Consumer Interactions and Peer Effects in Socially-Connected Digital Products
(A2025-124067)

AUTHORS

Yulia Nevskaya, Queen's University; Yijun Chen, Imperial College London

ABSTRACT

In-product social interactions are integral to many digital products, e.g., online games and social media. Social interactions happen during engagement with the product, making them correlated with product use intensity and user churn. This endogeneity makes it challenging to measure peer effects. Causal peer effects are needed for marketing strategies employing social influence. Using data from an online game, we estimate a joint model of network formation and peer effects on user churn. Our findings are: (i) Social ties are driven by unobserved gamer traits. Ignoring this endogeneity leads to underestimating peer effects. (ii) Networks with diverse gamer characteristics have less balanced social ties. (iii) Preference for diversity of ties depends on the CRM objectives. A less diverse network minimizes negative impact from popular user’s churn, while a more diverse network leverages positive influence of popular users. Firms must consider network composition in their product design.