#364 2015 · Uber · Transport
Uber paid drivers to wait for riders who hadn't arrived yet
the problem
No drivers without riders, no riders without drivers
background
Marketplaces need both sides live simultaneously to work: a rider app with no nearby cars is useless, and a driver staring at zero ride requests quits within days. A startup entering a new city had neither — no drivers logged in yet, no riders trusting an app showing an empty map — unlike incumbent taxi dispatch, which already had cabs idling at stands and standing driver rosters to draw on.
The obvious fix — market hard to riders first, or let supply and demand find each other organically — fails because whichever side arrives to an empty marketplace first rarely returns. A driver who opens the app to silence, or a rider who waits fifteen minutes for a car that never comes, treats that as proof the service doesn't work; each side's tolerance for emptiness is roughly one bad experience long.
what everyone would do
Market hard to one side first — riders or drivers — and let the other side follow organically once demand or supply is visible, the standard way marketplaces try to bootstrap. It fails because whichever side shows up to an empty marketplace first has almost no tolerance for the experience: a driver staring at a silent app or a rider waiting for a car that never comes treats one bad session as proof the service doesn't work, and neither side waits around for the other to catch up.
what they saw
Uber saw that the problem wasn't attracting either side individually, it was that each side's value depended entirely on the other already being present — a genuine chicken-and-egg trap that patience alone couldn't solve. Rather than wait for real rider demand to justify driver earnings, they could pay drivers as if the marketplace were already liquid, buying time for actual demand to build without losing the supply side to a single bad first experience.
the move
Uber recruited drivers in new and existing markets by guaranteeing minimum hourly earnings regardless of whether ride requests actually came in — most clearly documented in January 2015, when Uber cut rider fares in 48 US cities while simultaneously guaranteeing driver earnings so take-home pay wouldn't fall, keeping drivers logged in and available before demand alone could pay them.
why it works
Drivers quit fast when the app shows silence, because an empty request queue means the marketplace isn't working for them that day; guaranteeing minimum earnings regardless of actual rides removes that financial consequence, so drivers stay logged in and visible even before real demand exists to justify their presence. Riders opening the app then see nearby available cars — which is what actually drives adoption — with no way to distinguish subsidized supply from organic demand already at work. That adoption generates real ride requests, which gradually replace the subsidy, so the guarantee only has to cover the specific window between launch and organic liquidity, not run indefinitely.
the payoff
Uber's own January 2015 announcement and contemporary press coverage confirm the guarantee-plus-fare-cut mechanism across 48 cities; the same supply-first logic is described by former Uber growth staff as recurring through the company's earlier city-by-city launches from 2011 onward, though a precise, independently citable figure for any single 2013 city launch could not be confirmed live.
where it breaks
The tactic needs deep enough capital reserves to fund real cash payouts for as long as it takes local demand to build — undercapitalized competitors who copied the same city-launch playbook without matching capital depth often ran out of subsidy before liquidity took hold. It also depends on the subsidized side's presence being visible to the other side; paying drivers to stay logged in only works because riders can see the cars on the map, so subsidizing a side whose presence isn't observable does nothing to build the trust the mechanism relies on. And withdrawing the guarantee has to be managed carefully — drivers can anchor to income expectations formed under the subsidy, and behavior built around a guaranteed floor doesn't automatically convert cleanly into organic, demand-driven behavior once the floor is removed.
what came after
The tactic was later named and codified as the standard 'cold start' playbook for network-effect marketplaces in Andrew Chen's 2021 book The Cold Start Problem, drawing on his time on Uber's growth team, and versions of paid, guaranteed-minimum supply recruitment have since been used by delivery, grocery, and other two-sided marketplace startups launching into new geographies.
references
- [1]Uber's Clever, Hidden Move: How Its Latest Fare Cuts Can Actually Lock In Its DriversForbes, 2015forbes.com
- [2]Price Cuts for Riders with Guaranteed Earnings for DriversUber Newsroom, 2015uber.com