Direct ordering vs DoorDash and Grubhub

Every restaurant operator we talk to in the US has the same gut feeling about marketplaces: the orders are real, the volume is real, and the margins are not. Yet most operators keep paying a commission on every single delivery order to DoorDash, Grubhub and UberEats — up to 30% on the top tiers — because they think the alternative is zero orders. That binary is wrong. Direct ordering is not a replacement for marketplaces — it’s the channel you build underneath them so the same customer eventually becomes yours instead of theirs.

What do DoorDash, Grubhub and UberEats actually charge?

All three publish their rates on their own merchant pages. As of August 2026: DoorDash sells three Marketplace plans — Basic at 15%, Plus at 25% and Premier at 30% on delivery orders — with pickup at 6% on all three plans, provided your DoorDash pickup menu prices match your in-store prices. Grubhub publishes a marketing commission of 5%, 15% or 20% by plan (Basic, Plus, All-access), with its own delivery fleet starting at a further 10% on top, and states that those rates exclude payment processing. UberEats lists Lite at 20%, Plus at 25% and Premium at 30% on Marketplace orders, 15% if you deliver with your own staff, and 7% on pickup when you validate that in-app pickup prices match in-store (10% if you don’t). Rates vary by city in regulated markets — check the plan page for yours.

Take a $50 order. Here’s what actually lands in your bank account on each channel:

• DoorDash Marketplace, Premier tier (30% delivery): $50 gross, minus $15 commission. You keep $35.00.
• DoorDash Marketplace, Basic tier (15% delivery): $50 gross, minus $7.50 commission. You keep $42.50— but Basic buys you the least discoverability, so you’re paying the marketplace tax with the marketplace benefit turned down.
• DoorDash pickup (6%, all plans): $50 gross, minus $3. You keep $47.00 — the cheapest thing a marketplace sells you, and the one their reps mention least.
• Direct ordering on your own storefront: $50 gross, minus 2.9% + $0.30 if you run the card on your own Stripe account, zero commission to anyone. You keep ~$48.25.

That’s a $13.25 swing on every $50 order between a Premier delivery order and the same order placed direct. Run 80 orders a week through direct instead and that’s roughly $55,000 a year you don’t hand to a third party. For most independents, that’s the difference between “tight month” and “hire another line cook.” (See FoodyOS vs Grubhub for a side-by-side breakdown of the same math at the platform level.)

Customer data is the bigger story

Commissions are the obvious cost. The hidden cost is that the customer you served isn’t your customer— they belong to the marketplace. You don’t get the phone number. You don’t get the email. You can’t text them next Tuesday with a slow-night offer. You can’t see that they’ve ordered the same Pad Thai 11 weeks in a row and proactively send a free spring roll on visit 12.

Marketplaces deliberately hold that data. It’s their core asset. DoorDash spent the last decade building a logistics business on top of your food, and the part they kept was the relationship. Direct ordering inverts that: every order writes the customer’s phone, email and order history into your CRM, where you can use it for retention, win-back and dine-in upsells without paying a per-message toll to a third party.

What the marketplaces’ own “direct” products are really doing

DoorDash, UberEats and Grubhub all noticed that operators were trying to escape commissions and shipped their own “direct” products: DoorDash Drive / Storefront, UberEats Webshop / Uber Direct, and Grubhub Direct. The pitch is the same in each case — run a commission-free site under your own URL, we just handle the delivery driver and (sometimes) the payments.

On paper this looks like the answer. In practice it’s a soft trap. The delivery layer is locked to a single courier network you can’t price-shop, and it’s priced per drop rather than as a commission you can shrink: Uber publishes Uber Direct — on-demand delivery from your own site — as starting at $7.99 per delivery, and its commission-free Webshop at 2.5% + $0.29 per order (August 2026). On a $30 ticket, a $7.99 drop fee alone is over a quarter of the order — even though you technically “own” the channel. We break the trade-off down in detail in in-house delivery vs DoorDash Drive. The short version: a marketplace-owned direct product solves the commission problem for the marketplace, not for you.

The first tell is the data layer, and it’s a question you can settle before you sign anything. Ask the rep, in writing: on the day I leave, exactly which fields do I get back, in what format, and can I pull them myself without opening a ticket? With a true direct stack the answer is trivial, because the customer list, item-level order history and delivery preferences sit in a database you control. With a marketplace-run “direct” product, that question is the whole negotiation. If the answer is vague, you’re renting a checkout button, not owning a channel.

Routing is the second. A marketplace-run direct product still routes the delivery through the marketplace’s own courier network at a fixed per-drop fee, with no ability to swap in a cheaper local courier on slow nights or to dispatch your own driver on dense routes. A standalone direct stack lets you pick the courier per zone, batch deliveries, or run dine-in pickup at zero variable cost. That flexibility is where the real margin lives, and it’s precisely what the marketplace direct products are designed to prevent.

Marketplace cannibalization is real but overstated

The argument from DoorDash reps is always: “Customers don’t even know you exist without us. We bring you incremental orders.” That’s true the first time a customer orders. It is not true the third, fifth, or fifteenth time.

