On-demand ride-hailing platforms with heterogeneous quality-sensitive customers: Dedicated system or pooling system?

We examine how an on-demand ride-hailing platform chooses between the dedicated system and the pooling system when facing heterogeneous customers and drivers. In the dedicated system, there are two independent queues where quality-sensitive and quality-insensitive customers can be matched with drivers of low-or high-quality service, respectively; while in the pooling system, both types of customers are randomly matched with all drivers without distinguishing the service quality. We characterize the heterogeneity of customers and drivers, and we analyze the platform's optimal pricing and wage decisions in the two systems, respectively. Comparing the two systems, we find that when the driver's basic opportunity cost is not quite large and the difference in opportunity cost between two types of drivers is relatively small, the pooling system can achieve a win-win-win outcome for all participants (i.e., customers, drivers, and the platform). Furthermore, a high level of customers' sensitivity to service quality can increase the platform profit in the dedicated system but has no impact on the pooling system. However, when the level of customers' sensitivity to service quality is within a certain range (which is relatively small), quality-sensitive customers may pay a higher price to join the pooling system compared with the case in the dedicated system.