Using Social Network Structures to Predict User Engagement: The Importance of Social Ties in Influencer Marketing (with Remi Daviet, UW Madison and Seungbae Kim, UCLA), Under review at Marketing Science September 2022. SSRN
Best paper award for the Marketing Analytics, AI, and Machine Learning track, 2022 AMA Summer Academic Conference.
As collaborations between brands and social media influencers become increasingly popular, predicting and understanding the capacity of an influencer to generate user engagement (such as likes and comments) has garnered increasing attention from researchers. Not surprisingly, managers have been relying on follower-based statistics to identify individuals with potential to reach a vast number of users on social media. However, this approach may often direct managers to accounts with millions of followers accompanied with high recruiting costs. In this paper, we argue that the network structure of influencers is a useful measure for capturing an influencer’s ability to generate engagement. Using Instagram data, we perform a deep-learning analysis on the social interaction network of influencers and show that the network structure alone explains a large share of the variations in user engagement, even outperforming traditionally used variables such as the number of followers in the case of comments. We also show that many insights can be obtained from the network structure. Notably, we find that high-performing influencers form elite sub-communities that may not be central to the larger social network. This study contributes to the emergent literature on the importance of social ties in the digital environment.
Modeling Co-Engagement Patterns in Brand Information Networks (with Yaxin Cui, Northwestern University and Keran Zhao, University of Houston), Invited revision at Journal of Marketing, August 2022. SSRN
The rise in electronic interactions has made information networks ubiquitous. Information networks result from users’ activities on information systems (or on electronic platforms) and can include various types of social structures of firms/brands. There has been a growing consensus among researchers to understand the structural relationship among members of an information network as a first step to utilizing these networks for marketing purposes. A better understanding of information networks can help marketers develop a clearer overview of the interests of their brand communities on social media and effectively predict marketing outcomes. In this paper, we use extant statistical models, in particular Exponential Random Graph Models (ERGM), to understand the drivers of co-engagement patterns within brand networks on Twitter and to predict the future connectivity patterns between brands. Unlike conventional social networks that involve direct interaction between individuals, edges within a brand network arise due to common followership activity between Twitter users. The ERGM model reveals a mix of network and individual level brand characteristics responsible for network formation, thereby disclosing a list of significant brand (and network) features likely associated with users co-following brands on social media. Marketing implications of the work are also discussed.
Keywords – brand networks, social media, network formation, generative modeling, network characteristics.