Lyft Driver Pay Pricing Case: Choosing the Per-Ride Take With a Churn Model

In this pricing case, Lyft operates in Toledo, Ohio, where every ride costs the rider \$25 and the fare is split between Lyft and the driver. How should Lyft change what it pays drivers per ride to maximize its net revenue over the next 12 months? Work from these figures: | | Current split | Tested split | |---|---|---| | Lyft keeps per ride | \$6 | \$3 | | Driver earns per ride | \$19 | \$22 | | Share of ride requests completed | 60% | 93% | - Each rider requests one ride per month, and acquiring a rider costs \$15 on average. - A rider whose request was completed that month churns at 10% per month; a rider whose request was not completed churns at 33%. - Each driver completes 100 rides per month, costs \$500 to acquire, and churns at 5% per month. ```hint Value a completed match Write Lyft's monthly net revenue per active rider as a function of its cut, and remember that the completion rate changes both how many rides earn money and how many riders must be replaced. ``` ### Clarifying Questions - Should net revenue be Lyft's cut from completed rides minus rider and driver acquisition costs, or are other costs in scope? - With only two tested splits, may I assume the completion rate moves in a straight line between them, or is there more test data? - Is Lyft keeping its current rider and driver base in the city, or should I model building the market from zero? - Do extra completed rides require recruiting new drivers at 100 rides each, or can current drivers absorb them? ### What a Strong Answer Covers - A precise definition of net revenue and of which costs move with the split - A model linking Lyft's cut to completion rate, rider churn and driver needs, with each assumption stated - Correct arithmetic at both tested splits and an optimum found between them - A recommendation with its sensitivity to the open assumptions and a plan to validate it before full rollout - Risks the model leaves out, such as driver behavior over time and competitor responses ### Follow-up Questions - If a test at a split between the two showed completion well below the straight line, how would your recommendation change? - How would the answer change if the goal were profit over three years instead of net revenue over 12 months? - Could targeted driver incentives reach the same completion rate more cheaply than a flat per-ride increase, and how would you test that?