Integrated UCD focuses heavily on various strategies for optimizing deployment. It uses technology to move from a largely manual process to one that is much more automated and efficient. The various approaches are discussed in detail in Chapter 5. The case study in this section gives you an idea of what kind of optimization we're talking about. The example looks at three elements of carrying out UCD: getting candidate participants, carrying out a study, and analyzing results (see Figures 2.16 and 2.17). Ways of using the Internet and intranet to optimize these various elements will be shown.
Figure 2.16 UCD can be optimized with tools that automate the process of recruiting, conducting studies, and carrying out data analysis.
Figure 2.17 Desktop online survey to assess key attributes of operating system use. (Courtesy of IBM.)
CASE STUDY: OPTIMIZING UCD
Time and resource limitations can be overcome by using automated user feedback gathering tools.
Survey data can yield new information, validate known information, and dispel widely held beliefs about user behavior.
Using the Internet to gather user feedback gives a team access to an international sample of users.
A team at IBM needed information rapidly on the way users interacted with the Microsoft Windows operating system. The team wanted to know such things as how many windows users typically had open at the same time, how they normally started programs, how they switched between programs, and whether they ordinarily ran programs maximized or with multiple windows showing on the desktop.
However, the team was on a very tight development schedule and decided that it didn't have the time or resources to conduct the required study. It was suggested that the team consider using a Web survey. The team decided to follow the suggestion. The survey was constructed, the participants were selected and invited, and the data was collected, analyzed, and presented back to the teamall within three days. The results yielded important insights for the team (Figure 2.18). It learned that the majority of the participants do not change settings on the interface from the defaults and that little use was made of advanced features. For example, the majority left the task bar at the bottom of the screen with auto-hide turned off. Results also showed that despite having higher resolution monitors, a large number of participants maximized windows and switched between them using the task bar.
Figure 2.18 Summary of study results showing the typical number of open windows, ways of beginning a task, method of switching windows, types of shortcuts used, and method of creating a new document as well as the percent who use windows maximized and the drag-and-drop technique.
Results like these had a major impact on subsequent design direction. Comments from teams who have used this approach have been extremely positive. They realize that the use of mechanisms like these for rapidly collecting input from representative customers not only increases the knowledge they have of their users but it also allows them to collect this information when they don't have the time to use traditional methods. The input is also from an international sample. Despite their speed and power, the team knew that these methods should not be used as a replacement for all of the typical UCD methods and techniques. The team knew that it should augment the traditional methods in critical projects when there is extreme time pressure, a larger sample is desired, or a worldwide audience is important. In circumstances in which a project would normally get no customer input, these mechanisms can be used to ensure that some customer input is collected.
Despite the speed and power of these low-resource methods, it should be pointed out that they should not replace the traditional UCD methods and techniques. They should augment the traditional methods in critical projects, in cases where there is extreme time pressure, where a larger sample is desired, and where a worldwide audience is important. In certain circumstances where a project would normally get no user input, these mechanisms can be used to ensure that some user input is collected.
CASE STUDY: USING INTERNET-BASED USER EVALUATION METHODS
The Internet affords easy access to a large and diverse group of representative customers.
Iterative design can be done over the Internet using surveys, bitmapped low-fidelity prototypes, high-fidelity executable code and Web-based prototypes, and other tools.
UCD activities can be modified to exploit the medium of the Internet.
The Internet offers a tremendous opportunity to reach a wide variety of representative users for input to various UCD activities. UCD methods were used in the design of the conference registration Web site for the Usability Professionals' Association. Volunteers were solicited by posting a notice to a popular forum for usability professionals. A note was sent to all volunteers explaining that they would be participating in the design of the Web site. Then, a few volunteers were selected for each of the activities. Task analysis was performed by sending participants a list of tasks and asking them to rate the tasks and to add missing tasks. The conceptual model was evaluated by having users review text and drawings of task groupings and flow. These were translated into hand-sketched, low-fidelity prototypes. (See Figures 2.192.21.) High-fidelity prototypes were evaluated via the Web. (See Figure 2.22.) All these activities involved iterations of design based on user feedback.
