Just Eat, Uber Eats or Deliveroo: which is worth it for your restaurant?

Most of this decision gets argued the wrong way round. Operators compare the marketplaces against each other first and work out the economics afterwards, when the useful order is the reverse: find out whether an order through a marketplace makes you money at all, then decide which ones to be on.

Rates are not published in any stable way, they differ by contract, market and the size of your business, and they get renegotiated. So this article will not quote you a rate for anybody. Take yours off your own statement, and if you have not read one closely, that is the first job. Look for the commission percentage, whatever fixed admin or service charge sits alongside it, any joining or hardware fee, and what the customer is being charged on top, because that last one shapes your basket size whether you like it or not.

Work out whether an order pays

The number that matters is marginal: what one extra meal costs you to make. Not rent, not the fit-out, not your licence, because those are being paid whether that order arrives or not.

Take one representative dish and work down. The commission below is an illustration at 30% rather than a rate anybody is quoting, so substitute yours.

1. Menu price on the marketplace, say ten pounds.
2. Less commission at 30%, leaving seven pounds.
3. Less ingredients, say four pounds.
4. Less the kitchen labour that dish consumes, say two pounds.
5. Less packaging, which delivery adds and the counter does not.

On those illustrative numbers you are left with under a pound, against roughly four pounds on the same dish sold at your counter. Run it on your three best-selling dishes rather than an average, because the average hides the ones losing money.

Two things people leave out and should not. Packaging is a real per-order cost and it is not small. And refunds and cancellations land on you, so take the percentage off your own statement rather than assuming it is negligible.

The marketing argument, and when it holds

The usual defence of an unprofitable marketplace order is that it is marketing: you lose a little now to gain a customer who comes back. It is worth being honest about the mechanics of that, because the marketplace is built to prevent it.

Their business depends on owning the customer relationship. The customer opens their app, not your site, and next week the app decides who they see. You are paying to acquire someone who is not being handed to you. That does not make the channel worthless, but it does mean the acquisition only pays if you have some way of getting a second order without going through the marketplace again.

If you have no direct channel at all, the marketing argument does not hold, because there is nowhere for the acquired customer to be acquired to.

Should you leave?

Usually not, and certainly not as a first move. The marketplaces put you in front of demand you cannot reach otherwise, and an empty kitchen at two on a Tuesday costs you something too. Walking away from volume before you have built anywhere for it to go is how operators end up with neither.

The realistic version is to fix the economics rather than the principle. Price the delivery menu separately from the counter menu, since the two are not selling the same thing. Take off the dishes that travel badly or price out worst. Set a minimum order that reflects your packaging and prep cost. None of that requires leaving anybody.

Should you join all of them?

Being on more marketplaces adds demand, and it also adds a tablet, a second menu to maintain, another set of prep times, and another queue for the same kitchen. The failure mode is not commercial, it is operational: menus drift out of sync, an item sells out on one channel and keeps selling on another, and the kitchen loses the thread on a Friday.

So the honest test is whether you can run another channel without the kitchen noticing. If every marketplace order lands in the same queue as everything else, the answer is often yes. If it means a fourth screen by the window, be more careful. Deliverect is the usual route for collapsing marketplace orders into one queue.

Where a direct channel fits

Not as a replacement, as the place repeat business goes. The customer who found you on a marketplace and liked the food is the one worth moving, and the move is won on the small things: an insert in the bag, a better price direct, collection that is genuinely faster, a menu you actually control.

The economics differ because there is no commission on an order that came from your own site, only card fees and whatever your platform charges. Ours are on the pricing page rather than in this article, for the same reason we are not quoting anybody else's.

Delivery on your own orders is a separate question from taking them. Orkestro and Shipday connect courier networks to direct orders, so you can run delivery without a marketplace in the middle, and delivery zones keep the radius honest about what you can serve hot. If you would rather not do delivery on your own orders at all, click and collect is the version with no last mile in it.

After you decide

Whatever mix you land on, two habits keep it working. Review the numbers per channel every quarter, on your actual statements rather than on what you were told when you signed, because rates and fees move. And keep the menu, prices and availability coming from one place, because the moment they are maintained separately per channel they start disagreeing, and the disagreement always shows up as a refund.

Our side of it is direct ordering for takeaway and delivery, and the takeaway system is where the whole thing sits together.