Adapting Websites to Users
Technology Review (06/09/08) Naone, Erica

Researchers at the Massachusetts Institute of Technology’s Sloan School of Management are working to enable Web sites to automatically adapt to each visitor so the sites present information in a way that each user wants to see it. Although some Web sites already offer personalized features, they primarily use information taken from a user’s profile, stored cookies, or lengthy questionnaires. The Sloan system adapts to unknown users within the first few clicks on a Web site by analyzing each user’s choices.

Sloan professor John Hauser says a Web site running the system would detect a user’s cognitive style, watching for traits such as whether or not they are detail oriented, and would adjust accordingly. Every time the system learned something new about the user the Web site would make a subtle change until the Web site suddenly feels more natural, comfortable, and easy to navigate. Hauser says users should not even realize the Web site is being personalized.

A prototype developed for British Telecom’s Web site is designed so that the first few clicks visitors make are likely to reveal aspects of their cognitive style. For example, the first page users see asks them to choose to compare plans using a chart or to interact with a broadband advisor. Within about 10 clicks, the system understands the user’s cognitive style and morphs the Web site. In addition to guessing each user’s cognitive style, the system can track which versions of the Web site are most effective for each cognitive style.