Third-party delivery is sold as incremental revenue. Sometimes it is. Often it is the same customers, ordering the same food, at a materially worse margin — and the only way to know which you have is arithmetic.
The number that matters
Commission on delivery orders typically runs somewhere between the high teens and thirty per cent of the order value, depending on the platform and which service tier you are on.
Run it against a dish rather than in the abstract.
A $20 dish with a 30% food cost has $6 of food in it, leaving $14 before labour, rent and everything else. At 25% commission the platform takes $5. You now have $9, and you still have to cook it, package it and pay for the packaging.
So a dish contributing $14 in the dining room contributes about $8 through delivery once packaging is counted. That is not a disaster — it is just a different business, and it has to be priced and portioned as one.
Whether it is incremental
The whole case for delivery rests on it being extra orders you would not otherwise have had. Test it rather than assuming.
Did your dine-in covers fall when delivery started? If delivery is doing forty orders a week and dine-in dropped by twenty-five covers, most of it is cannibalisation at a worse margin.
Is delivery volume concentrated in your quiet hours? That is genuinely incremental and the best case — it fills kitchen time you were paying for anyway.
Is it concentrated in your peak? That is the worst case. Delivery orders competing with dine-in tickets during your busiest hour slow your dining room service while paying you less per dish.
That last pattern is common and it is worth checking specifically, because the fix is easy — turn delivery off during your peak two hours — and almost nobody does it.
The costs that get left out
Packaging. Real, per order, and rising.
The labour to assemble and hand off. Somebody stops what they are doing for every collection.
Remakes and refunds. Delivery generates more disputes, and the platform frequently resolves them in the customer's favour out of your money.
Kitchen disruption. Hard to quantify and real. Tickets arriving out of your own service rhythm slow the whole line.
The pricing question
Many restaurants price delivery higher than dine-in to recover commission. That is legitimate and most platforms permit it, though the terms vary and are worth reading rather than assuming.
Two cautions. Customers do notice and some resent it, particularly if the gap is large. And a price increase does not fully recover commission, because commission is charged on the higher price too — raising a $20 dish to $24 at 25% commission gives the platform $6 instead of $5, so you keep $3 of the $4.
Work out the increase that actually gets you where you want to be, rather than adding a round percentage and assuming it is covered.
The part worth being strategic about
Every delivery order gives the platform a customer relationship and gives you a transaction. You generally do not get the customer's details, cannot market to them, and cannot bring them back directly.
So the honest framing is that you are renting access to customers, permanently, at twenty to thirty per cent. That can be a reasonable deal — it is a marketing channel that only charges when it works — but it is a marketing cost rather than a sales channel, and it should be budgeted like one.
Direct ordering is the counterweight. Your own online ordering, or the phone, carries no commission and gives you the customer record. It will never match a platform for reach, and every order you move across is worth roughly a quarter more to you.
Worth doing the arithmetic on what that migration is worth before deciding how hard to push it. If a fifth of your delivery volume moved to direct, what does that recover annually? For most restaurants the number is larger than they expect.
The phone is the cheapest direct channel you already have
Phone orders carry no commission, and they are the channel most restaurants let ring out — because the phone rings hardest during service, when nobody can pick it up.
A phone order lost to a busy line frequently becomes a platform order twenty minutes later from the same customer, at twenty-five per cent less to you. That is the same revenue arriving through the expensive door because the cheap one was unattended.
What a restaurant phone setup actually has to do covers the options, including deciding deliberately whether you want phone takeout at all rather than leaving it to whoever is nearest the handset at seven on a Friday.
The three numbers to pull
Delivery orders per week, by hour. Dine-in covers over the same period, before and after. Average order value on each channel.
Those three tell you whether delivery is incremental, cannibalising, or actively getting in the way — and the answer changes what you should do about it.