Are DoorDash and Uber Eats worth it for a small restaurant?

Yes, on two conditions: the orders have to be ones you would not otherwise get, and the price on the app has to carry the app's fee. Checked against the platforms' own pricing pages today, DoorDash takes 15% to 30% of each delivery order, Uber Eats 20% to 30%, and Grubhub 5% to 20% in "marketing commission" plus a delivery fee that starts at 10%. Set against that, a Stanford study of Visa transaction data on what the apps do to a restaurant's sales found that roughly half of every dollar spent through a delivery app is new money and the other half replaced sales the restaurant would have made anyway (Collison, Stanford, 2020). So the honest shape of the answer is this: the lowest-commission plan, priced so the guest pays the fee, is usually worth it for the customers you would never have met; the 25–30% plans, feeding orders from people who already know you, usually are not. The rest of this page gives you the numbers behind that and a way to move your regulars to a cheaper door.

What do the delivery apps actually charge a restaurant?

These are the published rates on each platform's own pricing page as opened on 2026-09-07. Your contract can differ ("pricing in select markets may differ," Grubhub notes twice), and some cities cap what the apps may charge (see below).

Sources: DoorDash merchant pricing, Uber Eats merchant pricing, Grubhub pricing and fees. None of the three charges a monthly fee or a contract on the marketplace plans; DoorDash's page says "no activation fees, subscription fees, software fees, or cancellation fees," with a tablet at "$6/week after your trial" as the one optional hardware cost. Grubhub adds "an order processing fee to cover credit card processing and fraud monitoring" that its pages do not put a number on (Grubhub FAQ).

Two things on that table are easy to miss. Uber Eats' cheapest marketplace plan is 20% on its page today; if you were quoted 15% somewhere else, check the live page. And Uber Eats charges a lower pickup fee (7% rather than 10%) only if your app prices match your in-store prices, which matters for the pricing question below.

Some cities cap these fees by law. New York City's administrative code, amended June 30, 2025, limits a delivery fee to 15% of the purchase price, a basic listing fee to 5%, an optional enhanced-service fee to 20%, and a transaction fee to 3% (NYC Admin. Code § 20-563.3; NYC DCWP fee-cap flyer, April 2026). If you are in a city with a cap, the plan prices above do not apply to you as written; ask the platform for your city's schedule. The Independent Restaurant Coalition is asking Congress for a 15% cap nationally, along with a ban on the apps restricting what restaurants charge on their own menus (IRC, 2025).

Do app orders bring new customers, or just move your own?

Both, in roughly equal halves. The clearest study on the question used Visa's transaction data, about 1.47 million cards and 60.78 million restaurant transactions from 2014 to 2017, and compared what those cardholders spent at restaurants before and after delivery platforms arrived in their area. Its finding: "roughly half of each dollar spent on online food delivery services is new, whereas the remaining half is converted from in-person restaurant sales" (Collison, 2020).

The halves are not evenly spread, and the pattern is useful. Rural areas saw almost entirely new money (99.7 cents per dollar); urban areas about 47 cents. Weekday app sales were about 47 cents new; weekend sales only 29 cents, meaning a Saturday-night app order was more likely to be someone who would have come in. Lower-income zip codes saw far more new money (about 90 cents) than higher-income ones (about 46 cents). And the share that was new fell over the study period, from about 57 cents to about 46 cents, as people got used to ordering in. The paper's own back-of-envelope estimate is that the platforms raised restaurant revenue about 1.2% and lowered profit about 1.8%, because the fee applied to all of the order and only half of it was new.

The data is from before the pandemic, and delivery is a bigger part of dining now. The National Restaurant Association's 2025 off-premises report puts off-premises traffic at 30% for full-service restaurants in 2024, up from 19% in 2019, and at 83% for limited-service, up from 76%; 37% of consumers order delivery weekly, and 82% say they would order delivery more often if they could afford it (Restaurant Dive on the NRA report, April 2025). Demand is there. The question is only what each order leaves you.

What does one app order actually leave you?

Do the arithmetic on your own numbers, but here is the shape of it. Take a $40 order on a 25% plan. The platform keeps $10. If your food and packaging on that order run 35 cents on the dollar (your figure will differ), that is $14, which leaves $16 before labor, rent, and everything else. The same order at your own counter, paid by card, leaves you $40 minus roughly $1.50 in card processing (your rate will differ) minus the same $14, about $24.50. The app order is worth having only if it would not have happened at the counter. That is the whole test, and Collison's halves say about half of app dollars fail it.

Two more costs sit outside the commission. DoorDash's Plus and Premier plans buy a wider delivery radius, lower delivery fees for the guest, and on Premier an in-app ad, which is how DoorDash's own page describes the difference from Basic; Uber Eats' Premium matches ad spending up to $100 a month. The cheaper plans get you listed, not promoted. And an order the app cannot get picked up is yours to eat. Javier Trujillo, who runs five places in Omaha and dropped the apps after paying $188,000 in a year to services that "push our menu to their customers," described it plainly: "Maybe there was not enough Dashers in the area and they took the order, we get the order, we made the food, but nobody comes to pick it up. That's losses on us" (Entrepreneur, April 2026). Georg Liakopoulos, whose group runs The Hollywood Grill, put the same feeling in one line: "they charge you 20, 25%. They're partners in your business without them working" (WNUR, April 2026).

