Social Learning and Diffusion of Pervasive Goods: An Empirical Study of an African App Store

In this study, the authors develop a structural model that combines a macro diffusion model with a micro choice model to control for the effect of social influence on the mobile app choices of customers over app stores. Social influence refers to the density of adopters within the proximity of other customers. Using a large data set from an African app store and Bayesian estimation methods, the authors quantify the effect of social influence and investigate the impact of ignoring this process in estimating customer choices. The findings show that customer choices in the app store are explained better by offline than online density of adopters and that ignoring social influence in estimations results in biased estimates. Furthermore, the findings show that the mobile app adoption process is similar to adoption of music CDs, among all other classic economy goods. A counterfactual analysis shows that the app store can increase its revenue by 13.6% through a viral marketing policy (e.g., a sharing with friends and family button).

💡 Research Summary

The paper presents an integrated structural framework that simultaneously captures the macro‑level diffusion dynamics and the micro‑level discrete choice behavior of consumers in a mobile app store. At the aggregate level, a Bass‑type diffusion model is employed to describe the overall adoption curve of apps over time, with the key driver being “social influence,” operationalized as the density of adopters in a consumer’s vicinity. This density is measured along two dimensions: offline (physical proximity) and online (digital network connections). At the individual level, a multinomial logit choice model incorporates conventional product attributes—price, category, rating—together with the social‑influence variables to estimate the probability that a given consumer selects a particular app.

The authors analyze a large, transaction‑level dataset from a major African app store, comprising millions of download events, user location data, timestamps, and app characteristics. Bayesian estimation via Markov Chain Monte Carlo (MCMC) yields posterior distributions for all parameters, allowing a rigorous assessment of uncertainty. Prior distributions are informed by existing literature on Bass diffusion rates and price elasticity in discrete‑choice contexts.

Empirical findings reveal several important insights. First, offline adopter density exerts a significantly stronger effect on app choice than online density, indicating that physical proximity remains a potent conduit for social contagion even in digital markets. Second, omitting social influence from the choice model leads to biased estimates of price sensitivity and app quality effects, underscoring the risk of mis‑specifying demand models for policy or managerial decisions. Third, the temporal pattern of app adoption closely mirrors classic diffusion curves observed for physical goods such as music CDs, suggesting that the underlying adoption mechanism is not fundamentally altered by digitization. Fourth, a counterfactual simulation of a viral‑marketing intervention—adding a “share with friends and family” button—projects a 13.6 % increase in store revenue, demonstrating the tangible profit potential of leveraging social networks.

The study acknowledges limitations, including the geographic concentration of the data (which may affect external validity) and the coarse measurement of online network density. Future research directions proposed include extending the model to multiple regions and platforms, integrating richer social‑media data to refine online influence metrics, and exploring dynamic pricing or recommendation policies within the same structural framework.

Overall, the paper makes a methodological contribution by marrying macro diffusion and micro choice modeling within a Bayesian paradigm, and it provides actionable evidence that app store operators should incorporate offline social signals into marketing strategies and avoid ignoring social influence when estimating consumer preferences.