, 349, 356, 391-392, 398, 405, 455, 495-496, 502, 504, 514-524, 582, 595-609, 638, 662, 664, 669, 673, 679, 691 E-communications, 251 Economic, 669, 673 Economic benefits, 645 Economic transactions, 590 Editing data, 187 Education, 157, 165, 253, 638 Educational applications, 499 Educational collection, 442 Educational contexts, 169, 441-452 Educational environments, 451 Educational objectives, 434 Educational resource sites, 441-452 Educational technologies, 505 EEG (electro-encephalogram), 431 Effectiveness, 135, 140, 170, 194, 213, 264, 309, 359, 362, 364, 394, 410, 443, 448, 450, 454, 462, 465-467, 469, 498, 501, 529, 531, 534, 554, 556, 558, 567, 598-599, 603, 606-607, 609, 617, 650, 662-663, 672, 682, 689 Efficiency, 168, 264, 358, 446, 468, 524, 531, 547, 650 E-health, 658-673 Eisenhower National Clearinghouse, 445 Elderly people, 137, 246, 487, 645, 660 (see also, aging) E-learning, 148, 251, 454-469 (see also, distance education) Electroluminescent display, 61-62 Electronic Electronic commerce, 212, 539 Electronic data interchange (EDI), 665, 669 Electronic health information, 660 Electronic junk mail, 487 Electronic media, 575, 589
724
SUBJECT INDEX
Electronic medical records (EMRs), 662, 663, 670, 672 Electronic transactions, 613-624 E-mail, 13, 21-23, 30, 36, 41-42, 45, 100, 139, 185, 313, 323-324, 342, 345, 374, 396, 450, 452, 455, 459-460, 466, 472-473, 486, 487-488, 489, 497, 507, 519, 521, 523, 532, 536, 544, 551-552, 554, 559, 566-567, 580, 590, 598, 603-604, 636, 664, 667, 672, 682, 692, Emergent properties, 93, 410-411, 416, 420 Emotion(s), 212, 228 Emotional communication, 98 Emotional impact, 143 Empirical search, 417 Encryption, 617, 619-620, 622, 640, 668 End user, 500, 554, 568, 574, 688 (see also, users) Engineering psychology, 3, 5, 11 Engineering, 136, 241, 673 Enterprise resource planning (ERP), 497-499, 504 Entertainment, 147, 149, 251, 638 Entry fields, 515 Environment, 86, 137, 144, 178, 213, 225, 231, 245-246, 251-254, 262, 267-268, 275, 286, 288, 305-306, 308, 311-312, 318, 393, 408-412, 422, 427, 431, 434-435, 444, 461, 463, 498, 500, 529, 543, 546, 562-564, 567, 574, 579, 584-585, 590, 603, 605, 632, 634, 641, 649, 653-654, 670-671, 689-690 Environmental changes, 670 Environmental conditions, 53, 391 Environmental constraints, 76 Equivalence relation, 183 Equivalent class, 183 Equivalent property, 183 Ergonomics, 3, 31, 52, 95, 542, 654 Ergonomic algorithms, 689-699 Ergonomic chair, 64-65 Ergonomic criteria, 324 Ergonomic guidelines, 64 Ergonomic principles, 8, 11 Ergonomic tables, 64 Errors, 4-5, 96, 110, 113, 128, 202, 204-205, 289, 310, 313-314, 321, 379, 380, 394, 411, 425, 443, 474, 476, 480-481, 486, 489, 500, 510, 515, 521, 529, 531, 546, 620-622, 634, 637, 639, 662-663, 669, 670, 672-673, 680, 682, 689 Error messages, 128, 333, 354 Error recovery, 277 Estimating costs, 681, 687 Ethics, 334, 488-489, 640, 668 Ethnography, 94-95, 286, 332 Ethnographic methods, 329 Ethnographic studies, 334, 509 Ethnographic work, 553
E-trade, 149 E-training, 540 Eudora Web, 636 Eureka reports, 330 Europa, 443 European Union, 693 Evaluation, 71, 249, 322, 340, 361-362, 366-367, 369-371, 373, 380, 386, 388, 461-462, 465-467, 548, 563-564, 579, 603, 605, 607-609, 647, 650, 670, 698 Evaluating business web applications, 506, 510 Evaluating web usability, 333, 509 Evaluation iterations, 350 Evaluation metric, 96 Evaluation process, 142 Evolving technologies, 411 Executable content languages (XCLs), 614 Execution context, 247 Executive Process-Interactive Control (EPIC) architecture, EPIC, 78, 81, 425,427-428, 670 Expanding and contracting outlines, 110 Expectation, 116-117, 141, 375, 377, 379, 555, 578 Expectation-Disconfirmation Model, 575, 577-578 Experience, 123, 196, 251, 257, 305, 568, 578, 587, 592, 596, 649, 695 Experiments, 471-490 Expert(s), 167, 171, 251, 257, 357, 359, 373-374, 413, 422, 429, 443-444, 456, 461, 503, 505, 531, 547, 589, 602, 632, 636, 642, 663, 680, 683, 688 Expert and customer reviews, 519 Expert knowledge, 172, 434 Expert inspections, 380 Expert interview, 650 Expert problem solver, 433 Expert review, 510, 515 Expert systems, 169, 171 Expert user, 426, 466 Expertise, 195, 246, 253, 254, 264, 305, 312, 315, 530, 546, 602, Explicit knowledge, 530 Exploration, 112 Exploratory learning, 330, 390 Extended Task-Action Grammar (ETAG), 78, 80, 83 Extensible markup language (XML), 26, 35, 177, 180, 187, 214, 216, 229, 231, 261-262, 451, 495, 615, 622, 644, 668-669, 690 Extensible Stylesheet Language Transformation, 35 External links, 117 Extranets, 662 Eye movements, 404, 425-426, 431 Eye-hand coordination, 59-60 Eye-tracking deices, 61, 118
F
Familiarity, 148 Fatigue, 54, 146, 654 FAX, 454 Federal Aviation Administration, 388 Feedback, 43, 54, 107, 123, 128-129, 135, 144, 146, 148-149, 151-152, 166, 196, 200, 205, 216, 223, 227-228, 277, 339, 365, 410, 415, 431, 456, 459-460, 462-466, 468-504, 508, 510, 515, 529, 548, 555, 564, 578, 591, 652, 690 Feedback sounds, 140-141 Field labels, 121 Field level help, 127 Field methods, 323, 329 Field of view (FOV), 652 Field studies, 329, 331, 502 Field trials, 330 Field of search, 198 Files, 142 File format, 203 File sharing, 497 File transfer protocol (FTP), 21-22, 24, 473, 544, 629, 636 Filters and options, 203 Financial transactions, 602 Firewall, 515, 534, 539, 624 First-time users, 444 Pitts'law, 31,425, 639 Fixed size, 36, 38 Fixed versus fluid layout, 113, 115-116 Flash, 32, 36, 38, 277, 345, 516, 524, 690-691 Flash applications, 147 Flicker, 122, 276-277 Flight simulators, 147, 149 Flow charts, 74 Flow diagrams, 404 Fly-out menus, 110 Focus groups, 303, 306-307, 323, 331-332, 387, 509, 515 Focus question, 161 Font, 120, 122, 134, 203, 279, 287, 349, 523,644 Font size, 36, 120-121, 290, 292, 324, 524, 691 Font style, 121,292,524 Foot-operated devices, 61 Force, 146 Force feedback, 147 Forced-choice judgment, 151 Form controls, 125 Formalization of requirements, 32 Format, 128, 274-281, 289, 296 Formative evaluation, 35 FORTRAN (Formula translation), 16-18 Forward function, 197 Frame of reference, 394, 649 Frames, 37, 39, 107, 109-110, 117, 38, 345, 347-348, 637, 644, 649, 652, 653
SUBJECT INDEX
Framework, 35, 97, 143, 158, 194, 254, 256-258, 262, 264, 358-382, 427, 529, 578, 582, 595, 604, 638, 649, 697 Framing, 603 Frequency, 137, 151, 293, 388, 467, 475, 499-500, 509, 551, 568, 590, 596, 608, 632 Frequency distribution, 475 Frequency of use, 290 Frequency spectra, 141 Frequency spectrum, 143 Frequently asked questions (FAQs), 36, 127-128, 324, 465 Frequently used data, 500 Frequently used functions, 507 FRIEND21, 241 Frustration, 588, 592 Function, 378, 380 Function errors, 379 Function identification, 388 Functional capacities, 538 Functional goals, 415, 417 Functional properties, 183, 245 Functionality (semantic), 35 G
Games, 149 Gaze direction, 431 General guidelines, 149 General packet radio system (GPRS), 634 General principles, 11 General recommendations, 392 General rules, 434 General site help, 127 General-purpose browsing, 197 Genetic algorithms, 85-86 Genetics Based Policy Capturing (GBPC), 87 Genre of UI, 693 Georgia Board of Regents, 448 Gestures, 294 Glare, 55-56 Global nature, 504 Global positioning system (GPS), 261, 638 Global system for mobile communications (GSM), 261, 630-631, 634 Glove, 147, 152 Goal(s), 33, 71-77, 49, 83, 91, 93, 135, 148, 166, 168, 177, 197, 255-226, 228-229, 241, 247-248, 253, 275, 303, 305-307, 309, 313, 315, 317-318, 321, 322-323, 353, 360-362, 364, 370, 388, 390-393, 396, 401, 402, 405, 409, 411, 414-415, 422, 424, 426, 428-431, 449, 465-466, 502, 505, 528-529, 534, 537-538, 540, 545, 551-552,
