What Third-Party Delivery Actually Costs an Independent Restaurant

The published rates, as of 26 August 2026. DoorDash charges 15%, 25% or 30% commission on delivery orders depending on which of its Basic, Plus and Premier plans a restaurant is on, and 6% on pickup orders across all three. Uber Eats charges a 20% marketplace fee on Lite, 25% on Plus and 30% on Premium, with pickup at 7% where in-store pricing is validated and 10% where it is not. Grubhub charges a marketing commission that ranges from 5% to 20% depending on the package chosen, plus 10% on orders its own drivers deliver, plus an order processing fee of 3.05% + $0.30. Every figure in this piece was read off the platform's own merchant page and is linked at the point it appears.

//A worked example, with the assumptions on the table

To turn a commission percentage into a margin effect you need a cost structure to apply it to. The National Restaurant Association, writing on 8 July 2026, describes a typical independent restaurant before the pandemic as running a pre-tax profit margin of roughly 5%, with food and labour each accounting for approximately 33 cents of every dollar in sales and remaining expenses such as utilities and occupancy representing about 29% of sales. The same commentary estimates that total expenses for an average restaurant rose 36% between 2019 and 2026, that restaurant employee hourly earnings are up 41% and wholesale food prices up 35% over that period, and that 42% of operators said their restaurant was not profitable in 2025.

That table is a model, and it is wrong in at least three specific ways that cut in both directions. Labour on a delivery order is not the same as labour on a dine-in order: there is no table service and no bussing, but there is packing, bagging and order handling instead. Occupancy cost does not rise with an incremental delivery order, so charging the full 29% against it overstates the loss on marginal volume for a restaurant with spare kitchen capacity. And many operators raise menu prices on delivery channels, which the table deliberately does not model, because doing so would require inventing a markup figure. What the arithmetic does establish without ambiguity is the scale of the thing: against a 5% pre-tax margin, a 25% commission is five times the entire profit on the order. A delivery order can only work if at least one of the other lines behaves differently from the dine-in case.

It is worth saying plainly that there does not appear to be a published, methodologically described study isolating the margin effect of third-party delivery on independent restaurants. Restaurant Business, citing Technomic, reports delivery running at around 9% of traffic at limited-service chains and just under 4% at full-service chains, and the National Restaurant Association reports the profitability decline, but neither isolates commissions from food and labour inflation as the cause. Anyone quoting a precise figure for what delivery does to your margin is modelling, as the table above is, rather than measuring.

//The regulatory ceiling, where one exists

New York City is the only major US market with a hard statutory cap on what a delivery platform may charge a restaurant, and the cap is considerably higher than most coverage of it suggests. Section 20-563.3 of the New York City Administrative Code limits a third-party delivery service to 15% of the purchase price of an online order for delivery fees, 5% for basic service fees, 3% for transaction fees, and 20% for enhanced service fees — 43% in total where all four apply. The section was enacted by Local Law 2021/103, effective 24 January 2022, and amended by Local Law 2025/079, effective 30 June 2025, which is the amendment that added the 20% enhanced-services tier on top of what had previously been a 23% ceiling. The statute also requires a platform to refund fees charged above the caps within thirty days.

Outside New York City there is no federal cap and no equivalent statute in most jurisdictions, which means the published rate is the rate. A restaurant in a city without a cap is negotiating against a rate card, and the rate card is the one linked at the top of this page.

//Who owns the customer

The second cost of marketplace delivery is not expressed as a percentage. When an order arrives through a marketplace, the customer's name, delivery address, email address, phone number and order history sit with the platform rather than with the restaurant that cooked the food. New York City attempted to change that with a 2021 law requiring delivery platforms to hand that data to restaurants on request. DoorDash, Grubhub and Uber Eats sued. In September 2024 Judge Analisa Torres of the Southern District of New York held the law unconstitutional as compelled speech under the First Amendment, as Restaurant Dive reported at the time. On 5 August 2026 the Second Circuit affirmed, with Judge Park writing that the law compelled speech for what amounted to economic favouritism, and faulting an opt-out mechanism that required customers to decline data sharing order by order rather than once and permanently.

The practical position as of August 2026 is that no US jurisdiction compels a delivery platform to give a restaurant its customers' contact details. A restaurant on marketplace channels is buying orders, not customers. The order arrives, the margin on it is thin by construction, and the next time that person is hungry the relationship they have is with the app. That is not a hidden cost or a loophole. It is the product the platforms are selling, and it is precisely why the same platforms also sell commission-free direct ordering.

//Direct ordering, and what it actually costs

DoorDash's Storefront product carries 0% commission on orders placed through a restaurant's own website, with payment processing at 2.9% + $0.30 per order on its Boost and Pro packages and 3.3% + $0.30 on Starter. Grubhub markets a comparable commission-free product under the name Grubhub Direct. On a $30 order, 2.9% + $0.30 is $1.17 against $7.50 at DoorDash's 25% Plus rate — cheaper by a factor of six.

What those products do not include is demand. A marketplace commission buys placement in front of people who were not looking for you specifically; a direct-ordering page costs almost nothing because it does nothing at all to find the customer. Substituting one for the other only works to the extent that a restaurant already has some way to reach people — signage, a mailing list, regulars, a neighbourhood, a reason to be searched for by name. Treating a 0% ordering page as a replacement for a 25% marketplace, rather than as a cheaper channel for demand you already have, is the most common way this maths gets done wrong.

The pickup rates deserve a second look for the same reason. DoorDash charges 6% on pickup orders on every one of its plans, and Uber Eats charges 7% where in-store pricing is validated and 10% where it is not. On a $30 order, pickup at 6% costs $1.80 against $7.50 at the 25% delivery rate. Any share of marketplace demand that can be shifted from delivery to pickup is a four-fold reduction in commission on those orders without leaving the platform at all, and without needing a mailing list to do it.

//Four things worth checking on your own account

• +Which plan tier each platform actually has you on, and what the tier below it costs in visibility. The gap between DoorDash Basic at 15% and Premier at 30% is half your commission bill.
• +Whether your Uber Eats pickup orders are being charged 7% or 10%. The lower rate is conditional on validated in-store pricing, which means your platform menu prices matching your in-store prices.
• +Whether your menu prices differ between channels and by how much, since that markup is the single largest lever on whether a delivery order clears its own commission.
• +What share of delivery orders are repeat customers you could reach directly if you had their contact details — which, on the law as it stands after the Second Circuit's ruling of 5 August 2026, you generally will not receive from the platform.