Measuring Product Type with Sequential Dynamics of Online Product Reviews: Theory and Applications

The distinction between experience products and search products (“product type”) is increasingly important in online commerce due to the way products of different types interact with Internet-based technologies such as online retailing, search engines and online reviews. One challenge in understanding the role of product type in online commerce is that it is difficult to measure objectively. We propose an approach where product type can be inferred based on the statistical analysis of online product reviews. Our theoretical analyses indicate that for pure search products, the variance of the mean rating should decrease as more consumers rate the product; for experience products the variance of the mean rating may remain constant or increase depending on the importance of the experience attributes in consumer utility. We demonstrate the use of this classification approach at category, product and attribute level using archival data from Amazon.com, Yelp.com and Ctrip.com. Implications of this analytical tool and empirical findings for research, theory and managerial practice are discussed.