Investigating Consumers’ Engagement on Social Media using Image and Text Mining: Evidence from Facebook
(A2019-8651)

AUTHORS

Iman Ahmadi, Warwick Business School, the University of Warwick; Christian Janze, Independent Researcher; Adrian Waltenrath, Goethe University Frankfurt, Faculty of Economics and Business Administration

KEYWORDS

consumer engagement; social media; machine learning

ABSTRACT

Firms take advantage of social media to engage with their customers. Broadly speaking, the content they create and post for this purpose is of textual and visual nature. While firms recognize the importance of visual content (images, videos), previous research focuses almost exclusively on textual content. Therefore, we develop a conceptual framework allowing for the holistic examination of the relationship of the composition of both the textual and visual content of firm’s posts and the level of consumer engagement. Based on our conceptual framework, we conduct a study covering more than 67,000 posts of 347 members of the S&P 500. We find that (i) a higher emphasis on emotional rather than informative topics in the textual and visual content of a post is associated with higher levels of consumer engagement, and (ii) the presence of human faces, the presence of joyful facial expressions, and more printed text on the visual content is associated with lower levels of consumer engagement.