Mining Consumer-Generated Product Reviews to Automate Market Structure Analyses
University of Cincinnati
Date: March 14, 2014
Time: 10:00 am - 11:30 am
Location: Room 245
The increasing popularity of social media has led to exponential growth in user-generated content, such as online product reviews, which could be highly beneficial for market intelligence. This source creates both a unique opportunity and notable challenges for academic researchers and marketing practitioners, due to the large volume and unstructured nature of the resulting data. This article presents an innovative method to automate market structure analyses by integrating text mining and sentiment analysis techniques to analyze the linguistic structure of review sentences. A multilevel, hierarchical structure of product attributes and consumer sentiments toward product attributes emerges from this method. This study also uses the derived consumer sentiments to assess the market structure. A prototype system supported an empirical investigation of more than 20,000 tablet reviews on Amazon.com, though the proposed method is generic and applicable to free-form reviews of any product or service category. The findings thus offer insights for ongoing marketing research efforts.
For more information, please contact Cynthia Wong at: cynthia [dot] wong3 [at] mcgill [dot] ca.