Optimal composition of solo and pool services for on-demand ride-hailing

On-demand ride-hailing services are becoming a ubiquitous means of mobility in major cities. While convenient and accessible, they are costly for many, and thus ride-hailing providers are often offering pooling options that allow passengers to share a ride and split the fare. The downside of pooling is the extra waiting time passengers may experience as drivers make detours to find and board additional passengers. This paper assesses the competitive advantage of solo (single occupancy) and pool services and investigates the potential benefits of offering both. We characterize the search friction using bilateral meeting functions and present a closed queueing system to capture the pooling effects. We then derive the optimal fare, vehicle occupancy and fleet size for solo and pool services. In symmetric-elasticity meeting functions, we analytically show that (i) the pool service is profitable only when there is increasing returns-to-scale in matching, (ii) the solo service is profitable when the marginal fleet acquisition cost is low, and (iii) the provider should offer both solo and pool services only when there is decreasing returns-to-scale in matching. These observations are further validated numerically for more general cases.