Should you charge more on the app than in the restaurant?

Most restaurants do, and guests know it. Across 30 large national brands the average price premium on delivery marketplaces versus dine-in was 19% and had nearly doubled since 2020, per Gordon Haskett Research Advisors as reported in Nation's Restaurant News (NRN, July 2024). A 2025 secret-shopper study of more than 300 delivery orders found third-party entrées priced $1.85 above the restaurant's own menu (Restaurant Business on the Intouch Insight study, 2025). On the guest side, 40% of the 850 people Toast surveyed said higher prices are what frustrates them most about delivery apps (Toast, August 2024 survey).

The workable position for a small place is the one the price-premium figures above suggest most restaurants have settled on: the app price carries the app's fee, and your own menu, at your counter and on your own page, stays at your price. Two cautions. Uber Eats charges its lower pickup fee only with "validated in-store pricing," so raising app prices can cost you three points on pickup orders there; check your plan. And the gap should be honest and consistent, because a guest who orders from the app and then sees your menu at the counter will notice, and the fee is the reason you can give them.

What goes wrong that you will be blamed for?

Wrong and late orders, and the guest blames you, not the app. In a 2022 consumer survey, 63% of delivery users had received an incorrect order; 66% blamed the restaurant for an incorrect, late, or missing item; and afterward 24% asked for a refund, 20% said they would not reorder from that restaurant, and 17% posted a negative review (Restaurant Dive on the eduMe study, May 2022). The 2025 secret-shopper study found accuracy about the same on both channels (89%) but batching, where a driver carries several orders at once, on 19% of third-party deliveries versus 4% of first-party ones, and delivery times of 35 minutes or more versus about 30 (Restaurant Business, 2025).

The part that hurts most is that you usually cannot reach the guest to fix it. DoorDash's merchant help page describes contacting a customer about an order by entering your own number and having DoorDash connect the call; it does not describe giving you the customer's number or email (DoorDash merchant help). For most app guests you will never have a name or a way to say sorry, which is why so many wrong orders end as a review instead of a phone call.

How do you get regulars to order from you directly?

Give them a door that is at least as easy as the app, and tell them about it in the bag. Most guests say they would rather: 58% prefer a restaurant's own app or website for delivery, against the third-party apps, with convenience (65%), easier customization (50%), and loyalty rewards (36%) as the reasons (Restaurant Dive on the NCR Voyix 2025 report). What they do is different: in Toast's survey 48% said they use DoorDash most often and 33% a restaurant's own app, and the split runs by age, with 54% of guests over 54 preferring the restaurant's own site and 59% of 18-to-24-year-olds preferring DoorDash (Toast).

The cheap way to close that gap is to use the direct-ordering tools the platforms themselves now give away, because they want the delivery fee even when they lose the commission. DoorDash's Online Ordering is "commission-free" on orders from your own site and from the "Order online" button on your Google Business Profile, with only payment processing charged; it requires an active DoorDash store (DoorDash Online Ordering; DoorDash help: Google ordering). Uber Eats' Webshop charges "2.5% order processing fee + $0.29 per order" (Uber Eats pricing). Grubhub Direct advertises "zero marketing commission for online orders" and no platform or hosting fees (Grubhub pricing). If you would rather not run your own door through a marketplace, ChowNow's plans are $249, $349, and $449 a month, plus a $119–$499 setup fee and 2.95% + $0.29 per transaction (ChowNow pricing); Square Online has a free plan at 3.3% + 30¢ per online card payment, and $49 and $149 a month plans at 2.9% + 30¢ (Square Online plans).

Then tell people. The guest who ordered through the app is holding your bag; a card in it with your own ordering link and a reason to use it (your price, no markup, a free side next time) is the one message the app cannot intercept. Expect the shift to be slow, and be honest with yourself about why guests use the apps: the secret-shopper study found third-party orders came with a promotion 52% of the time against 32% for direct, and the guest's delivery fee was actually lower on the apps ($5.14 against $5.96 direct) (Restaurant Business, 2025). Your direct door wins on price of the food and on the relationship, not on the delivery fee.

So should a small restaurant be on the apps?

Be on them the way you would rent a billboard: for reach, at the cheapest rate that gets you seen, with prices that cover the cost. Start on the lowest-commission plan and use the 0% introductory weeks to see how many orders are from postcodes and names you do not recognize. Price the app menu to carry the fee and keep your own menu at your price. Turn on the platform's commission-free direct ordering and put the link in every bag and on your Google profile. Reconsider the 25–30% plans only if the extra placement demonstrably brings orders you could not get otherwise; the guarantee on DoorDash's Premier plan (a month's commission refunded if you get fewer than 20 orders, with conditions) tells you the platforms know that is the question. And keep the thing the apps cannot give you: a way for the guest to reach you, in your own words, when something goes wrong.

That last part is what our own product does, for what it's worth. Brief First gives your restaurant a small card and a page of its own that you edit yourself, with your menu at your prices and a link out to wherever you take direct orders, and a private line on that page where a guest can tell you, in their own words, that the order was wrong, so a mistake becomes a callback the same day rather than a review. The page and the private line are free, with no card and no clock. What costs money, from $30 a month, is the AI host on the card that answers menu questions in the guest's own language and sends guest requests to your team. It never takes payment and touches no commission. And the advice above works whether you use us or a printed card in the bag.