You can test the claim yourself in about twenty minutes, and it costs nothing. Pull a month of marketplace orders and count how many names, phone numbers or addresses show up more than once. Every repeat in that list is a customer who already knows your restaurant and would order direct if a direct option existed and were visible — which means that order isn’t incremental discovery, it’s the marketplace taxing one of your regulars. There is no public figure for what share of marketplace orders are repeats, because the marketplaces don’t publish it. Your own order file is the only honest source, and you already have it.

Regulators have already picked a side

Major US cities are increasingly skeptical that 30% marketplace commissions are economically sustainable for independent restaurants. New York City made its pandemic-era third-party delivery fee cap permanent through Local Laws 52 and 88 of 2021, and the Council has revisited the numbers since — the caps in force at any given moment are the ones the city’s Department of Consumer and Worker Protection publishes on that page, so read it before you model anything. San Francisco and Seattle have their own caps. California passed AB 286 in 2021, which forces delivery platforms to itemize the breakdown of what they charge restaurants and customers, prohibits markups above the restaurant’s menu price without consent, and requires 100% of tips to flow to the courier — closing several of the games marketplaces used to play with prices and gratuities.

None of this fixes the underlying commission economics outside the capped cities, but it tells you which way the wind is blowing: even legislators who don’t love regulating prices have decided 30% commissions are extractive. Building a direct channel now is the operator-side version of that same conclusion — and it works in every state, with no waiting for a city council to act.

The caution that goes with all of it: caps get amended, challenged and re-litigated, and they only bind in the cities that passed them. Treat any specific cap figure as a moving target and confirm it against the city’s own page before you put it in a spreadsheet. The part that doesn’t move is the one you control — on an order placed on your own storefront, the take rate is zero by construction, in every state, with no city council involved.

The long tail belongs to whoever owns the relationship

A regular orders 30 times a year. At an average ticket of $40, that customer is worth $1,200 in revenue. On DoorDash’s Premier tier at 30%, you net about $840 from that customer over the year. Direct, you net about $1,160. That’s $320 per regular per year — and a single restaurant has dozens to hundreds of regulars.

The play isn’t to ban DoorDash. The play is to build a direct channel so when that customer orders for the second time, they have somewhere to go that doesn’t cost you $12.

How direct and marketplace coexist

The framing that works: marketplaces are an acquisition channel, direct is a retention channel. You pay DoorDash a finder’s fee for the first order, then you pull the customer onto direct as fast as possible. Tactics that work in practice:

• Insert in every marketplace bag:a printed card that says “Order direct next time, save 15% with code DIRECT15.” That discount is half what DoorDash’s top tier charges you, and it converts.
• Loyalty tied to your direct channel only: points accrue when they order from your site or WhatsApp, not from DoorDash.
• Slightly higher prices on marketplaces:many operators price the marketplace menu above the direct menu, and the gap is itself an advertisement for ordering direct. Read the agreement before you set it — DoorDash’s 6% pickup commission, for example, is conditional on your pickup prices matching your in-store prices — and check your state’s menu-pricing disclosure rules.
• QR code at every dine-in table that goes to your direct ordering page, not a Yelp listing or marketplace.
• A reorder link on the receipt— WhatsApp or a short URL back to your own storefront: the fastest path from “they ate” to “they reorder direct.”

What does moving volume to direct actually save?

Do the arithmetic on your own numbers rather than trusting anyone’s projection, including ours. Take a restaurant doing $1M a year in delivery and move a quarter of it — $250,000 — off a 30% marketplace tier and onto its own storefront. On the marketplace, that volume costs $75,000 in commission. Direct, running the same cards on your own Stripe account at 2.9% + 30¢, it costs about $7,500. The gap is roughly $67,000, and nothing about it requires losing a marketplace customer — the marketplace channel keeps running the whole time.

The shape holds at any volume, which is why it’s worth running with your real mix in front of you: the marketplace take rate is a percentage of everything, and your processor’s rate is a percentage of a much smaller number. Every point of mix you move is a permanent line-item reduction, not a promotion that expires.

Direct ordering doesn’t kill DoorDash. It just stops you from paying DoorDash for customers who were already yours. The full playbook for migrating volume off marketplaces lives in how to stop paying third-party delivery fees, and the all-in cost of running direct is on our pricing page — flat per location, no per-order commission.

Sources

1. DoorDash for Merchants, “Merchant Fees & Pricing” (Basic 15% / Plus 25% / Premier 30% delivery; 6% pickup on all plans), accessed August 2026 — merchants.doordash.com/en-us/pricing
2. Grubhub for Restaurants, “Grubhub pricing and fees” (Basic 5% / Plus 15% / All-access 20% marketing commission; Grubhub Delivery from 10%; rates exclude payment processing), accessed August 2026 — get.grubhub.com/grubhub-pricing-and-fees/
3. Uber Eats Merchants, “Pricing” (Lite 20% / Plus 25% / Premium 30% Marketplace; 15% self-delivery; 7% pickup with validated in-store pricing; Uber Direct from $7.99 per delivery; Webshop 2.5% + $0.29 per order), accessed August 2026 — merchants.ubereats.com/us/en/pricing/
4. NYC Department of Consumer and Worker Protection, “Third-Party Food Delivery Services” (Local Laws 52 & 88 of 2021; current fee caps as published by DCWP) — nyc.gov/site/dca/businesses/3rd-party-food-deliverers.page
5. California Legislature, AB-286 — Food delivery: purchase prices and tips (2021) — leginfo.legislature.ca.gov — AB-286