Relative Superiority of Key Centrality Measures for Identifying Influencers on Social Media

Abstract

Marketers have been increasingly turning to social media for marketing campaigns, including viral marketing. A key step in viral marketing is to identify influencers in order to maximize the reach of a marketing message. Existing research shows that centrality measures, such as degree and betweenness, are effective methods for influencer identification. However, viral marketing models used in different studies vary greatly, making it difficult to compare findings across the studies. In this paper, the authors built an agent-based framework of viral marketing that supports different experiment settings, such as different network structures and information diffusion modes, and used it to study relative superiority of various centrality measures. The results show that relative superiority of the measures are affected by some factors, but not as much by others. Practical implications of the results are discussed.