Delivery Commissions That Kill Restaurant Margins: Myth vs Reality

How much does the app really take from a normal order?

The app takes a percentage of a ticket that is usually small, which is why every point matters:

in New York the average subtotal of a third-party app order was 28.12 USD between January and March 2024 (NYC Department of Consumer and Worker Protection). With tickets like that, the commission is charged on the full sale while your profit lives in a few dollars of the plate.

For example, if your average app order is 30 USD and the platform keeps 25 %, 7.50 USD leaves before you count packaging, the promotion you paid for yourself or the refund for the burger that arrived cold, and at forty orders a day that line weighs more than many invoices you do check every week.

The right number is not in the contract but in the payout statement, so pull it order by order for a full month and work out what share of gross sales got lost along the way.

Is there a law that limits what delivery apps can charge?

Yes, but it is local:

in New York City the delivery fee cannot exceed 15 % of the price of each online order, and the transaction fee has its own ceiling, according to the delivery app requirements published by the city's consumer regulator (rule in force when the source was consulted; confirm it at the official link). That cap protects whoever operates inside that jurisdiction and nobody else.

Outside it, the commission is whatever you sign, unless your municipality or country has passed something similar, and you confirm that with the industry association or a local lawyer before taking it for granted, never with what a group of owners says on social media. Watch one detail the cap does not touch: paid campaigns inside the app and visibility plans usually sit apart, and that is where the bill grows again without breaking any rule.

Why does a platform offer me several plans with different commissions?

Because the platform sells visibility as well as delivery, and each commission tier buys a different spot inside the app.

Restaurant Dive documented in 2020 that DoorDash launched three levels (Basic, Plus and Premier) with commissions of 15 %, 25 % and 30 %, and the underlying logic still holds: a lower commission means less exposure or less reach, and the other way around. The most frequent mistake is picking the plan out of fear of vanishing from the listing, when the right choice comes from the contribution margin of the dishes that actually sell through that channel.

If the star dish leaves little after recipe cost and packaging, the expensive plan only multiplies orders that lose money, and more volume on a negative margin is the fastest way I know to burn cash. Margin per dish first, plan second, in that order and no other.

What commission can a dark kitchen pay without losing money?

A dark kitchen can carry whatever commission its margin per dish absorbs after food cost and packaging, and at the high end of Deliveroo, which a Dephna guide places between 25 % and 35 %, that margin runs out fast.

Run the numbers with example figures. If a dish sells on the app for 100 pesos, with food cost at the 32 % ceiling that Masterestaurant treats as a MAXIMUM and 5 pesos of packaging, 63 pesos remain before the platform; what would happen if the commission were 35 %?

You would have 28 pesos to cover payroll, rent, utilities and this month's promotion, and if you also fund a 20 % discount on that same dish, 8 pesos survive and the order no longer pays even its share of the rent. That is why a dark kitchen designs its app menu for the channel, with portions, packaging and price built for that commission instead of inheriting them from the dining room.

Is it fine to charge more on the app than in the dining room?

Yes, and in most cases it is the only way an app order leaves any profit, as long as your contract with the platform does not force price parity, which is worth reading before you move a single number.

The tension is real: raising the price protects margin but can lower conversion, and leaving it flat keeps volume while giving the profit away. Diego F. Parra settles that clash at Masterestaurant with menu engineering by channel, which is something different from a flat increase.

You do not raise the whole menu, you raise what the commission destroys; the dishes with the best margin can stay close to the dining room price and act as a hook, while thin-margin dishes go up as much as needed or leave the app. For years I believed a single price was a matter of honesty with the guest, and I was wrong: the app customer pays for convenience, and the kitchen has no reason to subsidize it.

How much money do delivery commissions move in a big city?

Far more than an owner senses by looking only at their own place:

the NYC Department of Consumer and Worker Protection puts it at 15 million USD per week paid by New York merchants to the apps during the first quarter of 2024. That sum explains why the commission can no longer be treated as a minor sales expense. It is a structural transfer of margin from kitchens to platforms, week after week, and whoever has it spread across several apps rarely sees it consolidated in a single line of the income statement.

The advice is easy to say and tedious to do: open separate accounting lines for commissions, funded promotions, refunds and delivery packaging, and review them every Monday next to prime cost. What does not show up on the income statement does not get managed, and in delivery the invisible part is usually the most expensive one.

Do I need to sell through delivery apps for my restaurant to grow?

No: almost half of US operators, 47.2 % according to the Restaurant365 2026 mid-year report, get between 0 and 10 % of their revenue from third-party delivery.

App delivery works better as a complementary channel, with a short menu and its own prices, than as the main engine of a business whose cost base was built for the dining room. My position here is firm: the app helps people discover you and your own channel helps them buy again, whether through WhatsApp or your own website, where the commission does not exist and the customer data is yours.

The exception is the dark kitchen, which lives off the channel and therefore must be designed with the commission as part of the recipe from day one. For everyone else, the useful question is what share of your sales can depend on a third party without a fee change keeping you up at night.

Five myths your payout statement disproves

Myth: the commission is the contract rate. Reality: in the U.S. the most common average commission band is 20 to 24.9%, where 28% of restaurants using third-party delivery sit, per the survey the National Restaurant Association filed with the FTC in 2026. Payment fees, in-app campaigns and complaint adjustments come on top, so the rate that matters is the EFFECTIVE one: total deductions divided by gross sales for the month.

Myth: apps charge little in Latin America. Reality: the president of CANIRAC told El Universal San Luis (2026) that commissions in Mexico run 15 to 35% of total sales, and for a ghost kitchen selling almost entirely through apps that upper band becomes the second-largest cost after ingredients. Myth: more volume offsets the commission. Reality: if each order carries negative margin, selling more only speeds up the loss. What happens if a brand runs a self-funded 20% promotion to double orders?

Five myths your payout statement disproves — in practice

For example, with a 100 dish, 30 food cost, 25 commission and 6 packaging, 39 is left for labor and overhead; with the discount the price drops to 80, the commission to 20 and 24 is left, so doubling orders yields 48 instead of 39, nine more for twice the line work, twice the packaging and twice the complaint exposure. Myth: raising app prices scares customers off.

Reality: the delivery guest compares against the convenience of staying home, not against your dine-in menu. The tension (higher price versus algorithm visibility) is real, and it resolves by splitting menus: a shorter app menu of high-margin dishes does not compete with the dining room. Myth: opening a ghost kitchen solves the problem.

But a ghost kitchen built from scratch with no direct channel depends on a third-party algorithm for every order, and ghost kitchen vs restaurant is won or lost on app pricing, not on rent savings; ghost kitchen marketing starts with your own customer base and direct ordering from month one.