Click (Bother You?), Click (Bother You?), Click...

It used to be that "clicks" were irritating sounds, not core business objectives. But that was before the commercialization of the web.

Nowadays, being able to generate clicks, or more precisely "click-throughs," can make the difference between eating at Chez Panis or McDonald's. And that's just one reason why research into click-through patterns and browsing behaviors is vital to online business and therefore a key focus of research.

For instance, Eugene Agichtein, Eric Brill, Susan Dumais, and Robert Ragno, members of Microsoft Research's Text Mining, Search, and Navigation Research Group, have been investigating user interaction with search engines. What they've found, and what they report in their paper "Learning User Interaction Models for Predicting Web Search Result Preferences," is that accurate modeling and interpretation of user interactions with a search engine can significantly improve search-result ranking, the detection of "click-spam," Web search personalization and, ultimately, the overall Web search experience. While user interactions with the Web search engines are plentiful, new robust techniques are required to understand the relationship between user interactions and result quality.

Likewise in their paper "Improving Web Search Ranking by Incorporating User Behavior," the research team discusses utilizing the discovered patterns of user behavior to improve algorithms that rank search engine results. This work demonstrates how to add a new rich set of user behavior features into ranking for large relevance gains.

"Most search engines today use a somewhat two-dimensional approach, matching user queries with the content and link structure of Web pages to return a list of results," says Agichtein. "We’re looking at how to add a third dimension--the users themselves--to improve the search experience. By examining click-through and browsing patterns across a large number of users, we are able to learn a great deal about how people interact with search technologies and can thereby improve our accuracy dramatically."

In other words, not to put bad puns in Agichtein's mouth, we can learn what clicks with users, and what doesn't.

Posted by Jon Erickson at 09:26 AM  Permalink