Examining the Impact of Content Features on Customer Engagement in Social Media: A Data Mining Approach on Instagram

Objective With the ever-increasing role of social media in the growth of digital marketing, which is one of the most important yet unpredictable trends today, platforms like Facebook and Instagram have become indispensable tools in advertising and marketing. Instagram ranks second after Facebook in terms of leading social media platforms used by marketing professionals and brand owners. The large amount of advertising and marketing content produced on social networks has led business owners to create content that is engaging for the target audience, thus increasing the engagement rate on their business pages. On the other hand, due to the significant increase in the number of followers on Instagram over the past years, as well as the interest of users in following businesses on this platform, the great importance of this network for marketing specialists is undeniable. Therefore, this study aims to investigate the effect of published content attributes on customer engagement on Instagram. Methodology Considering the huge amount of content produced on social networks, one of the best methods to analyze and identify patterns in these networks is data mining techniques. To achieve this goal, visual and textual features of 490 content pieces published on the commercial pages of cosmetics products from the Cinere, Lisel, Schon, and Bioderma brands were collected. Then, the Clementine data mining toolkit, along with three methods—Association Rules, Apriori Algorithm, and Decision Tree—were used to identify features affecting customer engagement in terms of likes, comments, and conversations, and to evaluate their effectiveness. Findings The results of this study indicate the effects of embedded visual and textual content features of published messages on customer engagement behavior. Based on these results, the use of persuasive textual features—such as Holiday Mention and Remarkable Facts—along with informative textual features—such as price and target audiences—affect customer engagement on social media. In addition, visual informative features—such as brand centrality and product centrality—affect customer engagement behavior to a certain degree. Conclusion Previous research on Instagram has mainly emphasized likes and comments as customer engagement behaviors. In addition to examining these two behaviors, the present research also evaluates the conversations raised in comments (conversations). Moreover, by considering all parts of the published content and defining thirteen textual features and two visual features, the present research provides a deeper understanding of the effective content features that influence customer engagement behaviors and fills the existing gap on how to create published content to increase engagement. The findings provide business managers and brands, especially in the field of cosmetics and hygiene products, with an efficient and effective customer engagement strategy for visual social networks.