Abstract Today massive on-demand ride requests are dispatched every hour on ride-sourcing platforms, but an unneglectable amount of confirmed orders is later cancelled by customers. Customers’ order cancellation not only wastes drivers’ efforts, but also lowers the availability of supplies on the ride-sourcing platform. Based on a two-month hourly-average dataset provided by Didi Chuxing, this paper makes the first attempt to look into customers’ cancellation behaviours of the confirmed-orders. The customers’ confirmed-order cancellation rate (COCR) is observed to be significantly and negatively correlated to the customers’ waiting time for pick-up time. We regard such counter-intuitive phenomena as an outcome of customers’ mode switch upon meeting vacant taxis while in waiting for ride-sourcing vehicles. The cancellation rate is thus characterized as a function of customers’ average waiting time for ride-sourcing vehicles and cruising taxis, the platform’s penalty strategy for cancellation of confirmed-orders, and the customers’ own characteristics (i.e., value of time, psychological cost of order cancellation). As both the customers’ waiting times for ride-sourcing vehicles and cruising taxis are endogenously determined by the demand and supply of ride-sourcing and taxi markets respectively, a system of nonlinear equations is proposed to depict the complex interactions between the two markets at equilibrium while considering customers’ order cancellation behaviour under a given penalty and compensation rule. With the proposed model, we theoretically establish the existence of the equilibrium under mild conditions, and examine the optimal pricing and penalty/compensation strategies that respectively maximize the platform’s profit and social welfare respectively. Numerical examples are presented to demonstrate the distinct optimal pricing and penalty/compensation strategies for different objectives and examine the various impacts of the social credibility.