User-centered design for personalization

Abstract

In chapter 1, I introduced the concept of personalization and showed how tailored electronic communication is the product of centuries of evolution. Personalization involves gearing communication towards an individual’s characteristics, preferences and context. User-Centered Design (UCD) was proposed as a means to achieve a good fit between personalized communication and the individual user. This means that design of personalization should include an initial focus on users and their tasks, studies should be conducted that focus on actual user behavior and perceptions, and finally, an iterative design approach should be applied. In this way, problematic issues related to specific, personalized usability issues, such as privacy or a need for control, can be prevented. Chapter 2 addressed an early stage in the UCD process of personalization to determine the role of trust in the organization providing personalization, trust in the technology, and perceived controllability in relation to the intention of potential users to use online content personalization. Using an online questionnaire, 1,141 participants were demonstrated four common approaches to online content personalization and a non-personalized baseline condition with respect to a fictive municipality. We assessed participant perceptions of the aforementioned factors and determined their influence on the intention to use the different approaches to online content personalization. Trust in the organization appeared to play no role in the decision to use online content personalization. Trust in the technology had a moderate effect on the intention to use, while perceived controllability was overall the most important antecedent. When designing online content personalization, it is therefore most important to provide users with the option to control personalization. Next, users should be assured that they are interacting with an organization in a secure electronic environment. The requirements engineering phase was focus of chapter 3. In that chapter, we proposed a user-centered approach to requirements engineering for personalized e-Government services and demonstrated its value by means of a case study. The approach utilized interviews and formulated requirements by focusing on concrete and measurable criteria, low-fidelity prototyping, and evaluating by means of a citizen walkthrough. The case study reaffirmed the importance of applying an iterative approach to design, as the translation of user input into system design may not align with the original characteristics, preferences and contexts of the user. Furthermore, using a citizen walkthrough, the proposed approach succeeded in making personalization understandable to participants, which is an important objective for evaluating personalization. Finally, the case study demonstrated that a multidisciplinary design team is a crucial aspect of creating personalized e-Government services. In chapter 4, we reviewed literature that focused on user-centered evaluation of personalization (i.e., evaluations that include an assessment of subjective criteria or the identification of usability problems). The findings indicate that current user-centered evaluations, as reported in the scientific literature, are not well-aligned with the principles of UCD. Questionnaires appeared to be exceedingly popular, while methods that have been found to identify usability problems well, such as thinking-aloud techniques, are only used sparingly. Specific usability issues for personalization are only rarely a topic of investigation. In the last few years, however, an increasing number of publications have reported on evaluations that focus on acceptance, iterative design or system trust. This trend suggests that personalization researchers are becoming aware of the added value of user-centered evaluations and are starting to make it part of their common research practice. Chapter 5 reported a comparison of the usefulness of three methods (i.e., interviews, questionnaires with open-ended questions and concurrent thinking-aloud techniques) for identifying usability issues in personalized systems. Thinking-aloud was the only method that uncovered all critical and serious problems related to personalization as well as usability problems not related to personalization. Furthermore, it was also the method that best elicited participant feedback on the perceived quality of personalized output. Comments on the specific usability issues for personalization were elicited best by the questionnaire. Therefore, when evaluating a personalized system in order to obtain input for redesign purposes, we recommend a combination of thinking-aloud techniques and questionnaires with open-ended questions that address specific usability issues in personalization.