Buyer behavior · Observational dynamic-network analysis with matching/statistical controls for homophily and contemporaneous adoption.
Distinguishing Influence-Based Contagion from Homophily-Driven Diffusion in Dynamic Networks
Proceedings of the National Academy of Sciences · 2009
Summary
Naive methods overestimate peer influence by 300–700%; homophily explains more than 50% of perceived contagion in the analyzed adoption network.
For your marketing
Possible founder application
Use it as a behavioral mechanism or experiment-design reference, not a direct forecast for a SaaS funnel.
Reported result
What the study found
Naive methods overestimate peer influence by 300–700%; homophily explains more than 50% of perceived contagion in the analyzed adoption network.
Research question
What the authors investigated
How much apparent product-adoption contagion is true peer influence versus homophily?
Design
Observational dynamic-network analysis with matching/statistical controls for homophily and contemporaneous adoption.
Sample and setting
A large online social network’s mobile-service adoption decisions and user demographic/behavioral similarity.
Scope and transfer
Limitations to keep in view
The network is one mobile-service setting and remains observational; results do not estimate a referral-program treatment effect.
Research sources
PNAS publisher abstract/results explicitly report the 300–700% overestimate and >50% homophily estimates. Direct publisher-page retrieval was recorded on 2026-10-09; this profile uses publisher metadata/abstract and does not claim independent review of unavailable full text.