Optimization in Online Content Recommendation Services: from Clicks to Engagement
Coauthor(s): O. Besbes, Y. Gur, and A. Zeevi
A new class of online services allows publishers to direct readers from online articles they currently read to other web-based content they may be interested in. Unlike many other online services (such as display ads) that terminate after a single "click," the content recommendation service encourages readers to use it repeatedly and often serves as a navigation tool for readers, shepherding them through the extensive and rapidly changing world of online content. We study the dynamic content recommendation problem, focusing on the questions of how to measure the performance of such service, and how to optimize its performance in light of the environment it operates in. Based on a rich data-set, we develop a representation of content along two key dimensions: clickability, the likelihood to click to an article when it is recommended, driven mainly by the attractiveness of the title; and engageability, the likelihood to click from an article when it hosts a recommendation, driven mainly by the "quality" of the content itself. Based on this representation we propose a class of user-path focused heuristics, with the purpose of efficiently integrating the objective of generating a high instantaneous click probability with one of preserving an ongoing user engagement. Work is based on collaboration with Outbrain, a leading provider of dynamic online content recommendations.
O. Besbes, Y. Gur, and A. Zeevi "Optimization in Online Content Recommendation Services: from Clicks to Engagement." , Columbia Business School, (2013).