Market segmentation based on customer experience dimensions extracted from online reviews using data mining

Abstract

Purpose This study aims to develop a practical and effective approach for market segmentation using customer experience dimensions derived from online reviews. Design/methodology/approach The research investigates over 6,500 customer evaluations of food establishments on Taiwan’s Yelp platform through the Latent Dirichlet allocation (LDA) data mining approach. By using the LDA-derived experience dimensions, cluster analysis discloses market segments. Subsequently, sentiment analysis is used to scrutinize the emotional scores of each segment. Findings Mining online review data helps discern divergent and new customer experience dimensions and sheds light on the divergent preferences among identified customer segments concerning these dimensions. Moreover, the polarity of sentiments expressed by consumers varies across such segments. Research limitations/implications Analyzing customer attributes extracted from online reviews for segmentation can enhance comprehension of customers’ needs. Further, using sentiment analysis and attributes of online reviews result in rich profiling of the identified segments, revealing gaps and opportunities for marketers. Originality/value This research presents a new approach to segmentation, which surmounts the restrictions of segmentation methods dependent on survey-based information. It contributes to the field and provides a valuable means for conducting customer-focused market segmentation. Furthermore, the suggested methodology is transferable across different sectors and not reliant on particular data sources, creating possibilities in diverse scenarios.