Third-Party Delivery Economics for Restaurant Owners

When delivery apps arrived, they looked like free incremental revenue: orders the restaurant would never have captured, fulfilled from the kitchen that was already running. Years on, the operators who survived the channel learned a harder truth, that a delivery order is the most expensive revenue in the building, and the order total on the screen is the most misleading number an operator can fixate on. The economics are not bad, exactly, but they are completely different from the dining room, and treating them as the same is how restaurants lose money on revenue that felt like a gift. This guide is the operator's framework for making delivery pay, and it builds on disciplined food cost percentage control, because delivery quietly distorts the food cost number operators rely on.

The Commission Is the Headline, but Not the Whole Story

Marketplace commissions on DoorDash, Uber Eats, and Grubhub commonly run 15% to 30% of the order total, scaling with the service tier the restaurant selects, where higher tiers buy more marketing placement and visibility. According to the National Restaurant Association, these fees are the single largest reason a delivery order can be unprofitable even when it presents as incremental revenue. A 28% commission on a $40 order is more than $11 gone before a single other cost, and that is the part operators can see.

The costs they often miss compound the problem. Packaging adds real per-order expense that never appears on the plate cost. The extra labor to assemble, bag, and hand off delivery orders during a rush competes with the dining room. And critically, delivery menus rarely carry the high-margin drinks and desserts that make dine-in checks profitable, so the channel skews toward the lower-margin entree alone. A dish with a healthy 70% dine-in contribution margin can fall to breakeven or a loss after commission and packaging. The order total looks like a win; the channel margin tells the real story.

A $40 Order, Line by Line

The abstraction of a commission percentage hides how thin the channel really is, so it pays to trace one order all the way down. Take a $40 delivery order on a dish that carries a 70% dine-in contribution margin, meaning about $12 of food cost. A 27% marketplace commission removes $10.80. Packaging (the clamshell, the bag, the cutlery, the label) commonly runs $1.50 to $2.50 per order according to widely reported restaurant operating figures, so call it $2. The incremental labor to assemble and hand off the order during a rush is real but hard to isolate, so leave it at zero to be generous. After food, commission, and packaging, that $40 order leaves roughly $15.20, a contribution margin of about 38%, against the 70% the same plate earns in the dining room.

Now apply the missing-mix problem. The dine-in version of that guest would likely have added a $7 drink and a $6 dessert, both high-margin, lifting the check and its dollar contribution well above the entree alone. The delivery order rarely includes either, so the channel does not just shave margin off the plate, it strips out the most profitable courses entirely. This is why an order total that looks like found revenue can net almost nothing once it is fully costed, and why the National Restaurant Association repeatedly flags delivery as the channel most likely to be unprofitable while appearing to grow the top line.

Read the Fee Stack, Not Just the Headline Rate

Operators often quote a single commission number, but the marketplace deduction is usually a stack of separate fees, and reading the statement line by line is the only way to know the true take rate. A typical breakdown separates a base delivery commission, an optional marketing or sponsored-placement fee that buys higher search ranking inside the app, and payment-processing charges on the transaction. An operator who layers a sponsored-listing campaign on top of a mid-tier commission can push the effective deduction well past the headline rate they think they signed up for, which is how a 20% plan quietly behaves like a 30% plan.

The discipline is to reconcile the marketplace payout statement against gross order value every period, exactly the way a restaurant reconciles its merchant-services statement. The effective take rate, total fees divided by total order value, is the number that belongs in the channel margin calculation, not the plan's advertised commission. Operators who track the effective rate catch fee creep from promotions and placement spend that the advertised number conceals, and they can decide deliberately whether each layer of marketing fee returns enough incremental orders to justify itself.

Track Delivery as Its Own Profit Center

The reporting trap is that commission is a separate line item that never touches the kitchen, so the food cost percentage on a delivery plate can look perfectly normal while the true channel margin is dismal. Operators who blend delivery into a single food cost number are flying blind on the channel that is most likely to be losing them money. The fix is to track delivery as its own profit center, with commission, packaging, and the channel-specific labor and menu mix all included, so the real margin is visible rather than buried in a blended average.

