Topic Features in Negative Customer Reviews: Evidence Based on Text Data Mining

Abstract

Several studies have focused on the effects of online negative customer reviews on sales, especially pertaining to Internet shopping and e-retailing. However, there is mixed evidence and the theoretical studies have mainly focused on the volume and valence. To understand the effects of negative customer reviews on sales, the present study uses text data mining techniques to investigate how three factors, namely “content topic, proportion, and consistency,” bout the textual content of negative customer reviews influence online sales. Relevant data were collected from a large-scale online shopping platform. The results of content association and topic extraction reveal four topics—product quality, delivery service, cost performance, and taste. A new econometric model proposed in this study shows that different topics have different effects on sales. Negative customer reviews with a higher percentage or consistency about these four topics significantly jeopardize product sales. Theoretical and managerial implications and future research directions are also presented.