554, 556, 567, 581, 595, 597, 599, 601, 605, 608, 647, 658, 671, 680-687, 689, Goal-directed users, 605 Goal structures, 396 Goals Operators Methods and Selection Rules (GOMS), 72, 74, 78-79, 81, 83, 87, 323, 390, 392, 396-405, 398, 400-401, 403-404, 426-427, 670 GOMS analysis, 323 Google, 24, 177, 200-203, 487, 508, 574 Gopher, 23-24, 629 Government, 165 GPS, 640-641 Graphics, 34, 125, 113, 118, 124, 135, 163, 213, 251, 263, 271, 276, 279, 285, 287, 294, 454, 463, 471, 507, 511, 516, 524, 551, 600, 607, 615, 644, 651, 689, 692, 698 Graphic design, 36, 293, 296 Graphic imagery, 42, 507 Graphic images, 507 Graphic interface, 619-620 Graphic tablet, 152 Graphical interface, 44, 158, 273, 377, 415, 420, 624 Graphical and textual site maps, 126 Graphical browsers, 169, 449 Graphical displays, 160 Graphical images, 118, 297 Graphical representations, 417, 462 Graphical user interaction, 230 Graphically displays, 200 Graphical user interface (GUI), 19, 24, 31, 35, 37, 42, 121, 218, 289, 294, 323, 332, 338-339, 342, 347, 504, 349-350, 352, 506, 536, 559-560, 562-563, 568, 618, 622, 650, 688-699 GUI applications, 508, 697 GUI guidelines, 688-699 GUI-base browsers, 23-24 Graphics interchange format (GIF), 124 Graphs, 107, 119-120, 124 Grayscale, 124 "Greeked" Web page, 113-114 Group, 431, 555 Group decision support system (GDSS), 90, 93 Group dynamics, 91 Group size, 98 Group task analysis, 509 Grouping of signs, 141 Groupware, 90, 97-98, 100 Guideline(s), 118, 125, 252, 272, 281, 293, 297, 324, 338, 350, 352-353, 377, 400, 471, 502-504, 512, 545-546, 551, 564, 568, 575, 579, 597, 629, 644-645, 668-699 Guideline definition language, 697 Guideline for the use of laptop computers, 64 Guidelines for media selection, 150
725
Guidelines for the use of graphics, 124 Guidelines for UI design, 501-509 Guidelines for Web collaboration, 564-569 Guidelines on audio-visual synchrony, 150 H
Hacker, 613-615 Handspring Blazer, 636 Handwriting recognition, 635, 640 Haptics, 511,562-653 Haptic device, 144, 146-148, 151-52 Haptic displays, 146-147, 653 Haptic interactions, 151 Haptic interfaces, 62, 146 Haptic metaphor designs, 148 Haptic perception, 148 Haptic stimuli, 57 Hardware, 135, 149, 653 Hardware-software compatibility, 170 Headers, 39, 523, 697 Headings, 506 Head-mounted display, 56-57, 214, 652 Heads-up displays, 56 Health, 638 Health care, 658-673 Health care applications, 499 Health care costs, 663 Health care system, 658-673 Healthcare information systems, 659 Health Insurance Portability and Accountability Act (HIPPAA), 660-661, 666, 668-669 Health quality, 658-673 Hearing, 138, 268 Help, 117, 127,444 Help interfaces, 109 Help systems, 36 Heterogeneity, 35, 91, 97-98, 488 Heterogeneous, group, 421 Heterogeneous information, 251 Heterogeneous data, 220 Heterogeneous information, 212 Heterogeneous samples, 471 Heterogeneous teams, 91, 95-96 Heuristics, 33, 86, 159, 162, 358, 546, 605, 640, 644, 672 Heuristic evaluation, 35, 323-324, 327, 348, 351, 354, 505, 510 Heuristic search, 81 Hierarchy, 108 Hierarchical classification, 162 Hierarchical levels, 166, 541, 567 Hierarchical means-ends relations, 414 Hierarchical order, 534 Hierarchical structure, 39, 126, 157, 159, 178, 188, 193, 200-201, 214, 219, 229, 277, 286, 295, 329, 341-343, 413, 419, 426, 431, 497-498, 500, 531, 533, 543, 565, 600, 622
726
SUBJECT INDEX
Hierarchical task analysis (HTA), 33, 74, 390-392, 396, 400, 404-405, 670 Hierarchy of constraints, 413 Hierarchy of subtasks, 73 High functionality application (HFA), 415 High-contrast, 122 High-dimension space, 40 Higher education, 441, 454-469 High-fidelity displays, 652 High-resolution, 645 High-risk environments, 96 High-speed connections, 20 History, 39-40, 45, 567 History of computers, 13-26 History of Internet, 13-26 Holistic approach, 94-95, 97-100 Holistic measures, 99 Holistic outcome, 92 Home function, 197 Home link, 117 Homepage, 113, 126, 178, 188, 197, 287, 486, 504, 515-517, 539, 544, 555-557, 596, 615, 617, 681, 692 Homogeneity, 91, 97-98, 306, 417, 430 Horizontal layout, 110 Horizontal scrolling, 113 (see also, scrolling) Hotmail, 636 Human associative memory (HAM), 81 Human capabilities, 26, 83, 544 Human cognition, 29, 31 Human computer interaction (HCI), 30, 511,29-30,33,39,45,243, 247-249, 511, 618, 620, 622, 624, 658, 669-673, 688, 695 (see also, human-machine interaction) Human engineering, 4-8 Human factors, 3, 5-7, 11, 24-25, 90-91, 94-95, 100, 107, 121, 146, 177, 186, 189, 386, 388-391, 411, 422, 425, 441, 475, 478, 480, 495-499, 501, 503-507, 509-510, 537, 613, 617-623, 619-620, 622, 624, 629, 658-659, 669-673, 679-687, 683, 687, 690 Human factors and ergonomic guidelines, 240 Human factors and ergonomics, 334 Human Factors and Ergonomics Society (HFES), 6, 63, 693 Human factors intervention, 687 Human factors principles, 502 Human factors professionals, 3-4, 31, 167, 324, 327, 385-386, 679 Human information processing, 53, 79, 86, 295, 408, 425, 670, 672 Human judgment, 202 Human perception, 4, 28, 31, 604, 692 Human perceptual limitation, 297 Human performance, 3-4, 10, 28, 56, 58, 81, 83, 86, 388-389, 422, 427, 430, 652
Human processing capabilities, 39, 670 Human resource (HR), 498 Human subjects, 334 Human-centered design, 303, 360, 548 Human-character interaction communication, 225 Human-computer interaction (HCI), 28, 52, 57, 76, 83, 86-87, 92, 241, 284, 393, 397, 413 (see also, Human machine interaction) Human-computer interface, 424, 501, Human-machine interaction, 152, 213, 437, 673 (see also, Human computer interaction) Human-readable form, 176 Hybrid architectures, 428 Hybrid model, 85 Hybrid user interface, 645 Hyperlinks, 23, 515-517, 523, 536, 559-560, 568 Hyperlink multimedia, 169 Hyperlinked information, 157, 164, 176, 187, 324 Hypermedia, 169, 257, 264, 431, 434, 534, 648, 692 Hypermedia system, 197, 253 Hypertext, 39-40, 73, 118, 251, 286, 545, 692, 696 Hypertext editing systems (HES), 23 Hypertext links (see Hyperlinks) Hypertext markup language (HTML), 2, 23, 35, 42, 44, 109, 121-122, 125, 127, 152, 169, 176-177, 185-186, 188, 203, 252, 257, 269, 271, 274, 280, 345, 441, 443, 472, 476, 478-480, 485-487, 489, 495-496, 498-499, 501, 503-507, 510-511, 515, 534, 556, 614-615, 617, 619, 630, 634, 636, 688, 690 HTTP, 22-24, 41, 441, 483, 534, 536, 614, 630, 636, 641 Hyphenation, 127
IBM, 639, 692, 696 Icon, 37, 124, 139, 147-148, 163-164, 169, 171, 285, 290, 296-294, 349, 353, 403, 417-419, 479, 545, 559-560, 640 Icon identification, 425 Iconic level, 141 Iconic representation, 419 Iconic sign, 139 Iconography, 287 Ideas, 158 Identification, 41, 148, 360, 375, 621 IIS, 618-619 HS Lockdown, 624 Image(s), 122, 124, 157, 164, 171, 177-178, 181, 185, 196, 200, 203, 214, 218, 240, 273, 280, 293-294,
349, 487, 515, 519, 524, 597, 601, 608, 615, 636, 644, 652, 691-692 Image maps, 110 Image polarity, 121-122 Image recognition, 433 Image retrieval, 431 IMG tags, 125 IMAP4, 23 Immersive displays, 652-653 Implicit communication, 464 Implicit knowledge, 530, 542 Implicit learning, 428 IMS Global Learning Consortium, 441 Inconsistency, 114 Index, 139, 176, 193, 197 Index menus, 110 Individual team cognition, 92 Individual's perceptions, 575 Industry structures, 668 Inefficiency, 289, 553 InfoQuilt, 218 Information, 74, 172, 176, 178, 193-194, 223, 225, 239, 274, 308, 539, 543, 599, 647, 661, 663-664, 673 Information acquisition and retrieval, 434 Information architecture, 286, 295-297, 339, 341-346, 352-354, 359, 365-378, 447-448, 596, 601, 690 Information availability, 567 Information content and structure, 252 Information design, 509 Information display, 500 Information dissemination, 574, 663, 665, 690 Information division, 114 Information flow, 528-548 Information infrastructure, 659 Information integration, 220 Information management, 544, 562, 566-567 Information presentation, 431-432, 546, 600, 608 Information processing, 53, 75, 85, 92-93, 388, 604 Information processing models, 410 Information processing theory, 303 Information quality, 601 Information quantity, 107 Information representation, 603 Information requirements, 167, 392 Information retrieval, 193-207, 213, 258, 415, 538, 540, 543, 544, 567, 617, 670 Information retrieval systems, 194-196, 253 Information search, 597, 599, 603, 605-606, 609 Information security, 542, 613-624, 669 Information sharing, 157, 539, 546, 664-665
SUBJECT INDEX
