Most web sites and web services generate a wealth of attention data and make little or no use of it. Using attention data is an excellent way to provide high quality and relevant content recommendations to end users with fresh and relevant content in the form of recommendations. This talk will focus on how to quickly add attention-based recommendations to your web site or web service with only one day of work.

A primary benefit of high quality content recommendations is the creation of a user-generated bottom-up taxonomy. In this taxonomy every piece of content points to its next best matches. Having such a taxonomy enables a different kind of search and browse experience in which the user searches for the right “small world” and then refines results within this world using the related content recommendations. Since this small world inevitably spans multiple web sites and services, these taxonomies must deal with heterogeneous content types across a variety of web locations. This challenging problem and the best practices required to enable cross-site taxonomies will be described in detail.