Is Delivery Actually Profitable for Restaurants?

Revenue is not the question

Ask most operators how delivery is performing and you get a revenue figure. It is the number the platform dashboard shows, the number that appears in the weekly sales summary, and the number that makes the channel feel like it is working. It is also the one number in the entire discussion that tells you nothing useful.

Delivery revenue is gross. Between it and your bank account sit a commission rate somewhere between twenty and thirty-five percent, packaging that nobody costed, and promotional discounts that were agreed to in order to keep visibility in a crowded listing. By the time those have been taken, a channel producing a fifth of your sales may be producing a very small fraction of your profit — or none.

The question worth answering is not whether delivery sells. It plainly does. The question is what one order is worth once the channel has taken its share, and whether that is more or less than what you would have earned doing something else with the same kitchen minutes.

Run one order all the way through

Averages hide everything here, so take a single order. A $30 basket on a platform charging 30% commission, with a 15% off promotion you agreed to fund, a 32% food cost, and $1.40 of packaging.

The discount removes $4.50, so the guest pays $25.50. Commission of 30% on that takes $7.65, leaving a payout of $17.85. Food cost of $9.60 and packaging of $1.40 come out of that payout, not out of the $30. What is left is $6.85 — a shade under 23% of the price the guest saw on the screen.

Now sell the same food at the same price to someone at a table. No discount, no commission, no packaging. Gross profit is $20.40. The dine-in order is worth almost exactly three times the delivery order, from an identical amount of cooking.

That ratio is the most useful thing you can know about your delivery channel, and it varies enormously between dishes. A high-margin, single-container dish might land at 1.5 delivery orders per cover. A discounted, multi-container dish with an expensive protein can pass five, at which point every delivery ticket that delays a table is straightforwardly destroying money.

The three deductions dine-in never makes

Commission

The largest and the least negotiable. It is worth knowing precisely what your rate covers — delivery, marketing placement, payment processing — because rates quoted in conversation are rarely the rate on the settlement report once optional marketing spend is included. Read one full statement line by line before you accept the headline number.

Packaging

The cost operators forget most consistently, because it arrives on a different invoice and never appears on a recipe card. Containers, lids, bag, seal, label, cutlery and sauce pots regularly reach one to two dollars an order. On a small basket that is more than the remaining margin. Cost it as an ingredient on every delivery dish rather than carrying it as an overhead, because overheads get ignored and ingredients get priced.

Self-funded discounting

The one that does the most damage per dollar, because it compounds with commission. Cutting 20% off the price loses you a fifth of revenue and saves you only a fifth of the commission on that fifth. In the example above, adding a 15% promotion cost $4.50 and saved $1.35, wiping out nearly half the order's profit. It is entirely possible to run a successful promotion that raises order count, raises kitchen workload, and lowers absolute profit.

What delivery does not cost you

The comparison above is deliberately unkind to delivery, and it is worth being honest about the other side. A delivery order does not occupy a table, does not need a server, does not use glassware or linen, and does not generate a share of the laundry and breakage bill. It does not consume a cover during your two hours of peak demand unless you let it.

Most importantly, it uses capacity you have already paid for. Rent and the bulk of kitchen labour are fixed for the shift. An order arriving mid-afternoon when the dining room is a third full is close to pure contribution, and a 20% margin on that order is a genuinely good outcome because the alternative was zero.

This is why the honest answer to whether delivery is profitable is that it depends entirely on when the order arrives. The same $6.85 is excellent at three in the afternoon and expensive at eight on a Saturday, and no single channel-level margin figure can tell those two apart.

The incremental versus cannibalised test

The decision that matters is whether a delivery order is incremental — business you would not otherwise have had — or cannibalised, meaning it replaced a dine-in ticket that would have been worth several times as much.

There is no perfect way to measure this, but there are useful proxies. Look at delivery orders by hour against dining room occupancy by hour. If the two peaks coincide and your kitchen is at capacity during them, a meaningful share of delivery is displacing better business. If delivery peaks fill your quiet periods, it is doing exactly what it should. Comparing dine-in covers before and after a platform launch, using a like-for-like basis, tells you more than any platform report will.

The operational answer, once you know the pattern, is usually not to leave the platform. It is to cap delivery at peak, either by pausing the tablet or by lengthening quoted prep times, and to lean into it hard everywhere else.

Making delivery work

• Build a separate delivery menu. Not your whole menu with a price uplift — a shorter list of dishes that travel well, pack into one container and carry margin. Removing your three worst travellers usually improves channel margin more than any rate negotiation.
• Price delivery 10–15% above dine-in. Standard practice, broadly expected, and it recovers packaging plus a slice of commission.
• Cost packaging into every dish. If it is not on the recipe card it is not in your price.
• Discount narrowly or not at all. A promotion on one high-margin dish, or a spend threshold that lifts basket size, behaves nothing like a blanket percentage off everything.
• Watch the labour. Packing and handing over orders is real work. If delivery volume has grown and your rota has not, service is degrading somewhere you are not looking.
• Apply menu engineering to the channel. Run your delivery dishes through the same popularity and margin grid you use in-house. The winners are usually different dishes.

None of this requires leaving a platform, and for most independent restaurants leaving is the wrong move — the visibility is worth something real, and the orders that would follow you directly are fewer than optimism suggests. What is required is treating delivery as a distinct business with its own menu, its own prices and its own margin target, rather than as your existing business sold at a slightly worse rate.

Work out the per-order number for your five best-selling delivery dishes. It takes an afternoon, and it usually changes at least one decision immediately.

Frequently asked questions

What commission rate do delivery platforms charge?

It varies by market, by contract and by which services you take. What matters is not the headline rate but the effective rate on your own settlement reports once optional marketing spend, promotional contributions and payment fees are included. Divide total deductions by total order value for a quarter and use that figure.

Is it better to raise delivery prices or cut portions?

Raise prices. A modest uplift on delivery menus is expected and rarely commented on. Portion reductions are noticed immediately, get photographed, and turn a margin problem into a reputation problem that is far more expensive to fix.

Do platform promotions ever pay for themselves?

Sometimes, when they lift basket size rather than cut price — a free item on a spend threshold, for example. Blanket percentage discounts almost never do, because they reduce revenue by more than they reduce commission and the incremental orders rarely make up the gap.

How do I know if delivery is stealing my dine-in trade?

Plot delivery orders per hour against dining room occupancy per hour. Overlapping peaks in a kitchen that is already at capacity means displacement. Delivery filling your quiet hours means genuine incremental business. The pattern is usually obvious within two weeks of data.

Should delivery have its own food cost target?

Yes, and a tighter one. Because commission takes a fixed share of the top line, a delivery dish needs a lower food cost percentage than a dine-in dish to reach the same margin. Running the same target across both channels guarantees the delivery version underperforms. See how to price a menu item for the method.

Is a direct ordering website worth building?

The saving is the entire commission line, which is larger than almost any promotion you would run to attract the order. The catch is that traffic does not appear on its own, so budget for the marketing you will need. Treat it as a channel you have to build demand for, not a cheaper version of one that already has demand.

Does delivery affect prime cost?

Food and labour for delivery orders sit inside prime cost; commission usually sits below it as an operating expense. That accounting split is why delivery can look harmless in a prime cost review while quietly damaging net profit. Track the channel separately as well as within the total.

Run the numbers

Use the free Delivery Commission Profit Calculator to apply everything above to your own figures.