Building ease of use into the IBM user experience.

User experience is the day-to-day interaction that a customer has with all aspects of an offering. For many customers, the user experience is the offering. It is therefore critically important to ensure that customers have a positive experience. However, research from across the industry indicates that the user experience is too often less than positive due to a lack of ease of use. In one study, (1) participants searching for information at a Web site were only able to find the desired information 42 percent of the time. In another study, (2) 39 percent of participants were unable to complete an on-line purchase. It has been found (3,4) that 40 to 50 percent of users who have a negative experience at a Web site do not return to it. A study of software (3) found that the average program has 40 design flaws that impair the ability to use it. Alan Ganek summarized it best when he stated in a previous issue of this Journal, "The computer industry has spent decades creating systems of marvelous and ever-increasing complexity. But today, complexity itself is the problem." (4)

According to a recent Forrester report, "Improving user experience can increase both revenue and customer satisfaction while lowering costs." (5) A Garther study (6) concluded that customer satisfaction can be increased by 40 percent by applying appropriate user experience methods. Of companies that have used some ease-of-use methods, a recent study (7) found that 72 percent reported these methods had a significant impact on product development, and 82 percent reported the methods had improved the ease of use of products developed in their organizations.

In response to these challenges across the industry, we at IBM developed an industry-leading User-Centered Design (UCD) process in the mid-1990s to build ease of use into the user experience of our hardware, software, Web sites, and services. UCD has become a core enabler of our business process, and a corporate leadership team has successfully driven adoption and execution of UCD across all divisions of the company.

An enhanced version of our UCD process, called User Engineering (UE), was developed recently to further improve upon UCD and to optimize it for e-business on demand *. User engineering is the mechanism through which customer changes can be detected. Its processes and methods are used to understand market, business, and user requirements, which are subsequently rigorously modeled and then engineered with iterative user feedback. This approach is critically important to achieving the e-business on demand elements of integration and autonomic computing, ensuring that core business processes are integrated inside and outside an organization and that user tasks are made more efficient by having systems increasingly take over time-consuming and errorprone user tasks. Note that although UE is optimized for IBM's e-business on demand strategy, it is not restricted to being used with it.

The following section provides: (1) An overview of the process transformation we have introduced at raM, including a description of core characteristics that differentiate UCD from a traditional approach and how UE is a significant further enhancement of UCD with regard to these characteristics; (2) an outline of the key elements that were put in place to enable and guide the requisite organizational transformation to optimally execute the process; and (3) an outline of the papers in this issue of the IBM Systems Journal and how they relate to our overall strategy.

Process transformation

IBM used various isolated usability and human factors methods for decades, but these were not part of an overall integrated process. Typically, these methods introduced user testing late in the development cycle and, as a result, were not very successful.

The IBM version of UCD was developed in the 1990s, (8) based on the seminal work of Norman and Draper (9) and Hamel and Prahalad. (l0) Our version of UCD is an approach for designing ease of use into the total user experience with products and systems. The total user experience includes everything the user sees and touches. Unlike the isolated usability and human factors methods used previously, our UCD approach covers the entire design and development cycle with particular focus on the early phases. Most important, UCD is directly integrated into our corporate-wide offering development process. UCD has been used successfully across all our divisions on hardware, software, services, and Web projects for the past seven years.

Although minor updates and modifications were made to the UCD process over those years, a major set of enhancements has recently been made to yield a further "quantum leap" improvement in the process and the resulting user experience with our offerings. The new version of the process is called User Engineering (UE). To date, numerous early deployment projects have successfully used UE (or selected methods within it), and the new process will be used across the company as projects complete their current releases using UCD and begin their next releases with UE.

It is important to point out that our UCD and UE approaches are different from most in the industry. Other approaches are typically one-dimensional, emphasizing only one or two particular methods or types of method to the exclusion of others. For example, some focus on persona and scenarios,(11) while others focus on contextual design. (12) Yet others focus on rapid iterative development. (13) Each of these methods or systems has much to offer but, in our view, they do not offer a complete solution. Our approach has been to evaluate the relative benefits of each method and system across the industry, develop our own methods where there are gaps, and then incorporate the best methods into a comprehensive, integrated, multidimensional process.

A recent study (7) found that very few companies across the industry report having a process as complete as IBM's. Only 13 percent of companies surveyed had an end-to-end UCD process involving users in all phases, and only 5 percent used full multidisciplinary teams. None reported focusing on the total user experience. Benchmark studies carried out by the author with major IT vendors suggest that these companies tend to have a more complete process than the average company involved in the survey described above. However, these studies also reinforce the fact that the IBM approach, particularly with respect to the UE enhancements to UCD, is quite novel across the industry.

A detailed description of UCD and UE will not be provided here. For more details about UCD, see Reference 14. Further details about UE are provided in Reference 15, later in this issue. The following is a brief high-level general summary of the basic elements of the process that are common to UCD and UE. The next section will detail how UCD and UE differ.

Process overview. A simplified generic depiction of the design process is shown in Figure 1, illustrating the major phases of the process together with the type of activity performed within the phase and the basic question being addressed. The design process starts with the collection of relevant market definition information to answer the basic question, "Who do we think will use this offering?" This involves understanding the target markets, types of users, prime competitors, market trends, high level needs and preferences, and so forth. Next, detailed information is collected from representative users within the target markets to understand their goals and tasks to answer the question, "What are they looking for?" Following this, we attempt to understand how the tasks described in the prior step are carried out today either with a competitor's product or an analog method. This answers the question, "What else is out there?"

[FIGURE 1 OMITTED]

At this point, conceptual design of the user experience starts, and early feedback is gathered from users, answering the question, "How's this for starters?" This leads to several cycles of iterative detailed design and user feedback through design evaluation and validation sessions, answering the questions, "Does this work?" and "What would make it better?" At the end of the development cycle, a user feedback benchmark assessment session is conducted to answer the question, "How do we stack up?"

A variety of characteristics illustrate the progression from a traditional approach to our UCD approach and then to our UE approach. These characteristics are shown in Table 1. The three approaches will now be contrasted using these characteristics.