Information Design for On-Demand Service Platforms: A Queueing-Theoretic Approach

Abstract

Information design in on-demand service platforms matters in applications such as taxi services, ride-hailing platforms, and freight exchanges. Displayed service delay information significantly affects platform revenues, leading users to balk or renege. Information design is crucial for platforms with dynamic supply and demand; however, the effects of various information policies on user behavior are unclear. User arrival rates are not only influenced by the platform's information policy, but also by the perceived long-term matching probability in a model with multiple platforms. We use queueing theory to examine information disclosure policies for maximizing platform revenue in a marketplace featuring single- and double-sided queueing service systems. In a single-sided model, forming the queue on the side with the higher arrival rate generates higher expected revenue. The preferred information policy depends on the arrival rate and system load. In a double-sided model, hiding the queue-length information is preferred for the side with a lower arrival rate, whereas displaying it on both sides proves advantageous when both sides have high arrival rates. Considering the long-term influence of matching probability on user arrival rates, the recommendations for selecting the information policy remain qualitatively the same, but the revenue difference between information policies increases.

Keywords: sharing economy, information design, queueing system, balking and reneging