Lyft Airport Route Pricing: Trade Platform Take for Match Rate and Rider Retention

You are the pricing product manager for Lyft's scheduled-rides feature, and Lyft is launching in a new city; Toledo, Ohio is the example. Work on a single route only: a one-way trip between the airport and downtown, in either direction. The city has other routes, but they are outside this exercise. On this route, riders are used to paying \$25 and drivers are used to earning \$19. Lyft launches at exactly those numbers, keeping \$6 per completed ride. At that payout, roughly 60 of every 100 ride requests end up matched with a driver. | Side | Current unit economics | |---|---| | Drivers | Acquiring a new working driver costs about \$500. | | Drivers | At the prevailing \$19 wage, drivers churn 5% per month and each completes 100 rides per month. | | Riders | Acquiring a new rider costs between \$10 and \$20, and the cost rises the faster you acquire, because the current marketing channels have limited depth. | | Riders | A rider requests 1 ride per month on average. | | Riders | Riders who never experience a "failed to find a driver" event churn 10% per month; riders who experience at least one churn 33% per month. | Lyft has run one pricing experiment. For a few weeks it lowered its take on every ride from \$6 to \$3 (so drivers received \$22), and the match rate rose almost immediately from 60% to about 93%. Net revenue means the amount riders pay minus the amount Lyft pays drivers. Riders cannot be charged more than the prevailing \$25. **Your task:** maximize Lyft's net revenue on this route over the next 12 months. Should Lyft raise or cut the per-trip driver payout (by changing its take), and by how much? Justify the number. ```hint Look past the first ride The few-week experiment measured match rate, not what happens to a rider afterward. Ask what a failed match costs over the rest of the year. ``` ### Constraints and Clarifications - The rider fare cannot go above \$25; the lever in question is the driver payout, which moves Lyft's take dollar for dollar. - An unmatched request produces no revenue. - Match rates are known only at a \$6 take (about 60%) and a \$3 take (about 93%); any value in between must be estimated and labeled as an assumption. ### Clarifying Questions - Should rider and driver acquisition spend be subtracted, or is the objective strictly rider payments minus driver payouts? - Does the 33% churn rate apply only in a month in which the rider has a failed request, or in every month after a rider's first failed request? - Does a higher payout also change driver churn (5% per month) or rides per driver, or only drivers' willingness to accept this trip? - Is "1 ride per month" exactly one request per active rider, or an average with some riders requesting several? ### What a Strong Answer Covers - Why comparing margin per request alone, which favors keeping the \$6 take, is misleading once failed matches drive rider churn. - A 12-month model linking payout to match rate, match rate to rider churn, and churn to rides, with every assumption labeled, including how take levels between \$3 and \$6 are estimated and how long the higher churn after a failed match lasts. - A check of whether the timing of rider acquisition, rider and driver acquisition costs, and driver capacity change the recommendation. - A specific payout change with a sensitivity range, the conditions that would change it, and a plan to validate it. - Arithmetic that is shown and sanity-checked. ### Follow-up Questions - If you could pay extra only on requests that are not accepted quickly, or on airport rides scheduled well in advance, how would your recommendation change? - Design the next experiment to map the match-rate curve between the two observed points without hurting the launch. - Rider acquisition gets more expensive the faster you acquire. How does that interact with your choice of take?