Information society, 70, 240-241, 251-264 Information storage, 566 Information structure, 176, 198, 543 Information system design, 409 Information systems, 661, 672, 677 Information technology (IT), 52, 166, 441, 451, 455-456, 523, 528-529, 531, 534, 537, 540, 544-545, 553, 574, 583, 613, 659, 661-662, 666, 686 Information technology system, 577, 542 Information theory, 26 Information type, 600 Information visualization, 662 Infrastructure, 168-169, 261, 448, 530, 536, 539-540, 542, 546, 662, 669 Injury, 55 Input channel, 136, 146 Input device, 45, 57, 60, 65, 144, 147, 473, 477, 642, Input interface, 639 Input technologies, 642 Input/output, 147, 252-253, 511, 632, 642 Inspection, 324, 361 Instant messaging, 451, 459, 555, 566-567, 639 Institute of Electrical & Electronics Engineers (IEEE), 630, 633, 641 Institute of Medicine, 672 Institutional review boards (IRBs), 481 Instructional system, 433 Instructional text, 506 Integration, 507, 509, 563-568, 574 Integrated information, 166 Integrated delivery networks (IDNs), 659-661, 667 Integrated environment, 219 Integrity, 660 Intelligent adaptation, 252 Intelligent agent, 128, 178, 268, 272, 424, 435-436, 606-607, Intelligent information agents, 211-234 Intelligent interfaces, 86, 211-234, 430 Intelligent tutor, 432-434 Intended audience, 683 Intensity, 142-143, 151 Intentions, 411, 413, 416, 576 Interacting network, 13 Interaction, 91, 158, 231, 253, 296, 409, 443, 446-447, 455, 457-459, 462-464, 503, 539, 541, 551, 553, 555-556, 563, 568, 574, 578, 582, 589-590, 596, 598, 601-602, 614-624, 631-632, 638, 644, 647-649, 664-665 Interaction architecture, 361 Interaction design, 34, 502 Interaction pattern cards (ICPs), 232 Interaction pattern, 92, 694 Interaction platform, 241-242
Interactive interfaces and systems, 135, 338, 408, 464, 478, 499, 506, 511, 598-599 Interactive (software) artifacts, 240 Interactive applications, 503, 654 Interactive community, 460, 642 Interactive cycle, 178 Interactive environments, 650 Interactive information, 644 Interactive information agents and interfaces, 211-234 Interactive information retrieval model, 196 Interactive mock-ups, 515 Interactive prototype, 328-329 Interactive qualities, 463 Interactive tasks, 688 Interactive television, 454 Interactive web pages, 393 Interactive Web site, 176, 485 Interdisciplinary approaches, 166, 443 Interface agents, 231 Interface architectures, 43-44 Interface code, 35 Interface, 36, 128-129, 134, 136, 187, 325, 330, 359, 419, 469, 509, 579, 589, 648, 679, 690, 692 Interface development, 57, 253, 325 Interface elements, 280 Interface engines, 661 Interface guidelines, 125 Interface models, 502 Interface technology, 205 Interface testing, 430 Interface usability, 692 Interface design, 31, 41-42, 56-57, 75, 77, 99, 140, 193, 200, 202, 204-206, 341, 344, 359-382, 388-405, 412, 424, 458, 462, 585, 606, 679-687 Interface design guidelines, 128 Interface design principles, 343 Interface designer group, 359 Interface designer, 39, 152, 203 Interface designer/usability engineer, 348 Interlaced image, 124 Internal links, 117 Internal rate of return (IRR), 683, 686-687 International e-business, 357 International Ergonomics Association (IEA), 64 International organizations, 693 International Standards Organization (ISO), ISO, 135, 149, 303, 357-358, 360, 362, 688-699 Internationalization, 41 Internationally, 514 Internet Internet-based communication, 452 Internet-based research, 471-490 Internet browser, 146 Internet communication, 20
727
Internet connection, 18 Internet Explorer, 24, 122, 203, 345, 536, 615, 622, 624 Internet growth, 70 Internet Mail Extensions (MIME), 23 Internet Message Access Protocol (IMAP), 23 Internet protocol (IP), 620, 22, 14 IP address, 20, 312, 476, 485, 614, 634-635 Internet Service Providers (ISPs), 26, 473 Internet services manager (ISM), 619 Interpersonal exchanges, 442 Interpersonal skills and attitudes, 312 Interpretation, 196 Intersensory conflicts, 654 Interview, 32, 73, 86, 94-95, 158, 167-168, 242, 303, 306-309, 311, 313, 317, 323, 327, 329, 331-332, 334, 339, 343, 360, 387, 391, 393-394, 396, 417, 443, 497, 502, 509, 515, 662 Intranet applications, 19, 100, 495, 528-548, 662 Intraorganizational communication, 528-548 Intrusion detection, 621 Invisible images for spacing, 36 IS4ALL, 243, 249 ISDN (Integrated Services Digital Network), 18 ISM (industrial, scientific, and medical) band, 633 Item display, 204 Iteration, 34, 524 Iterative design reviews, 515 Iterative design, 389, 392 Iterative evaluation, 344, 348, 351, 354 Iterative process, 245 Iterative redesign, 638-639 Iterative testing, 325 I
Jargon, 91, 128-129, 289, 524, 695 Java, 22, 36, 188, 345, 478-479, 501, 504, 506, 516, 524, 617, 623, 635-636, 690-691, 697, 699 Java applet, 495-496, 614, 622 Java plug-ins, 495 Java Virtual Machine, 641 JavaScript, 32, 36, 39, 229, 324, 476, 478-480, 524, 614, 636-637, 690 JIET (just in e-time), 357-382 Job performance, 388, 404, 546, 578 Joint photographic experts group (JPEG), 124 Joystick, 59, 60, 61, 144, 147 Judgment, 123, 164, 196, Justification, 121 Just-noticeable difference (JND), 146
728
SUBJECT INDEX
Layout, 36-37, 214, 291, 296, 348, 377, 379-381, 396-397, 421, 500-501 Key sounds, 143 Layout and design of the content, 277 Keyboard, 20, 23, 58, 59, 63, 65, 134, Layout of augmented environment, 136, 144, 147, 276-277, 398, 431, 649 437, 499-500, 524, 631-632, 642 Layout patterns, 421 Chord keyboards, 58 LCD, 632 Dvorak keyboard, 147 Learnability, 142, 426, 462, 558, 650 QWERTY keyboard, 58, 147, 639, 642 Learner characteristics, 449, 461 Smooth top keyboards, 59 Learning, 10, 158, 160, 165-166, 172, Split keyboards, 58 223, 228, 233, 425, 427-428, Virtual keyboard, 642 441-452, 458, 461, 529, 545, 553, Keystroke-level model (KIM), 81, 555, 606, 664, 692 Learning application, 462 398-399, 401 Keyword, 123, 178, 194, 197-198, 200, Learning architecture, 428 Learning communities, 452 202-207, 216, 295, 418 Learning content management systems Kinesthetic, 653 (LCMS), 460 "Know thy user", 417 Knowledge, 81, 99, 158, 165, 167,170, Learning effects, 195 Learning environments, 457 177, 212, 387, 424, 428, 430, Learning goals, 434 432-433, 457, 501, 509, 531, 556, Learning management systems (LMS), 573, 587, 591, 595, 597, 599, 606, 460, 464 618, 661, 664-665, 671, 673, 688 Learning mechanism, 427 Knowledge acquisition, 72-73, 77, 80, Learning object, 445-452 92, 165, 171, 432, 435, 437, 442, Learning rate, 682 463, 529-530, 540, 543, 596-597, Legibility threshold, 292 651 Letter case, 121 Knowledge base, 542 Levels, 261 Knowledge capture, 168-169 Levels of abstraction, 178 Knowledge domain, 163, 619 Levels of expertise, 649 Knowledge elicitation methods, Lexicon, 44, 218 94-95,167, 171 Lexical knowledge, 44 Knowledge elicitation, 95, 98, 165, Lexical level, 253, 695 167-168, 171, 306, 331, 388, 394, Lexical relationships, 219 435, 617 LICAI, 426 Knowledge Elicitation of team Life cycle, 245, 252, 339, 690 cognition, 95 Light emitting diode, 61-62 Knowledge management, 165, 528, Lighting, 55-56 532, 543, 552 Line length, 121 Knowledge management systems, 447 Line spacing, 292 Knowledge preservation, 171-172 Linear data, 214 Knowledge representation, 168, 433 Linear or hierarchical path, 38 Knowledge requirements, 660 Linear structure, 341, 541 Knowledge sharing, 161, 165-166, Linguistic, 31, 290 171, 530, 534, 536, 538, 544, 697 Link(s), 31, 34, 39-41, 107, 117-118, Knowledge structure, 84, 96, 169, 124, 157, 160, 162, 177, 179-180, 401, 531, 547-548 185, 205, 286, 347, 349, 428, 431, 445, 472, 476, 486, 488, 519-520, L 568, 648, 663 Link affordance visibility, 118 Labels, 157, 203, 291, 293, 295, 352, Link analysis, 404 378, 501, 651 Link anchors or terms, 119 Laboratory usability tests, 330 Link destination, 119 Language, 180, 186, 188-189, 203, 241, Link geometry, 40 255, 261-263, 272, 278-279, 284, Link placement, 118 287, 289, 290-293, 296, 309-311, Link titles, 119 318, 442-443, 457, 464-465, 565, Linkage, 377, 379-381 568 Linked information, 38, 117, 163, 171, Multiple languages, 497, 524 177 Local Area Networks (LAN), 26, 43 Links per page, 205 Laptop computers, 60, 498, 630-631 Nontext links, 118 Laser scanning, 642 Unvisited and visited links, 40, Latent semantic analysis, 99 118-119, 126 K
Linksys, 630 Linux, 621 Liquid crystal display, 61-62 List, 112,214,636 List box, 125, 694 List style results, 127 ListServs, 21, 23, 305 Local area networks (LANs), 21, 535, 630-645 Local maxima, 35 Localizing, 138 Location, 515 Location restrictions, 203 Location-aware association, 649 Location-aware augmented reality, 648-649 Log file, 480, 483-484, 509 Logged data, 333 Login/register button, 117 Log-on, 620 LONGDESC tags, 523 Longitudinal study, 158 Loudness, 136, 141 Luminance contrast, 55 Lycos, 201 Lynx, 24 M
Machine learning, 212, 230 Machine readable, 176-177, 269, 274,
435 Machine-understandable, 179 Macintosh, 621 Macromedia, 327, 472, 515, 648, 690, Magnitude estimates, 145, 148, 195 Mailing Lists, 23, 486 Mainframe systems, 14, 25, 661 Maintaining context, 507 Maintenance, 7-8, 322, 392, 501, 506, 553, 565, 568, 624, 660, 664, 666-667, 698 Malicious mobile code attacks, 617 Managed care organizations (MCOs), 659 Management, 166, 213, 565 Management of information, 212 Management strategy, 540 Map(s), 160, 651 (see also, concept maps) Marketing, 497 Markup, 178-179, 274, 277-278 Markup language, 231-233, 495, 630, 634, 636 Multiple markup languages, 636 Masking, 138, 142, Massachusetts Institute of Technology (MIT), 241, 630, 640, 652 Matching, 196-197, 216 Mathematical logic, 194 Mathematical models, 427 Matrix representations, 160 Meaning, 213 Meaningful, 141
