DoorDash Gives Away Margin on Purpose. That's the Flywheel.

Pay DoorDash $96 a year and it promises to stop charging you for delivery. That sounds like a coupon. It is actually the most important thing the company does, because on the very next order DoorDash waives a fee it would otherwise keep — saving the member around $5 each eligible order⁸ — and it does this on purpose, millions of times a day, while admitting in its own SEC filing that those discounted orders can earn it less per order than a full-price one.⁴ A company that gives away margin by design should be a charity. Instead, in 2024 DoorDash turned its first full-year profit in history.³ The giveaway is the engine.

The official story is that DashPass is a loyalty perk — a nice discount for heavy users, the way an airline gives you free bags. That framing misses the whole machine. DashPass is not a reward DoorDash hands out after you become loyal. It is the device that manufactures the loyalty in the first place, and it works precisely because DoorDash is willing to lose a little on each trip to win the next hundred.

The loop: give up the fee, buy the frequency

Here is the mechanism, worked all the way down. A non-member treats delivery as an occasional splurge — the fee is a tax they feel on every order, so they ration their orders. A member has prepaid the tax. The marginal delivery now feels free, and that one psychological shift changes everything downstream: the member orders more often, hesitates less, and — because they have already paid for the year — has a reason to keep choosing DoorDash over a rival rather than letting the dollars they sunk go to waste. DoorDash confirms the result plainly in its 10-K: members retain better and order more frequently than non-members.⁴ So the company hands over per-order margin and receives, in exchange, two things that compound — duration and frequency. A customer who orders twice as often for three times as long is worth far more than the margin on any single waived fee. That is the trade, and it is not close.

Now watch the loop close. More members ordering more often means more volume flowing across the marketplace — over 42 million monthly active users and $80.2 billion of Marketplace gross order value in 2024.²³ That volume is the bait for the other two sides of the network. More orders make DoorDash the most valuable channel for restaurants, which sign up to be where the demand is; denser restaurant selection makes DashPass more useful, which pulls in more members; more members deepen the order density that lets DoorDash route Dashers efficiently and deliver faster. Each turn of the wheel makes the next turn easier — and cheaper. The membership giveaway isn't draining the flywheel. It's the hand on the rim that keeps it spinning.

Why the lead widened while the giveaway grew

If the trade were marginal, the share numbers would be tight. They aren't. Consumer transaction-panel data put DoorDash at roughly 60.7% of U.S. food delivery at the end of 2024 — more than double Uber Eats at 26.1%, with Grubhub at 6.3%.⁷ (The 67% figure that floats around the press comes from a different panel methodology and isn't from DoorDash's filings, so treat it with care.) The point isn't the exact decimal; it's the direction. DoorDash subsidized fees harder and grew share at the same time, which only makes sense if the subsidy is doing strategic work. A member is far costlier to poach than a fee-paying drifter — they've prepaid, they've built a habit, and the selection on the other side is denser than a challenger can match on day one. The flywheel doesn't just retain customers. It quietly raises the cost of competing for them.

The honest threat: a bundle DoorDash can't match

The fair objection is that this looks too clean — a giveaway that funds itself, a wheel that only accelerates. The real vulnerability isn't inside the loop; it's next to it. DoorDash sells a pure-delivery subscription. Uber sells a membership that bundles rides and eats into a single relationship, and a subscription you use to get to the airport and order dinner is structurally harder to cancel than one you use only for dinner. If that cross-modal bundle proves stickier at scale, DoorDash is defending a one-sided habit against a two-sided one. The numbers it leans on cut both ways here: members are already over half of its monthly active users²⁵, which is strength — and also a ceiling, because the easy conversions are largely done, and the next members are the ones a rival's broader bundle can court most cheaply. The flywheel is real. It is not invulnerable. A wheel that spins on the frequency of a single behavior can be slowed by a competitor who owns more behaviors.

DoorDash figured out that the most valuable thing it sells isn't a delivery. It's the moment a customer stops doing the math before ordering — and it bought that moment by giving away the fee it could have charged. Strip away the discount language and what's left is a company that decided, deliberately, to earn less on every trip in order to own every customer's next decade of dinners. It worked: more members, more orders, a wider lead, and finally a profit.³⁷ The genius was never the coupon. It was understanding that in a flywheel, the margin you surrender today is the speed you keep tomorrow.