Face It: Exploring the Power of Human Faces in User-Generated Photos in Online Reviews

Online reviews crucially shape consumer purchasing decisions, with visual cues receiving increasing attention. However, the specific impact of human faces within these visuals remains underexplored. This study investigates how user-generated photos (UGPs) and facial characteristics affect review helpfulness. Using 152,320 Amazon clothing reviews, we analyze the role of UGPs, face presence, quantity, quality, and expression across different valences. Our findings reveal that while both UGPs and human faces enhance review helpfulness, general UGPs are more effective in positive reviews, whereas human faces are more impactful in negative ones. Furthermore, results do not support the "more is better" assumption, showing that reviews with multiple faces are generally less helpful than single-face ones, a pattern reversed only when facial clarity is high, or the review is negative. Finally, non-smiling faces enhance the helpfulness of negative reviews, supporting the role of consistency. Theoretically, this study enriches media richness theory by highlighting the fit between content type and review valence, provides evidence that information-processing constraints limit the benefits of visual quantity, and extends internal consistency to expression-rating alignment. In practice, we offer guidance to reviewers on selecting credible images and to platforms on optimizing algorithms to prioritize high-value visual cues.