Smart Data Pricing

Abstract

Over the past decade, Internet adoption and applications have proliferated. As a result, data traffic has nearly doubled in volume every year, yet at the same time network capacity is not keeping pace with this growth. This dissertation argues that the very diversity in usage that is driving growth in data traffic points to a viable solution for this fundamental capacity problem.

Smart data pricing looks at the users who drive demand for data and examines the incentives that will alter user demand so as to reduce congestion. Simply raising data prices will likely drive down demand, but also lead to vast user dissatisfaction. More sophisticated pricing schemes may not work in practice, as they require users to understand the prices offered. We demonstrate the feasibility and benefits of a smart data pricing approach by considering three such solutions.

Empirical traces of mobile data traffic reveal that network congestion can vary widely over time. Providers can thus lower congestion by charging users lower prices during less congested times, incentivizing users to shift their usage to these times. We show for the first time that such time-dependent pricing works for mobile data, by developing new algorithms for time-dependent pricing and then demonstrating their efficacy in a fifty-person trial.

Offloading some data traffic to supplementary network interfaces, such as WiFi, can help reduce network congestion. Yet such networks are not always available, so users may not be willing to subscribe to them. We consider user subscription dynamics as influenced by supplementary network pricing and congestion. We show that service providers' actions can have unintended consequences: for instance, increasing the coverage area of the supplemental network can increase its congestion, inducing some users to leave.

Network neutrality refers to the general principle that all traffic on the Internet should be treated equally, yet content providers may still gain a possibly unfair advantage by subsidizing their content's traffic. We show that the reality is more nuanced: sponsored data disproportionately benefits wealthier content providers, but also favors cost-sensitive users. Moreover, these users benefit proportionally more than content providers.