Improving the textbook exchange system at MIT

Abstract

Textbooks have become one of the biggest expenses of a college education. We attempt to solve this problem by providing students with better information and tools to facilitate book exchange. More specifically, we focus on the subproblem of using technology to build an effective local textbook market. This thesis makes three main contributions. First. it integrates a local sales system into BooksPicker, making it the first tool that does price comparison between both local and non-local (online) sources. Second, it introduces automatic pricing strategies that dynamically price books based on current market prices and book condition. Third, it provides an offer prioritization mechanism that uses location and social information to sort competing offers. Our system was deployed at MIT during the Fall 2010 semester and measured in terms of traffic, local market effectiveness, and through user feedback. Traffic increased three-fold and the number of transactions increased six-fold when compared to each of the two previous semesters. Usage of automatic pricing increased the probability of sale while decreasing the time in market of local offers. We did not obtain enough data to properly evaluate the offer prioritization mechanism.