Lingnan study using AI analyses over 16,000 social media videos in order to show key marketing strategies that boost viewership

The competition in managing short video content on social media platforms and executing online marketing has intensified in recent years. Crafting eye-catching video titles and thumbnails to increase viewer engagement has become a crucial marketing strategy. A recent study by the Department of Marketing and International Business at Lingnan University titled “Clicks for money: Predicting video views through a sentiment analysis of titles and thumbnails” used artificial intelligence (AI) to show that thumbnails expressing strong sentiments tend to attract more views, while strong sentiments conveyed in video titles may have the opposite effect. The Lingnan research team suggests these findings indicate that specific strategies in producing video content can fully leverage their videos’ potential to boost viewership and enhance online marketing efforts.

The Lingnan research team analysed a total of 16,215 videos published on YouTube, covered six categories: music, film, sport, education, politics, and technology, all with English titles. Using AI to extract content from video thumbnails, and including specific elements such as facial expressions, vehicles, animals, and food, the study also analysed the captions and text descriptions embedded in the thumbnails, aiming to understand how these elements affect video viewership and influence viewers' decisions to watch.

Within the analysed videos, 38 per cent (6,168) included captions in the thumbnails, and 51 per cent (8,220) used thumbnails featuring human faces. The results revealed that videos with thumbnails expressing strong sentiments (18.3%), whether positive or negative, had higher view counts. Also, videos with simple and concise captions (17.5%), titles conveying positive energy (36.4%), titles featuring punctuation (83.6%) and emojis (11.1%), and those with captions in the thumbnails (38%) experienced more clicks.

Prof Peng Ling, Head of the Department of Marketing and International Business of Lingnan University, said that this innovative study applied AI for systematic analysis so as to gain insights into how sentiments expressed in video titles and thumbnails play a crucial role in viewer engagement. The research team also includes Prof Cui Geng, Professor of the Department of Marketing and International Business, and Prof Sebastian Chung Yu-ho, Assistant Professor of the Department of Marketing and International Business.

The study was recently published in the peer-reviewed Journal of Business Research, published by Elsevier. Click HERE for details.