Online Brand Community User Segments: A Text Mining Approach

There is a trend that customers increasingly join the online brand community. However, evidence shows that there are nuances between different user segments, and only a small group of users are active. Thus, one key concern marketers face is identifying and targeting specific segments and decreasing user churn rates in an online environment. To this end, this study aims to propose a UGC-based segmentation of online brand community users, identify the characteristics of each segment, and consequently reduce online brand community users' churn rate. We used python to obtain users' post data from a well-known online brand community in China between July 2012 and December 2019, resulting in 912,452 posts and 20,493 users. We then use text mining and clustering methods to segment the users and compare the differences between the segments. Three groups—information-oriented users, entertainment-oriented users, and multi-motivation users—were emerged. Our results imply that entertainment-oriented users were the most active, yet, multi-directional users have the lowest probability of churn, with a churn rate of only 0.607 times than that of users who focus either on information or entertainment. Implications for marketing and future research opportunities are discussed.