Influence in Microblogs: Impact of User Behavior on Diffusion and Engagement

Abstract

Feature-rich social media can reveal many facets of individual user behavior; yet it is difficult to model such behavior due to both the noise and the overwhelming volume and heterogeneity in the data. In this paper, we address these challenges by building a model of user behavior in social media to understand the impact of users' actions on key influence outcomes. We specify models, inspired by social learning theory, that examine the outcome measures of followership (audience), diffusion (reach) and conversations (engagement) as functions of individual actions taken by a focal user within the medium. Our unique panel dataset is prepared by capturing and mining a snowball network of several thousand Twitter users and their activities over an extended period of time.Our results confirm that users' choices about specific actions can indeed positively affect performance on the above outcome measures; and establish a significant role for reciprocity within each outcome. We report several nuances of individual user behavior that may inform managers tasked with designing and executing social media strategies for firms. We find that creating fresh content of specific types as well as engaging in conversations are most effective in improving diffusion as well as leading to conversations. Interestingly, diffusing messages of other users in a focal user's reciprocal network helps with the focal user's own diffusion and conversations; whereas doing so for users outside the reciprocal network tends to hurt own diffusion and conversations.