UNSUPERVISED MACHINE-LEARNING ALGORITHM FOR IDENTIFYING SEGMENT-LEVEL KEY DRIVERS FROM CONSUMERS’ ONLINE REVIEW DATA

Sunghoon Kim, Arizona State University

 The authors propose a new, automatic machine-learning methodology to apply a classic model-based segmentation method in marketing to unstructured online review data. With the proposed method, firms can focus on key drivers per each segment in their marketing activities (e.g., online banner advertising, search advertising); this automatic method will help them systematically keep track of periodic patterns of segment-level key drivers. Using unstructured data from a large educational service review site for rating professors, the authors validate the extracted independent variables from unstructured textual reviews through multiple validation studies and then show heterogeneous key drivers for service satisfaction across three derived segments. For the least satisfied segment, the proportion of reviewers is significantly higher from the Science, Technology, Engineering, and Mathematics education category than from the other segments, indicating that professors in this segment should focus on more diverse drivers of both core academic and atmospheric attributes.