An Empirical Analysis of User Content Generation and Usage Behavior in the Mobile Internet

We quantify how users’ mobile Internet use is related to some unique characteristics of the mobile Internet space like usage-based data pricing, the extent of their geographical mobility and their social network behavior. We use a unique panel dataset consisting of individual-level mobile Internet usage data encompassing their multimedia content uploading and downloading behavior along with detailed data on their voice and text calls, and demographics. We build and estimate an individual-level simultaneous equations panel data model using three-stage least-squares (3SLS) after controlling for the endogeneity of the social network. We find that there exists negative temporal interdependence between the content generation and usage behavior for a given user. While the extent of geographical mobility of users is positively associated with their mobile Internet activities, users more frequently engage in content downloading compared to content uploading when they travel. We find statistically significant social network effects that are robust across model specifications. We compare the impact of different kinds of social networks – voice call based social networks, text message-based social network and offline locationbased spatial network. The impact of the voice call-based social network on is higher than the impact of location-based spatial network. Furthermore, the impact of the overlapping network (that is, combination of the social and spatial networks) is larger than that of either the social network or spatial network. We use our empirical estimates to measure the incremental value to firms from negative temporal interdependence and mobility. Implications for stimulating users’ mobile Internet usage, mobile advertising, and targeting of social networks are discussed.