The Evolution of Influence Through Endogenous Link Formation

Abstract

Marketing researchers and practitioners are interested in targeting individuals in social networks who may have disproportionately higher level of influence over others in their network. While the extant literature suggests either individual characteristics and/or network position may explain one’s relative influence, our study bridges these two streams of work by investigating the endogenous acquisition of network position as a function of exogenous individual characteristics. Specifically, do those with higher expertise attain higher influence when people choose who they listen to endogenously? Using an agent-based modeling (ABM) simulation framework, we model the dynamics of two types of individuals: experts with exogenous information and non-experts who listen to others' information. Over the course of multiple diffusions, agents choose whom to “listen to” for information; dropping less useful ties and adding new ones. We find that experts can have less influence (out-degree) than non-experts who collect information from multiple sources. Furthermore, homophily amongst experts decreases their influence. However, noise in communication channels moderates these effects be- cause non-experts also aggregate errors. We show our results are robust to alternative dynamic networks. Our research suggests that marketers should consider the envi- ronment, community characteristics, communication medium, and product domains to maximize the spread of information.

Keywords: agent-based modeling, influence, communication channel, noise, diffusion, link formation, social networks

JEL Classification: C15, C99, M30, M31, M39