On the Spillover Effects of Online Product Reviews on Purchases: Evidence from Clickstream Data
58 Pages Posted: 14 Sep 2016 Last revised: 20 May 2018
Date Written: May 17, 2018
We analyze the spillover role of the online reviews of other co-visited products on the purchases of a focal product using clickstream data from a large retailer based on: (a) whether the related (co-visited) products are complementary or substitutive, (b) whether the related products are from the same or a different brand, and (c) the choice of media channel (mobile or PC) used. To identify complementary and substitutive products, we develop supervised machine-learning models based on product characteristics, such as product category and brand; and novel text-based similarity measures (i.e., product name similarity and functional similarity). To quantify the functional similarity of pairwise products, we use a text-mining approach of topic modeling on the product descriptions. To train and validate the machine-learning models, we use product-pair labels derived from Amazon Mechanical Turk. Our results show that the mean rating of the online reviews of substitutive products has a negative effect on the purchasing of the focal product, while the mean rating of complementary products has a positive effect on focal product purchases. Surprisingly, the magnitude of the spillover effects of the mean ratings of co-visited (substitutive and complementary) products is larger, notably even twice as large, than the magnitude of the effect of the mean rating of the focal product, especially for complementary products. Also, we find that the negative spillover effects of the mean ratings of substitutive products across different brands in the purchasing of a focal product is significantly higher than those within the same brand. Finally, the spillover effects of ratings are stronger for consumers who use mobile devices versus traditional PCs, which reflects the limited display capabilities of small mobile screens and the narrow consumer attention due to time pressure or multi-tasking while using mobile devices. Our study has theoretical, managerial, and design implications on leveraging the spillover effects of the online product reviews of other related products to encourage online purchases.
Keywords: Online product reviews, Substitutive products, Complementary products, Brand Spillover, WOM Spillover, Topic modeling, Machine Learning
Suggested Citation: Suggested Citation