Choice Overload in Online Platforms

Abstract

The choice overload effect refers to the experimental finding that a larger choice set may in fact yield fewer overall purchases than would be observed with a smaller set. The effect has received much attention, but empirical evidence has been mixed so far. This paper is the first study to test for choice overload in a large scale online experiment, involving thousands of households making major purchases. The experiment was conducted by a leading online travel and price aggregator company, which offered some users a limited set of hotels while others had access to the full list. Overall results are in the direction opposite to choice overload, i.e. purchase rates and click-through rates decreased with fewer choices. However, there are heterogeneous effects across users: while local users exhibit lower purchase rates, foreign users exhibit higher purchase rates when offered fewer choices. Therefore, familiarity is found to be a choice overload moderator. This evidence suggests that retailers can benefit from choice set customization.

Keywords: Choice Overload, E-Commerce, Online Controlled Experiments, Retail, Search

JEL Classification: C9, D12, M31