The Aesthetic Paradox: How Visual Aesthetic Quality Shapes Digital Engagement on Social Media

Visual aesthetic quality (VAQ) is one of the most salient image characteristics of social media advertising, yet its nuanced effect on digital engagement remains underexplored. Building on existing research on the “beauty premium” and the “ugliness premium,” this research posits that the effect of VAQ follows a U-shaped pattern. Accordingly, a multi-study framework is proposed to verify both correlational and causal relationships. Study 1 quantifies VAQ of 11,559 post images leveraging the enhancer for aesthetics-oriented transformers (EAT) model, achieving an accuracy of 90.56%. Study 2 explores the correlation between VAQ and digital engagement via econometric analysis. Results reveal a U-shaped effect: both low and high VAQ lead to higher digital engagement than middle VAQ. This effect is more pronounced for posts with informational (vs. emotional) advertising appeals. Study 3 examines the causality with stimulus sampling and a randomized experiment, corroborating the findings of Study 2. Study 3 also uncovers the underlying mechanisms of the U-shaped effect: high VAQ enhances digital engagement primarily through perceived attractiveness, whereas low VAQ operates through perceived credibility. These findings refine the “beauty is always good” view into a nuanced U-shaped relationship and demonstrate how deep learning facilitates large-scale image analysis in social media.