How to identify product defects and segment consumer groups on an online auto forum

Abstract

China's automotive industry has entered a stage of great growth. Owing to the widespread and rapid information dissemination of online automotive forums, they have become an important source for manufacturers to collect product defect information and understand consumer preferences. The purpose of this study is to investigate car defect information and consumer preferences from online automotive forums and segment consumers based on consumer preferences, which holds great promise as a means for companies to improve product design and consumer relationship management. Based on text data of China's top‐ranked Pacific Auto Network, we applied the LDA model, sentiment analysis, text classification, and cluster analysis methods to determine the influencing factors of automobile consumers’ purchasing behavior, identified automobile product defects, and subdivided consumer groups. The results show that the factors affecting consumers’ car buying behavior are low fuel consumption, affordable price, superior performance, fashionable appearance, and comfortable use. Car product defects are mainly reflected in the three aspects of the car's performance, comfort, and configuration; there are six kinds of consumer groups: price seekers, fuel consumption seekers, performance seekers, comfort seekers, cost performance seekers, and appearance seekers. Of these, cost performance seekers make up the largest proportion, and they constitute the largest group of intention purchasers. These findings can help enterprises develop targeted product improvement programs, but also provide a decision‐making basis for enterprises to achieve precision marketing.

It is about inclusion! Mining online reviews to understand the needs of adaptive clothing customers

“…The LDA algorithm has been used widely to extract meaningful insight from large corpora of textual documents in various fields. Sun et al (2022) Several studies have used the LDA algorithm to investigate customer reviews and other user-generated content. Sutherland and Kiatkawsin (2020) used the LDA algorithm to identify topics of interest that drive customer experience and satisfaction in a sharing economy based on Airbnb reviews.…”

“…The LDA algorithm has been used widely to extract meaningful insight from large corpora of textual documents in various fields. Sun et al (2022) applied an LDA algorithm model and sentiment analysis to determine the influencing factor of automobile consumer behaviour. Bíró et al (2009) modified the technique to successfully collect enough spam and non‐spam topics for web spam classification.…”