Influence of Customer’s Self-comments Versus Others’ Comments on the Customer’s Subsequent Repurchase Behaviour

Abstract

User-generated content (UGC) is an important type of customer feedback that provides reference for a future customer’s purchase decision and acts as a guideline for corporate marketing strategies. This research verifies the influence of a customer’s previous self-comments and others’ comments on the former’s future repurchasing behavior. To obtain observations on customers’ real purchases, we cooperate with one company that sells its product through the Amazon platform to get transaction data and collect comments and ratings directly from the platform. Excluding missing data, in the first round of data collection there are 1,366 transactions from this company’s focal products during the years 2019 to 2022. After following up on any repeat purchase behavior from those customers in the next year, final repurchase orders make up 139 repeat transactions. This research uses the Large Language Model (LLM) to transform customers’ text comments by natural language processing (NLP) into a score. Empirical results reveal that customers are both concerned about their own previous experiences and others’ comments and focus more on the text content rather than the title of the comments. For the moderating effect of product types, the more negative the contents of others’ comments and previous comments on utilitarian products are, the less a customer will repurchase from the brand. One interesting finding is that when making repurchase decisions, others’ comments outweigh a customer’s self-comments.