Captivating the crowd: unraveling sentiments in tourism short videos for effective destination marketing on social media

Tourism short video marketing, benefiting from the richness of on-site sentiments and the convenience of distribution, has become an important way of destination marketing. This study explores sentiments in tourism short videos and identifies their drivers. Based on emotion evaluation theory, this study constructed a research framework combining qualitative grounded theory with quantitative deep learning-based image processing approaches to extract the representative sentiment features embedded in the videos. Empirical analysis was conducted using 1707 short videos from Douyin (i.e. TikTok in China). Five kinds of representational features were uncovered: social humanities attractions, historical and cultural attractions, natural attractions, character attractions, and activity attractions. Sentiment expression in tourism short videos varies by content elements. Moreover, compared to natural attractions, social humanities attractions tend to elicit higher levels of sentimental arousal, and although activity elements are not the main focus, they effectively convey positive sentiment. This study expands on application research on destination marketing and tourism big data (especially for tourism short video data), providing valuable insights for destination marketing organisations to better understand tourists’ on-site sentiment, to enhance the persuasive ability of tourism short videos, and promote the positive emotional image of the destination.