Information or Distortion? The Effect of Customer Generated Images on Product Rating Dynamics

Customer Generated Images (CGIs) on e-commerce platforms have been widely adopted as a promotional tool to persuade customers into purchase. Despite its prevalent applications, the effect of CGIs on customers' post-purchase satisfaction has not been extensively examined. This study postulates that CGIs could have information effect and distortion effect that counteract each other to influence subsequent product ratings. We employ a difference-in-differences model to empirically test our hypotheses, and find that CGIs lead to a decline in subsequent ratings compared with product ratings not exposed to CGIs. Furthermore, the second CGI brings about the most negative effect with a decrease of ratings by 0.14 stars. To evaluate the aesthetics of a CGI, we apply a deep learning based aesthetic assessment model, namely CGI-MobileNet, and validate its superior performance to state-of-the-art models. Additional analyses demonstrate that the negative effect is attenuated if the CGI has a higher aesthetic score where the image contains more relevant information and less distortion. Moreover, the disclosure of reviewer faces offsets most of the negative effect. Considering that reviews on the first pages of a platform have a larger probability to be observed, we further leverage the linguistic features of the review texts and design a pairwise learning algorithm to estimate the review ranking algorithm of the platform. By only considering CGIs appearing on the first pages, similar estimation results are obtained, further proving the robustness of our results. These findings suggest that e-commerce platforms and retailers should be aware of the potential negative effect of CGIs on the rating dynamics and take appropriate measures to circumvent them