Does Lowering Barriers to Rate Improve the Informativeness of the Rating Consensus on Online Platforms?

Abstract

Platform providers such as Amazon, Google, and Glassdoor have to design policies for submitting ratings. We investigate how lowering the barriers to rate affects the informativeness of the rating consensus. In 2020, Amazon.com introduced a new one-tap rating system, whereas before a written text review was required to rate a product. Anecdotal evidence suggests that the goal of this change was to reduce the relative influence of paid ratings by substantially increasing the number of authentic ratings. We use a diff-in-diff approach and compare the rating consensus of the same books on Amazon and other platforms. Our analyses show that after the policy change, the average rating increases, and the standard deviation across products decreases. Thus, the rating consensus becomes less informative for platform users to discriminate between products. A potential explanation is that lowering the barriers makes it cheaper to provide paid ratings which may outweigh a potential increase in authentic ratings. The results of several additional analyses are consistent with this explanation.

Keywords: Wisdom of crowds, platforms, non-financial information, informativeness, rating manipulations, rating management, fake reviews, Amazon

JEL Classification: D20, D82, M10, M31, M41