The Role of Spatial Proximity in the Adoption of a Digital Product

Abstract

We adopt a model-based approach to assess whether and to what extent spatial proximities between members of a target population influence the adoption of digital products. Our data come from a company that sells software that enables car dealers in Germany to list used cars in an online market. As compared to other studies in the diffusion literature, our data are unique because we have the exact date and time of adoption for every adopter, as well as the geo-coded location data and descriptors for every potential member of the target population. Further, because the product was digital, it was available for adoption instantaneously by everyone in the target population, which ensured there were no frictions because of unavailability in any specific region. Thus, if we find that spatial effects influence the diffusion process, we can ascribe those effects to local contagion (due to communication, observation, or competitive pressures). We estimate an Accelerated Failure Time hazard model with time-varying covariates and shared frailty. We find that spatial proximity matters in a significant way in the adoption of the software, even after accounting for the effects of a number of covariates (observed heterogeneity) and for frailty (unobserved heterogeneity) on the times to adoption. Specifically, adopters who were located closer to non-adopters had a significant effect on hastening adoption time of the non-adopters. Our results suggest that although Internet and other communication technologies may have greatly diminished the role of distance in the availability of information about a product as well as in the availability of the product itself (for digital products), they have not eliminated the importance of local contagion or trustworthy personal information.