SUBJECT INDEX
Measurement, 5, 94-95 Measuring benefits, 681-682, 687 Measuring team cognition, 93 Mechanoreceptor types, 144 Media, 135, 149-150, 164,172, 176, 178, 213, 223, 251, 259, 275, 313, 345-346, 348, 377-378, 380-381, 443, 447, 449-451, 457, 461, 463, 469, 471, 487, 511, 514, 529, 538, 587, 603, 636, 649, 654, 662 Media-enhanced or hybrid courses, 447 Mediated information retrieval, 194 Media richness theory, 97 Medical errors, 658-673 Medicine, 90 Member profiles, 448 Memorability, 650 Memory, 81, 604-606, 616, 619 Long-term memory, 426-427, 600 Working memory, 39, 425-427 Mental model, 72, 80, 318, 339, 341-342, 380, 409, 603, Mental transformations, 161 Mental workload, 81, 388, 650 Menu(s), 24, 35, 37, 39, 112, 142, 147, 286, 293, 295, 342, 346, 412, 427-428, 431, 475, 483, 504, 507, 510, 515, 519, 555, 559-560, 568, 623, 692 Menu configurations, 126 Menu design, 293 Menu headings, 110 Menu hierarchies, 107 Menu search, 425, 427-428 Menu selection, 60 Messaging, 554 Meta analysis, 485, 579, 585 Metacognitive skills, 165 Metadata, 177, 212, 218, 259, 278, 441, 445, 449-450, 555, 566 Metadata standards, 450 Metadata-based search, 216 Metaphor, 149, 287, 318, 417, 534, 649, 692 Metasearch, 206 Metatags, 486-487, 518 Methods 79, 329 Metrics, 100, 687 Microsoft, 231, 472, 618, 624, 636, 692 Midicomputer, 14, 19 Military, 165, 404, 435 Military aviation, 4-5, 8 Military systems, 90, 386, 388 Minesweeping, 118 Minicomputers, 19 Mission control, 96 MMS language, 635 Mobility, 451, 498, 632 Mobile computing, 264 Mobile IP, 634-635, 641 Mobile phone, 28, 34, 257
Mock-up design, 35, 73, 167, 238, 242, 324-325, 327, 340, 342-343, 345, 347-348, 353, 361-362, 366-368, 370, 505 Modalities, 135, 149, 151, 234, 246, 252, 259,272,464,511 Model, 401,670 Model of the user, 85, 247 Model structure, 79 Model tasks, 84 Model-tracing method, 433 Model Human Processor (MHP), 78-79, 81, 402-403, 425-427, 670 Modeling, 10, 71, 78, 151-152, 212, 228, 261, 590, Modem, 18, 23, 487 Moderator, 166 MOMIS (mediator environment for multiple information sources), 218, 220-221,219 Monaural, 138 Monitor sizes, 486 Monochromatic display, 123 Morphology, 158-159 Morse signs, 139 Motion platform, 147-148 Motivations, 177 Motor system, 79 Motor outputs, 84 Motor impaired, 254, 257, 264 Mouse, 20, 24, 59, 60, 144, 147, 152, 203, 275-276, 398, 401-403, 431, 437, Mouth-operated devices, 61 Movement, 277 Movie, 200, 273, 307 Multidimensional, 194 Multidimensional scaling (MDS), 94, 408, 421, 601 MDS-interactive interface, 422 Multidimensional scaling interface, 419-422 Multidimensional space, 650 Multidisciplinary, 245 Multidisciplinary teams, 166-167 Multilayer backpropagation network, 85 Multilingual, 443 Multimedia, 128, 134-135, 144, 148, 152, 168, 171-172, 178, 200, 202, 213, 229, 246, 273, 431, 447, 454-455, 460-461, 535-536, 606, 637, 644, 648, 662, 665 Multimedia Educational Resources for Learning and Online Teaching (MERLOT), 445-449, 452 Multimodal, 135, 146, 151, 224-226, 230,234,249,511,650 Multimodal interaction, 213, 651-652 Multimodal markup language (M3L), 231 Multimodal presentation, 650 Multination, 496 Multiple devices, 498
729
Multiprogramming, 18 Multisensory 149,598-599 Multisensory event, 151 Multisensory interaction, 149 Musculoskeletal discomfort or disorders, 52, 54-55, 59, 64 Music, 200 Musical instruments, 149 Myopic, 643-644 N
Napster phenomenon, 451 Narratives and scenarios, 509 Narrow and deep structures, 39 Narrow connectivity, 631-632 NASA-Task Load Index, 85 National Aeronautics and Space Administration (NASA), 10, 21,165-166, 168, 655 National Learning Infrastructure Initiative, 444 National organizations, 693 National Research Council (NRC), 4, 6, 668 National Science Foundation (NSF), 22, 433, 629, 655 National Science Foundation (NSF) network (NSFnet), 21 Natural GOMS (NGOMS), 72, 85 Natural GOMS language (NGOMSL), 74, 78-81, 84, 397-398, 400-401, 404 Natural language, 177-178, 194-195, 212-213, 222, 224 Natural sound, 139, 142 Naturalistic observation, 73, 311 Naval Air Warfare Center Training System Division (NAWCTSD), 655 Naval Electronics Laboratory (NEL), 5 Naval Research Laboratory (NRL), 5 Navigation, 30-31, 37-40, 61, 86, 107, 109-110, 113, 118, 125-126, 147-148, 169, 171, 178, 196, 205, 223, 234, 255, 269, 275-277, 279, 285, 295, 297, 313, 322, 325, 338-339, 341-343, 349, 351, 396, 414, 416-417, 421, 431, 434, 450, 469, 498, 505-506, 508, 511, 514-518, 540, 555-556, 568, 589, 599, 602, 637, 643, 647-648, 650-651, 689-693, 697, 699 Lost user, 648 Navigation bar, 124, 507-508, 556, 692 Navigation controls, 108, 516-517, 523 Navigation icons, 419 Navigational aids, 651 Navigational links, 110 Navigational or learning tools, 434 Navigational paths, 197,431-432 Navigational structure, 33, 348-349, 545 Navy's Special Devices Center (SDC), 5
730
SUBJECT INDEX
Needs analysis, 376 Needs and preferences of users, 514-515 Net present value (NPV), 683-687 Netscape, 24, 536, 622 Network(s), 21, 24, 157-158, 200, 212, 228-230, 251, 256, 259, 261, 286, 341, 444, 460, 511, 534-536, 538-539, 543, 546, 553, 555, 613-624, 635, 659, 661-662, 665, 670, 683 Network architecture, 304 Network components, 36 Network connection, 345, 449, 629-645 Network Control Protocol (NCP), 22 Network delays, 535 Network diagrams, 34 Network service providers, 18 Network structure, 434 Network visualization, 511 Networked terminals, 239 Network-Solutions Incorporated (NSI), 22 Neural networks, 85-86 Neural representation, 134 Neuronal transmission, 151 New windows, 37 News, 638 Newsgroups, 177, 487-488 Nielsen Norman Group, 505-506 Nodes, 157, 160 NOGAP, 367 Noise, 56 Noncooperative or cooperative information agent, 212 Nondeniability, 613-624 Nonfunctional requirements (NFRs), 245, 247-249 Nonparticipatory observations, 311 (see also, observations) Nonrepudiation, 613-624 Notebook PC, 635 Novice users, 196, 246, 251, 429, 433, 466, 622 Novice/expert difference, 198 Nuclear power plants, 96 Numeric information, 119 O
Object(s), 142, 559, 650 Object coupling, 648 Object profiles, 143, 448 Object-based languages, 44 Observation, 30, 32, 94-95, 163, 220, 308, 313 323, 329, 343, 385, 387, 424, 471, 509 ODLi3, 219, 222 OFDM (orthogonal frequency division multiplexing), 633 Office of Naval Research (ONR), 6, 10, 655 Oklahoma Board of Regents, 448
Older users, 118, 640, 671 (see also, aging) Omnidirectional, 136 Online Online behavior, 595-609 Online collaborative technologies, 454 Online communication, 463 Online context, 578 Online courses, 454-469 Online discussions, 313 Online environment, 446, 580-581, 584, 597, 602 Online forms, 394 Online learning, 454-469 Online population, 70 Online retail, 514 Online sales, 514-524 Ontologies, 177-178, 180, 184, 187-189, 218, 261, 435 Ontological information, 179 Operating system (OS), 10, 17, 19, 31-32, 36, 480, 489, 621-622, 632, 635, 641, 692 Operation of complex systems, 8 Operational management, 4 Operational requirements, 692 Operators, 79, 199, 426, 559 Opinion polls, 334 Opportunity cost, 685-686 Oracle, 503-504, 508 Order of control, 3 Order status button, 117 Organizations), 160, 167-168, 311, 669, 673, 683 Organization of information, 176-177, 285-289, 377, 443, 446, 472, 534 Organizational behavior theory, 529 Organizational context, 40 Organizational culture, 42, 567 Organizational knowledge, 533, 541 Organizational learning, 528-548 Organizational memory, 529, 530, 534, 543 Organizational structure, 29, 531 Orientation, 145, 269, 277, 291, 654 Orientation compatible with language, 293 Outline, 112 Output devices, 147 Output, 147 OWL language, 180-184, 187 P
Page defacements, 615-617 Page design and layout, 40, 107, 113, 115-120, 123, 203, 279, 339, 346, 348, 501, 507-536, 545, 644 Page flows, 502, 694 Page format, 515 Page length, 112 Page load time, 108 Page size, 108, 634
