Most restaurant owners know marketplace commission is expensive. Far fewer have sat down and worked out what it costs them specifically — and the number is usually worse than the guess.

Start with one order, not the month

Take a typical delivery order. Say it comes to 1,600 in your local currency — the ratios are what matter, not the units.

A common structure looks something like this:

Line	Amount
Order value	1,600
Food cost (at 32%)	512
Packaging	60
Labour attributable to the order	190
Rent, utilities, overhead share	210
Contribution before commission	628
Commission at 25%	400
What you actually keep	228

Two thirds of the money that order was going to make you left in a single line item.

The comparison that matters is not commission against revenue. It is commission against the margin the order was going to produce.

Against revenue, 25% sounds like a cost of doing business. Against the 628 that order was going to contribute, it is a controlling stake.

Run the numbers on your own restaurant

You need four figures, and you probably already know three of them:

Your monthly marketplace order volume
Your average order value on those orders
Your food cost percentage
Your commission rate

Multiply volume by average order value by commission rate. That is your annual commission run rate divided by twelve. For a restaurant doing 900 marketplace orders a month at an average order value of 1,400 with 25% commission, that is 315,000 a month — roughly 3.8 million a year.

Now compare that against what a subscription costs. The gap is not close.

The part that compounds

The commission is the visible cost. The structural one is that on a marketplace, the customer is not yours.

You do not get their phone number. You cannot tell who ordered four times last month and has now gone quiet. You cannot run a win-back campaign, because you have nobody to send it to.

That means every order is acquired fresh, forever. You never build the asset that makes a restaurant defensible — a list of people who already like your food.

What actually works

Almost nobody should leave marketplaces overnight, and anyone telling you otherwise is selling something. Marketplaces are genuinely good at putting you in front of people who have never heard of you.

The workable path looks like this:

Keep the marketplaces for discovery. Paying to acquire a genuinely new customer is a reasonable trade.
Launch your own ordering channel on your own domain, so there is somewhere to send people.
Put it where customers already look — packaging inserts, receipt footers, a card in the bag, your Instagram bio.
Give repeat customers a reason to switch. A loyalty program that only works on your channel is the most reliable one.
Make delivery at least as good. Live tracking and accurate estimates are what stop people drifting back.

Then let the mix move. Restaurants that do this deliberately usually see direct orders go from nothing to a meaningful share of delivery volume within a few months — and every order that moves across improves margin without increasing what the software costs.

The honest caveat

Direct ordering is not free. You have to drive the demand yourself, which means the discovery marketplaces were doing becomes your job.

For a brand-new outlet with no customer base, that is a real disadvantage and marketplaces earn their fee. For an established restaurant with people who already order regularly, continuing to pay discovery rates on those customers is simply a leak.

The question worth asking is not "should I leave" — it is "how much of my volume is repeat business I am still paying finder's fees on?"