That separate accounting is what turns delivery from a guess into a decision. Once an operator can see the true contribution margin per delivery order, they can answer the questions that matter: which items are worth offering on delivery, what the channel actually contributes after all costs, and whether a given commission tier earns its placement. Run a representative delivery order, net of commission and packaging, through a profit margin calculator to see the channel margin in dollars, and the strategy questions answer themselves.

Channel Pricing and Virtual Brands

Many operators raise delivery-channel prices 10% to 20% to offset commission, which the major platforms permit, and the trade-off is price perception for any guest who orders both ways. The cleaner framing is a deliberate channel-pricing decision: price delivery to protect a target contribution margin after all delivery-specific costs, and accept that the channel economics differ from the dining room rather than pretending they match. This connects directly to menu engineering and pricing, because the delivery menu deserves its own contribution-margin analysis, not a copy of the dine-in prices.

Virtual brands, delivery-only concepts run out of an existing kitchen and listed separately on the apps, are the other lever operators reach for. The genuine appeal is filling idle kitchen capacity and spreading fixed costs across more orders, which improves the economics of a kitchen that has slack hours. The risk is real: a virtual brand still pays full marketplace commission and adds operational complexity, so it only works when the incremental orders truly use spare capacity rather than competing with the core business for the same line during the same rush. A virtual brand that slows down your main concept is a net loss dressed as growth.

Errors, Refunds, and Chargebacks: The Hidden Tax on Thin Margins

On a channel that already nets a fraction of dine-in margin, order defects do disproportionate damage. A missing side, a wrong modifier, or a late handoff frequently results in a refund or credit that the restaurant absorbs, and because the underlying order was already low-margin, a single refunded order can wipe out the contribution from several clean ones. The platforms also assign error charges and ratings penalties that, left unmanaged, push a restaurant down in app ranking and force more spending on marketing fees to stay visible, a compounding cost that never appears as a line on the plate.

The defense is operational, not financial. A dedicated expo step for delivery tickets, a sealed-bag policy that doubles as tamper evidence, and a short pre-handoff checklist catch the errors that turn into refunds before they leave the kitchen. Operators who measure their delivery error rate as deliberately as their food cost variance treat it as what it is, a direct tax on the thinnest revenue in the building. The same precision that closes the gap between theoretical and actual cost in food cost percentage control applies to delivery defects, where the dollars saved are worth more because the margin underneath them is so slim.

What Shifted in 2025 and 2026

The delivery landscape kept moving, and operators repricing the channel on old assumptions are working from stale numbers. Several major cities extended or made permanent caps on the commission a marketplace can charge, originally introduced as emergency measures, and the platforms responded by adding customer-facing fees and adjusting how marketing tiers are priced, which shifts cost around rather than removing it. According to National Restaurant Association reporting, off-premise demand settled at a structurally higher share of restaurant sales than before 2020 rather than receding, so delivery is now a permanent part of the operating model for most full-service concepts rather than a temporary add-on.

At the same time, first-party ordering technology matured, with commission-free or flat-fee direct-ordering tools becoming standard rather than a custom build. That lowered the barrier to the retain-direct half of the strategy, making it realistic for an independent operator to capture repeat customers off the marketplaces without an enterprise budget. The strategic picture is unchanged, the marketplaces are still customer acquisition, but the tools to convert that acquisition into owned, full-margin repeat business got materially cheaper and easier to deploy in this period.

The Real Strategy: Acquire on the Apps, Retain Direct

The reframe that makes delivery work is treating the marketplaces as a customer-acquisition cost rather than a sales channel. The apps own the discovery and the demand, so they are genuinely useful for putting a restaurant in front of customers who would never have found it. The commission is the price of that introduction. The mistake is paying that introduction fee over and over on the same repeat customer, when the goal should be converting that customer to first-party ordering after the first or second order.

First-party online ordering, where the guest orders directly from the restaurant website with the restaurant arranging or skipping delivery, avoids the marketplace commission entirely and keeps the customer relationship and the data. The apps are the top of the funnel; direct ordering is where the margin lives. A practical strategy uses the marketplaces for reach while steadily nudging regulars to direct, recovering 15% to 30% on every order that converts. Delivery economics sit alongside food, labor, and turns as a core part of the operating picture, and the hospitality operator toolkit with the restaurant profit margin benchmarks connect the channel to the full economics.