Network Effects and Personal Influences: The Diffusion of an Online Social Network



We study the diffusion process in an online social network, for which we know the individual connections between members. We model the adoption decision of individuals as a binary choice affected by three factors: (1) the local network structure formed by already adopted neighbors, (2) the average characteristics of adopted neighbors (influencers), and (3) the characteristics of the potential adopters. We focus on the first factor, to which end we test two hypotheses based on previous research from sociology. First, an individual who is connected to many adopters has a higher adoption probability (degree effect). Second, the density of connections in a group of already adopted consumers has a positive effect on the adoption of individuals connected to this group (clustering effect). We find empirical support for both the degree and clustering effects as well as for their positive interaction. We also record significant effects for influencer and adopter characteristics. Specifically, for adopters, we find that their demographics and position in the entire network are good predictors of adoption. Similarly, in the case of already adopted individuals, average demographics and global network position can predict their influential power on their neighbors. In particular, an interesting counter-intuitive finding is that the average influential power of individuals decreases with the total number of their contacts. These results have practical implications for viral marketing in a context where, increasingly, a variety of technology platforms are considering to leverage their consumers’ revealed connection patterns.
Contact information:
Dr. S. Puntoni