Page templates, 694 Page title, 113, 205, 279, 487, 692 Page-level functions, 107, 127, 508 Pages loaded incrementally, 108 Paging, 112 Pain, 63 PALIO, 252, 257-264 PalmOS, 631,635, 644 Paper mock-up, 94 Paper prototype, 325, 328, 377 Parameters, 81, 149, 240-241, 253, 256, 364, 426, 692, 698 Parsimony, 644 Participant observation, 311 Participatory design, 670 Password, 488-489, 519, 522, 614-615, 619-620, 644 Password attacks, 616 Password authentication, 636 Pathfinder network, 94 Patient satisfaction, 669 Pattern language, 232 Pattern language markup language (PLML), 697 Payment structure, 659 PCMCIA (PC card), 630 Peoplesoft, 502-503 Perception, 134, 143-144, 160, 579, 653 Perceived quality, 579, 582, 590 Perceived risk, 582-583 Perception and action, 380, 410, 415, 422 Perception and cognition, 93 Perception of sounds, 138 Perceptual and cognitive capabilities, 10 Perceptual and motor systems, 79, 425 Perceptual inputs, 84 Perceptual loudness, 137 Perceptual-motor characteristics, 252 Perceptual-motor skills, 53-55 Performance, 81, 95, 165, 169, 245, 307, 388, 394, 401, 410, 426, 429-430, 462, 467, 496, 586, 590-592, 603, 605,633,651,658 Performance enhancement or improvement, 51, 63, 71, 81, 385, 391 Performance measures, 99, 120, 248, 388 Performance modeling, 389 Performance requirements, 389, 405 Performance support system, 168 Performance testing, 323, 327 Perl script, 483, 618-619, 690 Permissions, 617 Persona, 30, 222, 233, 315, 322, 332 Personal characteristics, 592 Personal computer, 18-19, 135 Personal data assistants (PDAs), 14, 20, 28, 52-53, 60, 64, 203, 257, 267, 269, 294, 325, 443, 451, 498, 523, 630, 632, 637-640
SUBJECT INDEX
Personal information, 614-624 Personal susceptibility, 54 Personality, 212, 228 Personalization, 515, 519, 521, 524, 562-567, 607, 638, 644, 691 Personalized models, 430-432, 437 Pharmaceutical applications, 500 Photo(s), 177, 179, 181-182, 185 Photo properties, 179 Photographs, 124, 181, 184 Physical abilities, 254-255, 264 Physical disability, 268 Physical discomfort, 654 Physical environment, 55, 138 Physical ergonomics, 51-52 Physical object, 144 Physical properties, 151 Physics, 136, 144 Physiology and psychology, 144, 196 Physiological measures, 388, 431 Picture, 157, 178-179, 206, 274, 294 Pictorial integration, 9 Pictorial representation, 478, 622 Pitch/timbre, 136, 142-143, 152 Pixel, 62 PKSM model, 78, 81, 84-85 Plagiarism, 451 Plant operation, 90 Plasma display panel, 61-62 Platform(s), 19, 37, 230, 247, 252, 258-259, 264, 342, 348, 458, 478, 485, 501, 506-507, 535, 528, 536, 538, 546, 553, 631, 636, 644, 661-662, 691, 693 Platform capabilities, 352 Platform independence, 36, 38, 243-246, Pleasantness, 148 Plug-in(s), 149, 342, 479-480, 516, 636-637, 644 Plural forms, 127 Pointing devices, 59-61, 65 Polychronic, 286, 295 Pop-up window, 109, 507, 602 Pop-up ads, 603-604 Pop-up dialogs, 276 Pop-up lists, 443 Portability, 243, 246 Portable computing, 630 Portable devices, 152 Portable document format (PDF), 169, 434 Portal(s), 41, 177-178, 193, 497, 537, 539-540, 544, 551, 638, 647, 663, 690 Ports, 18 Position analysis questionnaire, 404 Post Office Protocol (POP), 23 POP2, 23 POP3, 23 Posture, 65 Practice, 168 Precision, 202
Precursor, action result and interpretation (PARI), 387 Predictive models, 424-430 Preferences, 246, 322, 387, 430, 434-435, 591, 647 Presentation of information, 35, 107, 110, 135, 149, 178, 195, 225, 229, 241, 269, 271-274, 278-279, 289, 291-292, 296, 314-318, 338, 346, 348, 377, 424, 431-432, 442, 462, 466, 495, 561, 588, 608, 617, 647, 665, 689 Presentation and layout, 292 Presentation communication, 452 Presentation modalities, 264 Preserve the skills, 168 Pressure threshold, 146 Pressure, 144, 146 Principle(s), 267, 272, 285, 338, 350-353, 358, 531, 538, 622, 644, 652,688-699 Principles and guidelines, 339, 349 Principles and recommendations, 206 Principles of design, 3, 8 (see also, design principles) Printing, 109, 112, 116, 124, 696 Prior experience, 604 Prioritizing search results, 127 Privacy, 435-436, 481, 499, 517, 519, 521, 523, 590, 603, 607, 613-624, 640-641, 661, 666-669, 689 Privacy police, 324, 616 Privacy requirements, 56 Probabilistic indexing, 194 Probability, 194, 261, 295, 391, 433-434, 587, 590, 642, 660 Problem solving, 165-166, 170, 198, 397, 403, 422, 442, 445, 467, 529, 531-532, 597, 604, 609, 691 Problem space, 81 Procedural knowledge, 79-81, 84, 94, 396 Procedural knowledge structure model (PKSM), 80 Procedural-declarative interaction, 84 Proceduralizing, 76 Procedure usability analysis, 83 Process control, 96 Process quality, 379 Product(s) Product catalog, 682 Product design, 498 Product life cycle, 303 Product page, 597 Product quality, 378, 606 Product representation, 598, 607 Production rule, 79,-81, 426-429 Productivity, 136, 670 Professional users, 500 Profile, 30, 167, 212, 242, 308, 361, 367, 369-371, 373-374, 376, 380-382, 445, 500, 502, 688, 690 Program language, 16-19, 22, 24-25, 187-188, 398, 478, 480, 690
731
Programming device, 14 Programming interfaces, 509 Project constraints, 339 Project management 538 Projection technologies, 62 Prolonged static positions, 54 Promotions, 523 Properties, 177-185, 188 Proposition, 157-164, 166, 168, 171 Protocol, 72, 231, 259, 413, 534-535, 539, 553, 630, 632, 641 Protocol analysis, 242, 325 Prototype, 15, 34-35,139, 242-243, 248, 307-308, 323-324, 327, 330, 340, 347-348, 351-352, 354, 356, 461, 502, 568, 574, 672 Psychoacoustics, 13, 137, 148 Psychology, 31, 196 Psychological methods, 97 Psychological principles, 92 Psychophysics, 137, 143-146, 149, 151 PUA, 85 Pull-down menu, 28, 110, 644 Punch cards, 14-17 Punctuation, 127 Purpose, 649 Push buttons, 37, 345 Pyramid style of writing, 644 Q
Qualitative methods, 303-318, 329, 465 Qualitative assessment, 406 Qualitative data, 330, 603 Qualitative factors, 426, 590 Qualitative measures, 95, 213 Qualitative research, 318 Qualitative structural models, 83 Quality, 142, 149, 368, 381, 386, 392, 396, 454, 461, 465, 467, 501, 511, 536, 547, 578-579, 587, 607-609, 631, 688-699 Quality assessment, 134, 361-362, 365-366 Quality attributes, 245 Quality control, 448, 690 Quality differences, 420 Quality evaluations, 148 Quality goals, 369 Quality level, 364, 368, 371-372 Quality of care, 660, 662, 666 Quality of education, 456 Quality of heath, 668 Quality of information, 199, 476, 664-665, 668 Quality of Internet-based research, 489 Quality of knowledge, 600 Quality of the result, 202 Quality of the sound, 137 Quality of work, 538 Quality pages, 690
732
SUBJECT INDEX
Quantitative information, 95, 401, 465 Quantitative aspects, 427 Quantitative assessments, 397-398 Quantitative data, 331, 603 Quantitative measure, 79, 99, 213, 396, 426, 590 Quantitative methods, 303-318 Quantitative relationships, 652 Quantitative structural models, 83 Quantity, 467, 544, 680, 683 Query, 199-202, 204, 213-214, 232, 261, 443, 506, 511, 540, 651, 664 Query input, 203 Query overview, 200 Query preview, 200 Query refining, 205 Query-based approach, 99 Querying, 187, 199, 419, 499 Question type, 310 Questionnaire, 73, 99, 233, 310-311, 323, 326, 331-334, 339, 393, 396, 424, 430, 476, 509, 523 QWERTY, 58, 147, 639, 642 R
Radio button, 125-126, 345, 473-474, 476, 485, 504 Radio frequency (RF) devices, 641-642 Range restriction, 183 Ranking, 195, 200, 202, 205 Rank order, 161-162, 204 Ranking in search output, 199 Rapid feedback, 36 Rapid prototype, 94 Rasmussen's taxonomy, 84 Rate of return, 685 Ratings task, 171 Rational information agents, 212 RBSE (repository-based software engineering) spider, 201 Reaction time, 425 Reading Readability, 450 Readable annotations, 177 Reading and scanning direction, 294 Reading efficiency, 292 Reading online, 112 Reading performance, 121 Reading speed, 112, 121-122, 146 Real time, 14 Real time intervention, 100 Reasoning, 168, 171, 177, 213 Receptors, 134, 652 Recognition system, 642 Recommendation(s), 140, 391, 394-395, 404, 436, 466, 603, 607, 613, 689-699 Recruitment of participants, 326, 332 RedHat Linux, 624 Reducing load, 53 Redundancy, 134, 150 References, 187
