None of these six lessons is complicated on its own. What actually produced the swing was doing all of them at once, after years of doing none of them.
A head-on launch in the biggest city, against an incumbent running tens of thousands of taxis, is close to unwinnable early. Win smaller markets first, then move up.
Years of a fully decentralized structure — one generalist per city handling onboarding, support, pricing, and incentives simultaneously — burned enormous money for a mediocre result, while a leaner competitor spent less and grew faster. The fix wasn't more budget. It was centralizing by function so expertise compounded nationally instead of getting reinvented from scratch in every city.
Staff the new central functional teams with people who came up through the local markets themselves. You keep the ground truth without the operational drag of a fully decentralized structure.
A fleet partner manages many drivers under one relationship, and paying them a small margin on top avoids onboarding thousands of individual drivers one by one. Leaning harder into fleet partners tends to raise efficiency. But concentrate too much supply into too few hands and those partners gain real leverage over you — enough to coordinate against you or defect to a competitor. A workable target: something like 50-100 fleet partners in a city with around 20,000 cars, not a handful of mega-partners.
A large share of individual drivers do a handful of trips and disappear. That's structural, not a sign the operation is broken — it's part of why a mix that includes fleet partners tends to be more resilient than relying on individual acquisition alone.
Across hundreds of cities on every continent — from some of the wealthiest markets in the world to some of the poorest places ride-hailing operates — the fundamentals never change. Be the cheapest, be reasonably reliable. What shifts by market is only which part of the pitch gets emphasized — more on safety where that's a real concern, more on price where budgets are tight. The core lever is identical everywhere.
The full, detailed version of how this actually played out — including the years it took to hit the low point and the six months it took to reverse it — is worth reading in full.
A short working call — I'll tell you which lesson matters most for you right now.
Book a working call