The Effectiveness of Online Reputation Score under Silence Bias

Online reputation systems suffer from the silence bias: Most of consumers do not report after shopping. The reputation score aggregated by consumer ratings may not reflect the real quality of the product. In this paper, we study whether the silence bias decrease the effectiveness of the reputation score. To answer this question, this study proposes a new criterion: If the reputation score accurately reflects the relative quality of competing products, it is effective. We formulate a model of consumer reporting behavior and compare relative reputation scores under different reputation models in use. We find that the effectiveness of reputation score depends on consumers’ reporting propensity.

I. Introduction

Reputation mechanism plays a major role in e-commerce transactions between strangers, and most of e-commerce platforms have established their own reputation systems. The system collects consumer reviews (involving ratings and text reviews) about their shopping experiences and evaluations of various aspects of products and sellers[1], [2]. Online reviews have a significant impact on consumers’ purchase intentions and performance of businesses[3], [4]. Consumer ratings are aggregated in a certain way to form reputation score of the product or seller, and high reputation score usually means high product quality and service[5], [6].