Online Serendipity : The Case for Curated Recommender Systems

Recommender systems are effectively used to provide users with suggestions based on their preferences, and first showed their value in e-commerce sites like Amazon and eBay that algorithmically provided recommendations. A key drawback with these systems is that some items need “personal touch” recommendations to spur on purchase, use, or consumption. A recommender system that facilitates “personal touch” recommendations by enabling users to discover good recommenders as opposed to focusing on algorithmically recommending items addresses this drawback. In this paper, we discuss such a system — the Curated Recommender System. The characteristics of this kind of system are as follows: the system discovers curators and curators make recommendations; a curator is typically another user, though it can be an expert or even an algorithm; curators recommend from curated, thematic, and persistent collections of items; the system needs to support social networking; and curation leads to more serendipitous discovery. It is this last characteristic regarding online serendipity that holds particular promise for Curated Recommender System to provide new value for Websites, especially those that sell books, stream content, or provide social networking platforms