Does Visual Review Content Enhance Review Helpfulness? A Text-Mining Approach

Abstract

The previous literature has explored the effects of text review contents and the characteristics of reviewers on the helpfulness of reviews. Moreover, recent research has addressed the role of photos as visual review content, which influences consumers' evaluation of review helpfulness and shapes their decision-making in the consumption of experience goods. From the perspectives of signaling, attribution, and media richness theories, this study aims to explore how photos, as visual review content, influence review helpfulness. To test the role of visual contents, this study utilized the Yelp Open Dataset and selected restaurants in Florida as research subjects, considering both text and photo content. The LIWC software was used to analyze the sentiment of review contents. The corresponding photos were matched to restaurant codes and then categorized based on food and ambiance. The main findings indicate that positive reviews negatively affect review helpfulness, while negative reviews are positively associated with it. Moreover, the presence of visual review contents (such as photos of food and the environment) moderates the effect of extreme reviews on review helpfulness. Photos attenuate the negative effects of positive reviews and weaken the positive relationship between negative reviews and review helpfulness. By confirming the boundary condition of visual review contents, this study provides new insights that are expected to be useful for future research and offers novel implications for online platforms and business management.