Refinement, 162 Refresh rate, 122, 652 Registered members, 516, 521 Registration, 127-128, 142-143, 515, 518-520, 522-524, 607, 669 Regulatory issues, 661, 666-667, 673 Rehabilitation Act, 645 Relations, 158, 579 Relationship, 162, 180, 182, 218, 228, 234, 239, 305, 310, 312, 503, 511, 528, 535, 539, 576, 580-593, 602-603, 607, 609, 618, 667, 681-687, Relevance judgment, 194, 198-206, 394, 467, 553 Relevant hits, 518 Reliability, 93, 334, 389, 404-405, 458, 462, 466, 598, 632 Repetitive motions, 54 Representation, 148, 160, 171, 179, 194, 287, 411, 413, 422, 564, 595-596, 648, 650 Representation languages, 231 Representation of abstract actions, 141 Representational notations, 245 Representative sample of users, 307, 672 Representing knowledge, 75,427 Requirement analysis, 345-346, 349, 352, 386 Requirement identification, 360-362, 366-367, 373, 377, 379-381 Requirements gathering, 509 Research applications, 471-490 Research context, 100 Research environments, 94 Resize the window, 36 Resolution, 116, 126, 135 Resource(s), 177, 179 Resource description framework (RDF), 179-180, 182-188 RDF Schema (RDFS), 179 Resource management, 666 Response categories, 310 Response device, 476 Response time, 224, 485, 523 Restricted interface, 433 Result item list, 204-205 Result presentation, 205, 376 Result ranking, 204 Retrieving Information, 14, 193-207, 398, 434, 647 Retrieval strategy, 429 Retrieving the instructions, 429 Retrospective queries, 99 Return-on-investment (ROD, 465, 543, 546,574,679-687 ROI calculations, 680-687 ROI case studies, 681 ROI measures, 680-687 ROI metrics, 680-687 Reusability, 245, 247 Reviewing, 197 RGB values, 123
Rhythm, 142-143 Rich representation language (RRL), 231 Risk, 590, 593, 597, 601, 604, 606-607 Robot, 193, 487, 606 Role distinction, 91 Roughness, 145 Route knowledge, 412 Rules, 689-699 Rule of thumb 162 Rule-based description, 412 S
Safety, 689 Sample size, 332 SAP labs, 504-505 Satisfaction, 113, 135-136, 140, 306, 358-359, 362, 373, 395, 430, 528, 536, 578, 588, 590-592, 595, 603, 609, 631, 650, 689 Satisfying, 409-410, 413, 415-416, 420, 598 Save model, 507-508, 511 Scalability, 47, 245-246, 458, 535-536 Scalable fonts, 122 Scale, 310 Scaling, 146 Scan, 118,292 Scanning, 79, 118-119, 123, 197, 240, 257, 295, 431 Scenario, 33, 99, 233, 263, 242, 304, 307, 315, 360-362, 368, 374-376, 379-381, 397, 421, 462, 561, 565, 597, 608, 671 Schedule constraints, 456 Schemas, 216, 218-219 SchockWave, 690 Scraping, 186 Screen(s), 500, 642-643, 645 Screen design standards, 349, 352-354 Screen design, 350-351, 692 Screen linking, 128 Screen shots, 128, 184, 218, 347 Scripts, 36, 39, 231, 291, 473, 478 Scripting, 37, 43, 230-231, 291, 637 Scripting engines, 639 Scripting language, 229, 615-616, 619, 634, 636, 644, 690 Scrolling, 31, 36, 38, 110, 112, 116, 118, 291, 293, 324, 474, 485, 506, 516, 520-521, 644 Sealing, 94 Search, 31, 107, 112, 126, 159, 169, 178, 193-207, 268-269, 273, 279, 286, 292-293, 341-342, 346, 348, 409, 418-419, 422, 442, 450, 517-518, 536, 544-545, 567, 581, 585, 597, 600, 603, 605, 608-609, 647, 653, 664 Search browsing, 197 Search domain, 195 Search efficiency, 199 Search entry field, 203, 205, 516
SUBJECT INDEX
Search entry interface, 203 Search logs, 127 Search performance, 121 Search preferences, 431 Search queries, 193-207 (see also, queries) Search results, 112, 126-127, 200, 218, 344, 349 Search result page, 203, 518 Search robot, 193 Search spider, 193 Search strategies, 193-207, 292 Search term selection, 194-197, 199-207 Search time, 116, 122, 427-428 Search tools, 193 Search topic, 198-199, 202 Search/categorization, 638 Searchable databases, 443 Searched-for keywords, 518 Search engine, 23-24, 38, 40-41, 109, 117, 127, 176-178, 193-207, 211, 271, 295, 322, 324, 419, 435, 444, 448, 451, 486-487, 536, 543, 574 Search engine design, 193-207 Search engine spiders, 109 Searching, 87, 119-120, 123, 213, 281, 284, 295-297, 323, 417, 420, 424, 428, 507-508, 524, 555, 598 Searching for information, 445 Searching for meaning, 408-422 Secondary window, 108-109, 127 Security, 108, 245, 307, 436, 458, 479-481, 489, 517, 521, 523, 534, 602-603, 607, 609, 613-624, 640, 661, 666, 668-669, 673 Secure electronic transaction (SET) protocol, 620 Secure environment, 500 Secure sockets layer (SSL) encryption, 618 Security attacks, 613-624 Security breaches, 614-624 Security guidelines, 621 Security issues, 641 Security policy, 519, 616 Security requirements, 616 Security risk, 622 Selection, 474 Selection rules, 79 Self-documentaries, 313 Self-efficacy, 578 Self-management, 467 Semantic information, 44, 139-140, 143, 158-159, 180, 188, 303, 410, 411 Semantic content, 417 Semantic feedback, 35-36 Semantic highlighting, 199 Semantic level, 253 Semantic networks, 160 Semantic relatedness, 158 Semantic relationships, 221 Semantic requirements, 140
Semantic Web, 176-177, 180, 186-189 Semantic Web documents, 178 Semantic Web Markup, 184 Semistructured interviews, 308 (see also, interviews) Sensory and perceptual motor skills, 65 Sensory capabilities, 53 Sensory capacity, 65 Sensory channels, 148, 151 Sensory impairments, 671 Sensory issues, 52-53 Sensory modality, 134-135, 653 Sequential flow, 74 Serendipitous browsing, 197 Server(s), 43-44, 128, 187, 259, 261, 280-281, 471-473, 479-482, 487, 489, 503, 510, 535-536, 573, 613-624 Server access, 481 Server log-files, 127, 485 Server technology, 497 Server-based technologies, 280 Server-based tools, 279 Session hijacking, 616 Severity of a failure, 140 SEWASIE (Semantic Web Agents in Integrated Economics), 220 Shape, 143, 145, 158, 642 Sharing Information, 481-482 Shopping carts, 117, 514-524 Sickness, 654 Side effects, 650, 653-654 Siebel, 503 Sign-object relations, 139 Simple Mail Transfer Protocol (SMTP), 23,26 Simplicity principle, 9, 310 Simulated personalities, 443 Simulation environment, 425-426 Simulation, 100, 228-229, 427, 436, 456, 458 Simulator, 73, 436 Site design, 285, 338, 595, 601, 683, 689 Site logs, 510 Site management tools, 40 Site management, 178 Site map, 444, 107, 126, 536 Site structure, 107-108, 515, 556, 558 Site-level issues, 107 Situated action/situated cognition theories, 92-93 Situation awareness, 95, 99 Size, 124, 143, 202 Skill, 161, 546 Skill acquisition, 604 Skills, rules, and knowledge (SRK) taxonomy, 76 Small screen, 644 Smart menus, 432 SmartKom system, 224-225 SOAR (States, Operators, and Results), 72, 78, 81, 231, 425, 428
733
Societal issues, 673 Social acceptance, 645 Social and situational theories, 97 Social context, 639 Social dynamics, 99 Social environment, 600 Social interaction 234 Society for Computers in Psychology (SCiP), 490 Sociology, 31, 94 Software, 135, 653 Software architecture, 245 Software development, 90 Software engineering, 243, 245-246 Software usability, 323 Solution assessment focus group, 307 Sorting, 171, 206 Sound, 136, 151-152, 318, 471, 642 Sound design, 139, 141, 143, 148 Sound engineering, 136 Sound frequencies, 137 Sound identifiability, 140 Sound level meter, 140 Sound pleasantness, 140 Sound pressure, 137, 140 Sound quality, 136 Sound suitability, 140 Sound transmission, 142 Sound cards, 135 Spacing, 120-121, 475, 483 Spam, 487, 489 Spatial ability, 649 Spatial organization, 161 Special information, 170 Specialized populations, 309 Specification language Q, 231-233 Spectral features, 138 Speech recognition, 275, 433, 632 Speed and accuracy, 116 Speed of retrieval, 201 Speed of transmission, 202, 396, 487 535, 547 Spell check, 128,277 Spelling mistakes, 127 SPES (situated and participative enactment scenarios), 242 Spiders and crawlers, 193, 201, 487 Spool, 18 Sport, 638 Spreading activation, 428 Spreadsheet, 563, 566, 568, 559 Sprint, 629-631,640 SPSS, 474 SSL encryption, 619-620, 636, 637 Standard(s), 36, 228, 287, 338, 350-351, 441, 443, 450, 472, 502, 545, 603, 605, 631, 635, 648, 660-661, 666, 669, 688, 692-699 Standard design, 386 Standardization, 669 Standards interaction, 501 Standard generalized markup language (SGML), 634 State management, 507
734
SUBJECT INDEX
Static design, 110, 135 Statistics, 421, 485 Statistical data, 214 Statistical techniques, 195 Stereopsis, 652 Stereotypes, 256 Storing Information, 14 STORM project, 171-172 Storyboard, 35, 73, 167-168, 307, 327, 361-362, 461, 588 Strain, 51 Strategy, 412, 416, 429, 433, 545, 608, 671 Strategic IT, 539 Strategic objectives, 575 Strategic uses, 537 Streaming video, 460 Structural ability, 143 Structure, 39-40, 43-45, 157, 165, 178, 182, 186-187, 198, 216, 220, 230, 257, 272, 274, 305, 310-311, 317, 325, 341-342, 362, 409-410, 412-413, 424, 442, 528, 530, 538, 546, 565, 567, 634, 637, 648, 667, 669 Structure and organization, 280 Structure of the information, 108 Structured and systematic analysis, 240 Structured context, 447 Structured interviews, 171, 308 Structured layout, 185 Structured query language (SQL), 206 Student modeling, 433 Study environment, 457 Style guide, 692-699 Stylesheets, 122-123 Subgoals, 84 Subject magnitude, 146 Subject matter, 166 Subjectiveness, 195 Subjective aspects, 409 Subjective estimates, 146 Subjective evaluation, 296 Subjective judgment, 195 Subjective magnitude, 145 Subjective measures, 388 Subjective method, 332 Subjective norms, 573-593 Subjective opinions, 333 Subjective perception, 146 Subjective ratings, 113-114, 116, 120, 122, 510 Subjective satisfaction, 200, 467 Subjective scaling, 146 Subjective user assessment, 247 Subroutines, 17 Subsymbolic representation, 428 Subtasks, 33 Suite or family of applications, 501 Summary of the search result, 204 Summative evaluation, 35 Supercomputers, 19