Figure 2.19 Low-fidelity prototype of Web site navigation bar. (Courtesy of the Usability ProfessionalsÕ Association.)
Figure 2.20 Low-fidelity prototype of registration form. (Courtesy of the Usability ProfessionalsÕ Association.)
Figure 2.21 Low-fidelity prototype of tutorial page. (Courtesy of the Usability ProfessionalsÕ Association.)
Figure 2.22 High-fidelity prototype. (Courtesy of the Usability ProfessionalsÕ Association.)
The respondents for this design effort were admittedly a unique group of users. They were capable of wearing two hats: user and usability expert. Having such a population to work with made many aspects of UCD easier. For example, when designing the site architecture, it was not necessary to explain the affinity grouping method we used. Other groups of participants might need more instructions for participating in remote studies. Phone calls, Web chats, and other tools such as whiteboards and video could be used to help explain and implement the UCD methods. (See Chapter 5.)
UCD is all about designing products that best suit customers. One critically important element of accomplishing this goal concerns designing for all members of a target market and not just those who reside in the United States and who have good vision and are otherwise healthy. The majority of information technology products are developed in the United States, whereas the majority of customers in most markets reside outside the United States. Most computer products assume excellent vision, even though many users don't have this, and with an increasingly aging population, this situation will get worse. Individuals with other physical characteristics that influence their use of computer technology should also be addressed. Currently, the World Wide Web Consortium (W3C) provides guidance for designing for accessibility. Additionally, in the United States, section 508 of the Rehabilitation Act Amendments of 1998 requires that all information technology products used or developed by the federal government be accessible to employees with disabilities. Similar policies have been or are currently being enacted in many countries.
As discussed further in Chapter 5, the selection of participants in UCD studies and the method of contacting them needs to take into account these "universal access" considerations. Not doing so will exacerbate an already less than optimal situation with products not appropriately designed for all users in a target market. Focusing on universal access, however, also provides a great business opportunity, especially given the direction of e-business applications, which by their very nature are worldwide in their availability but can also be made accessible to all potential users with the right design.
CASE STUDY: DESIGNING FOR UNIVERSAL ACCESS
It is important that user feedback study participants be representative of the target audience.
UCD activities can be modified to accommodate participants with special needs such as those with visual impairments.
Often, low-tech methods can be easier to modify to accommodate participants with special needs than high-tech methods can.
Designers often attempt to accommodate users with visual impairments by allowing users to control various settings such as font size. In some cases, however, a product may be designed specifically for a visually impaired target audience. Such was the case for a point-of-sale system for the visually impaired. A user research specialist colleague of ours participated in a task analysis session intended to gather tasks and requirements for the development of this system. Members of the design team and representative customers who were, by definition, visually impaired, participated in the session. During the session, activities that are usually employed during task analysis were modified to accommodate the visually impaired participants. For example, as in most task flow modeling exercises, tasks were written on sticky notes and posted onto walls. Participants organized these tasks by moving the sticky notes into groups. However, in this particular session, tasks were written in large, bold text, and were reread aloud frequently to help participants better manipulate them.
This session used an effective, low-tech method for gathering tasks. However, it would also be possible to gather tasks from visually impaired users using a decision support center (DSC). A DSC is an electronic meeting room. (See Figure 2.23.) Several PCs in a LAN configuration serve as workstations for meeting participants. Group collaboration software runs on the PCs. When gathering data from visually impaired users, large monitors running this software at a low resolution with a screen magnification utility are available to enhance participants' ability to participate. For more completely visually impaired participants, screen-reader programs are required. In such cases, more text-oriented software is preferable to GUIs. In such cases, reverting to older, DOS-based versions of groupware programs might be beneficial. In any case, it is critical that actual representative customers for such systems be accommodated.
Figure 2.23 Decision support center used in the study. (Courtesy of IBM.)