Recommendation Systems in the Online Classified Industry : Studying the Online Classified Businesses and the Potential Role of Recommendation Solutions in their Development

Abstract

This thesis aims to provide an overview on today’s online classified businesses and recommendation solutions based on a research conducted at the recommendation systems provider Gravity R&D and its various online classified partners. Classified sites have become prominent players of the E-commerce and Internet Advertising sectors. Considering their business models, there are a number of challenges in making a classified site profitable, but recommendation systems can potentially provide solutions to many of these issues. Recommendation systems are now widely applied in many of today’s data-rich industries as a result of a significant scientific development. These solutions, including the ones implemented at classified sites, are built up using several specific technologies. In fact, their utilization and goal can be very different site-by-site, largely depending on the maturity of the studied classified businesses and on the platforms’ actual interaction points with the different user segments. It is advantageous to study recommendation systems in the classified industry not only due to the wide usage and vast variety of the applied solutions. The large amount of items and users that can be found on classified websites, as well as the available experience and good case practices make it easier to get a complete overview on the ways of the implementation and on the possible performance of recommendation solutions. To gain this overview, this thesis will present both of the above subjects, and it will introduce and evaluate several examples and use cases of recommendation solutions implemented at different classified sites, illustrated with the necessary data, charts and figures.