Survey(s), 303, 306, 308-312, 393-394, 396, 413, 466, 473-490, 480, 502, 510-511, 538, 547, 555, 557, 566, 580-581,613 Survey design, 309-312 Survey knowledge, 412-413, 651 Symbian OS, 635 Symbiotic, 422 Symbol(s), 9, 139, 143, 288, 292-294, 411, 524, 530, 533 Symbolic commands, 419 Symbolic language, 639 Symbolic level, 141 Symbolic relation, 142, 428, 601 Symbology 9, 294 Symmetric properties, 183 Synchronous, 551-552 Synonym, 218 Syntax, 44, 158-159, 180-181, 183-184, 199 Syntactic elements, 183 Syntactic level, 253, 695 Synthetic sound, 142 System, 196-197 System administrators, 500 System architecture, 314 System capabilities or capacities, 247, 615 System design, 659, 670 System integrity, 613-624 System language, 194 System management, 76 System performance, 149 System programs, 17 System response time, 397, 496-497, 505 System status, 443 System to organize representations in meteorology local knowledge, 171 System's goal, 73 Systematic Human Error Reduction and Prediction Approach (SHERPA), 672 T
Tab(s), 118, 348, 350 501, 507 Tab folders, 110, 112 Tab label, 349, 510 Tab structures, 507 Tables, 36-37, 60, 107, 115, 119-120, 180, 185, 451, 499, 506 Table designs, 120 Table of contents (TOG), 110, 112, 126, 277 Table title, 523 Tabular form, 110, 127 Tabular, 127 Tablets, 65 Tacit information, 543 Tacit knowledge, 530-531, 650 Tactile channel, 135, 143, 149 Tactile discrimination, 146
Tactile displays, 56, 62, 65 Tactile information, 653 Tactile processors, 81 Tactile sense, 144, 146 Tactile sensitivity, 145 Tag(s), 181, 185,472,474 Tag structure, 180 Tagged image file format (TIFF) images, 443 Target users, 294, 307-318 Target audience, 321-323, 326, 314, 469 Target population, 457 Targeted user groups, 143 Task analysis, 10, 33, 73-74, 77, 82, 94, 242, 385-405, 412, 425 Task Analysis for Error Identification (TAPED, 672 Task, 649 Task characteristics, 391 Task completion time, 404, 680, 682 Task demands, 83 Task environment, 75, 85-86, 94, 97, 134-135, 267, 385, 396 Task flows, 318 Task goal, 84, 385-405 Task hierarchy, 254-256 Task integration, 561 Task knowledge, 78, 396 Task limitations, 395 Task models, 308 Task objectives, 390 Task performance, 96, 113, 306, 392, 396-397, 404-405, 426-427 Task requirements, 73, 87 Task scenarios, 343, 345 Task structure, 33, 254 Task support, 432 Taskonomy of WWW use, 323 Taskwork knowledge, 94, 99 Taxonomies, 76, 92, 245, 261, 386-405, 435, 495, 534, 617, 621 Teaching, 158, 165 Teaching strategies, 464 Team, 90-92, 94-95, 97, 99, 167, 553-554, 563, 565-566 Team awareness, 554 Team cognition, 90, 94-98, 100, 396 Team cognition measurement, 91 Team collaboration, 100 Team communication, 97, 100 Team coordination, 552 Team information, 568 Team knowledge, 167 Team performance, 90 Team situation awareness, 96 Team size, 91 Team tasks and environments, 100 Team usability, 96 Teamware, 551-569, 565, 567-568 Teamwork, 167, 457, 534 Teamwork capabilities, 468 Teamwork knowledge, 94, 97, 99
SUBJECT INDEX
Technology, 608 Technological infrastructure, 234, 672 Technological issues, 673 Technology Acceptance Model (TAM), 575, 577-583 Telecollaborative activity, 442 Telecommunication, 28, 136, 141, 257, 262,307, 511,645 Telecommunication context, 293 Telecommunications infrastructure, 259 TELNET, 22, 629 Temperature, 143-144, 146 Tempo, 143 Temporal factors, 151 Temporal and spatial resolution, 134 Temporal correlation, 151 Temporal cues, 145 Temporal resolution, 149 Terminal(s), 17-18 Terminological relationships, 218 Terminology, 128 Testing, 670-673 Test goals, 326 Test plan, 326 Test users, 326 Testing and evaluation, 671, 673 Text, 113, 118-124, 135, 143, 158, 164, 178, 196, 198, 203, 206, 214, 222, 272-274, 277-278, 291-292, 294, 305, 313, 377, 397, 401-402, 410, 417-418, 426, 454, 459, 461, 463, 472, 508, 511, 515, 530, 551, 607, 634, 636-637, 642, 652, 664 Text box, 125, 345, 472-473, 475-476 Text Box Sizing, 290 Text direction, 293 Text images, 449 Text input box, 475 Text links, 34, 118, 279 Text messaging, 34 Text presentation, 121 Text scrolling, 524 Text search, 201 Text size, 122 Text wrappings, 292 Texture of pages, 134 Theoretical constructs, 309 Theories, 424 Theory of action, 411 Theory of meaning, 411 Theory of Planned Behavior (TPB), 575, 577, 580, 582 Theory of Reasoned Action (TRA), 575-578, 580, 582 Theory of value, 683 Theory of volitional behavior, 575 Therbligs, 385 Thermal energy, 144 Thermoreceptors, 144-145 Thesaurus, 196, 200, 214, 218-220, 518 Think-aloud methods, 35, 94-95, 168, 171, 326, 387
Three-Dimensional (3-D), 352, 647 3-D digital worlds, 227-228, 647 3-D space, 650 3-D virtual scenes, 652 Threshold, 135-138, 141, 144, 150-151, 591 Thumbnail images, 199, 205, 291, 293 Timbre, 141-143 Time, 96, 599 Time cognition, 285-286 Time division multiple access (TDMA), 634 Time sampling, 312 Timeline analysis, 404 Time-sharing, 18, 21 Titled borders, 698 Titles, 277 Toolbar, 108, 507, 560, 563 Top-down process, 80 Touch, 136 Touch pad, 60, 147 Touch screen, 60, 147, 152, 431, 437, 645 Touch tablet, 147 Touchscreens, 645 Trackball, 59-60, 147 Tracking, 645 Tracking movements, 54 Tracking software, 464 Traditional index, 112 Trained, 91 Training, 4-5, 36, 74-75, 85-86, 91, 96, 168, 198, 200, 223, 251, 282, 314, 387, 392, 394, 404, 415, 417, 454-455, 458, 467, 496-497, 502-503, 506, 510, 531, 539-540, 544, 553, 637, 664, 667, 682, 688-689 Training simulation, 436 Training skills, 436 Transaction(s), 43, 539, 603, 609, 614-615, 617, 619, 624, 661-662, 665, 667-669 Transaction infrastructure, 672 Transaction logs, 196 Transfer time, 151 Transistors, 17 Transitive properties, 183 Translated text, 524 Translation, 168, 289 Transmission Control Protocol (TCP), 20-21, 534-535, 634-635, 641 Transmission Control Protocol and Internet Protocol (TCP/IP), 13, 441 Transmission medium, 632 Transport protocol, 634 Transportation, 90 Travel, 650 Tree structures, 158 Treo, 300, 631 Triads, 92 Troubleshooting, 7, 178
735
Trust, 233-234, 553, 582-583, 592, 597, 601-603, 607-608, 637 TSIMMIS, 220 Tutoring application, 231 Two-dimensional (2-D) spatial map, 420 U
U.S. Air Force's Personnel Training Research Center, 5 U.S. Army Air Force's Aero Medical Laboratory, 5 U.S. Army Air Force's Aviation Psychology Program, 4 U.S. Department of Commerce, 193 U.S. Department of Human and Health Services, 666-667 U.S. National library of Medicine, 664 Ubiquitous access, 506 Ubiquitous computing, 57, 178 UEML, 83-87 Unambiguous, 683 Uncertainty, 295, 543, 668, 688 Uncertainty avoidance (UA), 288 Understand business process, 509 Understand users, 34, 41, 51, 71-72, 74, 76, 78, 136, 303-318, 325, 329, 387-388, 539, 548 Uniform resource locators (URL), 20, 22-23, 25, 38, 44-45, 109, 179-181, 185, 187, 201, 204-206, 257, 312, 324, 333, 441 459, 473, 483, 534, 566,615,624,637,692 Uninhabited air vehicle (UAV), 95, 98-99 United Nations, 514 Universal access, 239-264, 671-672, 688-699 UNIX, 18, 23, 26, 621 Update(d) information, 42, 187, 202, 565, 691 (see also, maintenance) Upgrades, 405 Uploading, 142, 473, 478, 483, 487, 489, 561 Upper and lowercase, 290-291 Uptake of Web-based services, 573-593 Urgency, 140 Usability, 29, 35, 41-42, 109, 113, 118-120, 123, 125-127, 135-136, 172, 186, 197, 201, 204, 241, 243, 252, 267, 272, 274, 284, 289, 296, 303, 311, 313-314, 321-322, 324, 332, 335, 349, 351, 358-359, 361, 368, 380, 389, 396, 404-405, 424, 441, 458, 469, 496, 501, 510, 512, 514, 535-536, 546, 554, 563, 587, 606-607, 613, 617-636, 645, 648, 650-653, 668-699 Usability assessment, 365, 368-371, 394 Usability concerns, 379, 643, 381, 654 Usability design, 357-382, 622 Usability design guidelines, 388
736
SUBJECT INDEX
