Efficiently matching on-demand trip requests from riders to available drivers is a central problem for ride-hailing services like Uber (Uber Technologies, Inc.). An emerging concern in ride-hailing operations is that too many offers are dispatched for a single trip request, and this has an impact on the reliability of the service. The purpose of this study was to explore inefficiencies within ride-hailing dispatch mechanisms, with a particular focus on the misalignment between existing matching logic and optimal stopping. The matching process was framed as an optimal stopping problem, identifying when to stop sending out offers per trip request. Findings from the optimal stopping model suggest that driver acceptance rates decline as the attempt number increases so long waiting times do not mean a the ride is more likely to be accepted.  Second, by applying Optimal Stopping Theory to the rides matching process, the existence of an optimal stopping point for dispatching ride offers that occurs before each market occasion’s finite horizon was identified for each market occasion. Third, the cost of each matching attempt number is non-linear, given a pronounced increase in cancellations during the first 30-40 matching attempts (approximately 5 minutes) followed by a plateau. The results of this study imply that the current unfulfilled logic is suboptimal as it lets offers linger in the marketplace for longer than it should and that efficiently matching riders to drivers is crucial to preventing trip loss via rider cancellations.