Visual rightness: Unveiling the impact of visual aesthetics on destination marketing effectiveness in travel vlogs with machine learning

Abstract

Travel vlogs have become a significant tool for destination marketing. However, limited tourism research has empirically examined the visual aesthetic attributes of travel vlogs and how they drive viewers' behaviors in destination marketing. To fill this gap, this study measures the visual aesthetic attributes of travel vlogs with computer vision (CV). Then, we examine their effects on viewers’ behavioral outcomes, as reflected by social media performance metrics (including clicks, likes, and collects). The results from both economic (OLS) and interpretable machine learning (XGBoost & SHAP) methods indicate that (1) for vlog cover, color balance, the rule of thirds, and texture diff increase clicks, whereas diagonal dominance and color diff decrease it; (2) for vlog video, saturation, clarity, and texture diff increase both likes and collects, whereas color diff decreases both. In addition, color balance and size diff increase likes, whereas contrast decreases collects. These results offer insights for optimizing visual aesthetics in travel vlogs to achieve destination marketing effectiveness.