REVISITATION PATTERNS IN WORLD WIDE WEB NAVIGATION

Abstract

In this paper, we report user's revisitation patterns to World Wide Web (WWW) pages, and use the results to lay an empirical foundation for the design of history mechanisms in Web browsers. Through history, a user can return quickly to a previously visited page, possibly reducing the cognitive and physical overhead required to navigate to it from scratch. We analyzed 6 weeks of detailed usage data collected from 23 users of a commercial web browser. We found that 58% of an individual's pages are revisits, and that users continually add new Web pages into their repertoire of visited pages. People tend to revisit pages just visited, access only a few pages frequently, browse in very small clusters of related pages, and generate only short sequences of repeated URL paths. We compared different history mechanisms, and found that the stack based prediction method prevalent in commercial browsers is inferior to the simpler approach of showing the last few recently visited URLs with duplicates removed. Other predictive approaches fare even better. These results suggest new approaches to managing history in browsers.