Ride Solo or Pool: The Impact of Sharing on Optimal Pricing of Ride-Sharing Services

Abstract

A ride-sharing platform (RSP), such as Uber or Lyft, can sometimes offer passengers an option to share (pool) the ride with fellow passengers. On the one hand, a passenger who pools benefits from paying a lower fare and the RSP benefits from increasing occupancy per car, thereby serving more passengers. On the other hand, a passenger who pools takes more time, on average, to reach her destination and may have to share the ride with a stranger, and the RSP gets a lower profit margin per passenger than from solo rides. We develop a queueing model to find the RSP's optimal revenue at equilibrium when passengers are strategic, and drivers are independent agents, and design the RSP's revenue-maximizing price-service menu. We find that offering both solo and pooled rides is optimal when the distribution of passenger-type is not skewed, and congestion is not high. Counter intuitively, when congestion is high, the RSP benefits from offering only one ride choice, and simulation-based results extend these findings when more than one route exists. Revenue per driver can be non-monotonic with respect to the number of cars, even though total revenue is monotonic. Numerical analysis provides important insights when the number of drivers is endogenous. For instance, equilibrium revenue per driver can decrease when the passenger arrival rate increases. We find that when the driver supply side thickens, revenue per driver decreases. The compensation drivers receive (as a fraction of total revenue generated) increases with their reservation price and decreases with the arrival rate of passengers. When demand is low, a higher wage-payout fraction can increase RSP's equilibrium revenue.