ENHANCING USER ENGAGEMENT WITH ONLINE INFORMATION: THE IMPACTS OF ONLINE ADVERTISEMENT LAYOUT AND SERENDIPITOUS RECOMMENDATION

The massive amount of information on the Internet makes it difficult for online marketers to maintain consumer attention and website visitors to discover the products of their latent interest. This thesis focuses on the design of two information technologies that alleviate the concerns: online advertising and recommender system. Study One examines a fundamental, yet under-studied design factor of static online advertising: text placement. Results from a field experiment and two eye-tracking experiments suggest that text separation can increase consumer attention and enhance ad effectiveness. Study Two presents a new serendipity-related dataset and proposes novel serendipitous recommender systems to address the problem of overspecification in recommendation. By leveraging transfer learning and multi-objective optimization techniques, our models improve over state-of-the-art methods, in both recommendation relevance prediction and serendipity prediction. Taken together, this thesis contributes to Information Systems literature by exploring how to smartly design online information technologies to better engage consumers.

Author dc:creator

• CUI WEI

• online advertising, eye-tracking, serendipity, recommender systems, deep learning