An Analysis of Diffusion Speed Among Digital: Empirical Insights Derived from a Comparative Analysis Across Products on Steam

Abstract

This study sets out to explore the diffusion speed of digital products on digital platforms. It does so by applying a binary logistic regression across a dataset of Steam products. The dependent variable used is a new way of conceptualizing speed, by comparing how many observations it takes to reach a certain percentage of the total amount of diffusion, relative to the total amount of observations the product had available, this study used 80%. Several theoretically informed independent variables were inferred on the products based on keywords attached to each of them. These independent variables were tested against the dependent variable in the logistic regression, wherein 3 independent variables turned out to be significant. Those were “Free”, “Network effect”, and “Hardware requirement”. The study assumes that the free products are using a freemium business model and suggests that this effect comes as a consequence of these products being digital. Furthermore, it is suggested that while the other two effects are indeed significant for these digital products, they can be explained theoretically by theories that also pertain to physical products.