Pricing Policy Selection for a Platform Providing Vertically Differentiated Services with Self-Scheduling Capacity

Xiaogang Lin,Yong-Wu Zhou
DOI: https://doi.org/10.1080/01605682.2018.1487822
IF: 3.6
2018-01-01
Journal of the Operational Research Society
Abstract:In this article, we study three pricing policies for a monopoly platform, such as Uber or Gett, who offers vertically differentiated services to customers via multiple types of self-scheduling providers. Ideally, the platform can employ a "dynamic pricing" policy, which pays providers wages and charges customers prices for the transactions of different services that both adjust based on prevailing demand conditions, to maximize its profit. However, since it is challenging for the platform to implement and for providers to understand this policy, the other two pricing policies are commonly adopted in practice, that is, "surge pricing" policy (adopted by Uber) which pays providers a fixed commission of its dynamic prices, and "static pricing" policy (applied by Gett) which pays providers a fixed commission of its fixed prices. By observing these phenomena, we propose to study and discuss the platform's profit performance of these three pricing strategies. We show that the surge pricing policy does not always perform well, which can explain why some on-demand platforms would implement the static pricing policy in practice. Also, although the dynamic pricing policy will significantly improve the platform's profit, we find that the profitability of the static (surge) pricing policy would approach that of the dynamic pricing policy if the platform can balance the number of different types of providers and/or reduce the commission rate.
What problem does this paper attempt to address?