Identifying multidimensional effects of online reviews on consumers’ automobile purchase behaviours in China: linking observational learning with economic outcomes

Abstract

ABSTRACT Based on 1,087,248 individual online reviews for 2007–2020 collected from the websites of AutoHome and Sohu Auto using the Python method, rating, volume, variance, and many other characteristics of online reviews are generated for 808 automobile products sold in mainland China. Using the observational learning (OL) framework to establish the relationship between automotive consumers’ purchasing behaviours and the display of online reviews, the econometric results reveal that, at the aggregate level, automotive consumers seem to respond more significantly to the volume/variance of online reviews (i.e. hedonic cues) rather than the average rating (i.e. utilitarian cues) for producing adequate OL. However, at the sub-categorical level, consumers are more likely to produce adequate OL regarding ratings representing utilitarian attributes while effects of most of the variance (i.e. hedonic cues) are limited. These findings imply that consumers are more likely to perceive automobile products as ‘more hedonic and less utilitarian’, but an overly complicated online review system forces consumers to simplify the OL process that prioritizes utilitarian attributes, as hedonic information is difficult to be understood. This study provides theoretical and empirical insights about how consumers respond differently to various online review characteristics for products with no clear boundary in a hedonic/utilitarian distinction.