Usability design standards, 545 Usability engineering, 338, 344, 353, 356, 504, 509 Usability engineering lifecycle, 303, 339-340, 343, 350, 354-355, 358, 502 Usability enhancement, 683 Usability evaluation, 296, 329, 331-332, 466, 502, 562, 644, 689 Usability evaluation methods, 323 Usability expert, 638-639, 641, 644 Usability goals, 342, 360, 389 Usability guidelines, 694-699 Usability heurist evaluation, 443, 515 Usability inspection, 323, 327, 362, 365-371, 325, 378 Usability knowledge, 689, 697 Usability lab, 326, 329 Usability metrics, 248, 680-687 Usability perspective, 496 Usability principles, 462 Usability quality, 363 Usability recommendations, 680 Usability services, 505 Usability specialists, 327 Usability study, 35, 110, 141, 308, 323, 325-328, 330, 334, 339, 345, 348-349, 351-354, 361-362, 365-368, 370, 378, 380-381, 386, 389, 396, 502, 505, 524, 681 Usability test (see usability study) Usability walkthrough, 324 Usage contexts, 243, 361, 366-367, 369-371, 373, 380-381 Use of Web-based services, 573-593 Useful, 170, 172 User characteristic(s), 85, 181, 183, 196, 253, 255-264, 303, 314, 352, 388, 417, 424, 430, 434, 457, 511, 515, 578, 595-609, 649-650, 658, 670 User abilities and capabilities, 52, 54, 254-264, 280, 417, 436, 459-460, 530, 670-673 User acceptance, 234, 573-593 User accounts, 619 User actions, 33, 652 User attitudes and opinions, 306, 387, 573-593 User environment, 72, 313-318 User expectations, 96, 308, 509-510, 519 User experience, 41, 136, 196, 306-307, 495, 503, 506, 511, 546, 555, 601, 605, 607-608, 632, 641, 651, 681 User frustration, 128, 544, 670 User goals, 400-401, 435, 437 User groups, 305 User intentions and motivations, 573-593, 595-609 User interaction, 77, 87, 622, 624, 651-652, 671
User needs, wants, and preference, 42, 80, 287, 306, 318, 331, 416, 507, 648, 680-681, 689 User perception, 538 User performance and preferences, 80, 123, 127, 225, 254-264, 332, 432, 435,511,615 User physical abilities and limitations, 71 User population, 304-306, 339, 353, 415 417, 421, 508 User requirements, 257, 502, 505 User satisfaction, 194, 321, 462, 466, 535,573-593, 596-609, 680 User skill, 508 User task performance, 653 User testing, 32, 42, 330, 689 User vocabulary, 128 User walkthrough, 308 User workload, 670 Users' expectations, 510, 519 User Agent Accessibility Guidelines (UAAG), 280 User agents, 274, 280 User assistance, 127, 128 User centered design (UCD), 358, 365-382 User centered design process, 524 User control, 443 User diagram protocol (UDP), 635 User diary, 509 User Engineering, 303 User interface, 43, 136, 139, 141-143, 152, 200, 216, 220, 240-241, 252, 254, 256-257, 264, 284-285, 291-292, 295-296, 303-304, 313, 318, 338, 358-382, 394, 403, 433, 410, 496, 499, 501, 503, 505, 507-508, 510, 514-515, 517, 519, 524, 552, 556-560, 562, 564, 568, 598, 635, 642, 650, 652-653, 680, 688-699 User interface (UI) design, 123, 149, 201, 285, 287-288, 290, 314, 339-340, 351, 353-354, 357, 507, 514, 608, 692 User interface design principles, 568 User interface designer, 354, 358, 362, 373 User interface elements, 697 User model, 389, 424-437, 433 User modeling, 72, 76-78, 84-85, 256, 424-437 User knowledge model, 78, 434 User performance models, 78-79, 292, 296, 306, 396, 424, 670 User profile, 30, 213, 223, 253, 255-256, 313, 315, 322, 326, 332, 339, 365, 424, 430, 500, 503 User scenarios, 324 User task, 150, 341, 339, 395, 398, 515, 574
User-centered design (UCD), 200, 303-305, 318, 323, 339, 386, 528, 538, 574, 670, 689-699 User-centered framework, 496 User-environment modeling language, 83 User-interface engineering, 506 Utility, 389 Utility learning, 428
Validity, 93, 243, 334, 389, 404-405, 426-504, 689 Variable-width or liquid layout, 115 Various intelligent systems, 195 Velocity of sound, 149 Verbal protocol, 323, 327, 333, 509 Verbal-text communication, 652 Verizon, 630 Vertical scrolling, 644 Vibration 134, 144-145, 148-149 Vibratory feedback, 147 Video, 135, 164, 177, 251, 273, 279, 461, 464, 551 Video tools, 128 Videoconferencing, 149, 454, 459-460 Videotape, 454-455 Vignettes, 307 Virtual environments), 137-138, 146, 152, 214, 598, 638, 647 Virtual 2D or 3D spatial environment, 225 Virtual agent, 222-223 Virtual architectures, 442 Virtual Assistive Technology, 280-281 Virtual book, 418-419 Virtual characters, 212, 222, 224 Virtual city, 226 Virtual displays, 645 Virtual learning communities and classrooms, 441-452, 664, 672-673 Virtual library, 417, 419 Virtual objects, 648 Virtual reality, 56-57, 61,151, 647-655 Virtual reality modeling language (VRML), 648 Virtual retinal display, 62, 652 Virtual team, 436, 445 Virtual universities, 446 Virtual walkthrough, 652 Virtual workspace, 555 Virtual world, 225, 459, 648 Virtually distributed teams, 547 Vision, 134-135 Visual aids, 9 Visual and auditory contents, 523 Visual Basic, 229, 327, 614 Visual capabilities, 53 Visual channel, 246 Visual display, 61, 65, 141, 213, 292, 421 Visual fatigue, 55
SUBJECT INDEX
Visual feedback, 60 Visual function, 53 Visual impairment, 268, 671 Visual information, 79, 461 Visual language, 689 Visual modality, 81, 149, 246, 272, 274, 653 Visual overlay, 119 Visual presentation, 274, 349 Visual representations, 508 Visual search, 110 Visual streaming, 128 VisualHarness system, 216 Visualization, 119, 196 Visually impaired users, 125 Vocabulary, 177, 180-182, 196, 241, 456 Voice over internet protocol (VoIP), 459 Voice recognition, 62 Voice synthesis, 62
w Walkthrough, 74, 94, 345, 348, 351, 672 Wants and needs analysis, 42, 509 Warning, 140, 622 Wayfinding, 650 Wearable devices, 56-57 Wearable handle, 147 Web address, 20, 22, 109 Web application, 36, 38, 41, 44, 135, 141, 143-144, 148, 151-152, 231, 284, 288, 312, 353-354, 385-405, 431, 497, 613-624, 649 Web attacks, 615-624 Web based research, 471-490 Web bloopers, 71 Web casting, 451, 459-460 Web Content Accessibility Guideline's, 269 Web design, 25, 28, 39, 41-42, 71, 86, 100, 110, 123, 157, 177, 287, 289-296, 339, 350, 408-409, 411, 650, 679-687, 683, 686, Web design guidelines, 289-296 Web designer, 25, 41, 122, 144, 152 Web environment, 148, 471-490, 504, 693, 699 Web forms, 472-476 Web forums, 44 Web graphs, 146
Web interaction, 285 Web interface, 44, 52, 57-63, 65, 226, 680, 697 Web layout, 311 Weblog, 42, 305, 332-334, 451, 466, 481 Web motivation Inventory, 596 Web object, 647 Web page, 32, 36, 38, 107-110, 112, 116, 118, 120, 122-125, 157, 169, 176-177, 179, 184, 193, 197, 20-204, 267, 289, 377-381, 471-490, 502, 504, 506-507, 545, 563, 600, 615, 619, 631, 634, 636, 648-649, 683, 692, 695, 698 Web scraping, 185-186 Web service, 574 Web site, 70, 110, 116, 124, 126-128, 135, 178, 187-188, 203, 267, 284, 304-318, 338, 343, 345-347, 349-350, 355, 392, 396, 431-432, 436, 445, 449, 451, 460, 471, 478-479, 483, 486, 495, 497, 502, 505, 512, 545, 559, 563, 597-601, 603, 608-609, 616-617, 619, 636, 648-649, 652, 662-664, 680-681, 686, 688, 690-695, 698 Web TV 325 Web UIs, 690-699 Web vulnerabilities, 614-624 Web work, 157 Web-based Web-based domain application, 97 Web-based instruction, 449 Web-based interface, 224, 304 Web-based learning tools, 109 Web-based methods, 323, 332-334 Web-based programs and applications, 495-512 Web-based services, 573-593 Web-based systems, 78, 83, 86 Web-based user forums, 36 WebCrawler, 201 Web-enabled application, 338, 343, 345, 348, 350-352, 355 Web-enabled cellular telephones, 523 Web-enhanced courses, 447 Web-time Development, 41 What you see is what you get (WYSI-WYG), 461 White space, 113, 116 Wide area information server (WAIS), 24
737
Wide area networks (WANs), 21 Wide connectivity, 631-632 Width or height, 38 Wi-Fi networks, 630-645 Window size, 36, 116, Windows, 555, 617, 619, 621, 631 Windows Explorer, 620 Wireless computing, 629-645 802.11 family, 630-645 Wireless application protocol (WAP), 629, 636 Wireless area network (WAN), 535 Wireless communication, 629-645 Wireless Ethernet, 633 Wireless local area networks, 451 Wireless markup language (WML), 634 Wireless network, 257, 259, 441, 630, 663, 667 With-in subjects design, 325 WMI, 598 WML/HDML, 636 Word processors, 559 Word spacing, 292 Work flow, 528-548, 566 Work reengineering, 342, 349 Workgroups, 554 Working environment, 539 Working surface height, 63-64 Workload, 404, 574, 688, 691, 695 Workstation, 63, 388 World Health Organization, 658-673 World Wide Web Consortium (W3O, 177, 179-180, 188, 269, 280, 645 W3C-WAI guidelines, 240, 252 Wrapping links, 119 WS-FTP, 22 WSS environment, 559, 564 X
XCL, 617, 622-623 Xerox PARC, 31 XHTML, 635 XML, 495 Xyleme query language (XyQL), 214, 216
Y
Yahoo, 24, 178, 201, 204, 508, 539, 574
