The delivery app commission that ate the kitchen's margin

The tell is always the same. Sales are up, the kitchen has never been busier, the tickets do not stop, and the year-end accounts are worse than the year before. Everyone in hospitality has either had that year or watched somebody nearby have it, and delivery platforms are the most common single explanation.

The mistake is not signing up. For plenty of kitchens the apps are the difference between a viable business and an empty Tuesday. The mistake is treating platform revenue as if it were the same shape as counter revenue, and then measuring the wrong thing while the margin drains.

The number that is published, and the number that is not

Uber Eats publishes its UK merchant rates openly: 30% where Uber's couriers deliver, 13% where you deliver yourself, 13% for pick-up orders, and a £650 initial activation fee excluding VAT that covers the tablet, printer, menu build and photography. That is a genuinely useful reference point because it is on the record.

Deliveroo and Just Eat quote per restaurant rather than publishing a single figure, and the rate moves with the delivery model, the plan you are on, your location and whether you have agreed any exclusivity. Which means the only rate that matters to your business is the one written in your own agreement, and a surprising number of owners have never read it since the day the tablet arrived.

Read yours for three things specifically: the commission percentage and whether it is calculated on the order value including or excluding VAT, who funds promotions, and what happens to your money when a customer complains.

Where the money goes on a £25 order

Hot takeaway food is standard-rated for VAT at 20%, so a £25 order is £20.83 of sale and £4.17 that was never yours.

Take a platform-delivered order at 30%, applied to the £25 the customer pays. Commission is £7.50, plus VAT which a VAT-registered kitchen reclaims, so £7.50 of real cost. Food cost at 30% of the net sale is £6.25. Packaging, which counter service does not pay, is around £0.80. That leaves £6.28 of contribution before a single second of labour, rent or utilities.

The same £25 taken at the counter or through your own website: £20.83 net, £6.25 food, £0.80 packaging, and roughly £0.38 of card fees at a typical online rate of about 1.5% plus 20p. Contribution £13.40. More than double.

Now put labour against it. Two people on the delivery line at £13 an hour fully loaded is £26 an hour; at twelve orders an hour that is £2.17 an order. The platform order is down to about £4.11 before rent, utilities, insurance and the cost of the person who answers the phone. The direct order is at £11.23. If your rent and overheads run to £4,000 a month, the platform channel needs roughly 975 orders a month to cover them on its own. The direct channel needs about 356.

One detail worth checking in your agreement: if your commission is calculated on the VAT-exclusive figure rather than the £25, the commission in that example falls to £6.25 and contribution rises by £1.25. Same rate, different base, a fifth of a difference.

Delivery does not have a revenue problem. It has a contribution problem, and revenue is what hides it.

The discounts you fund yourself

This is where the arithmetic turns. Platform promotions — 25% off, buy one get one free, free delivery over a threshold — are typically funded by the restaurant, not the platform. The discount comes off your consideration, so your VAT falls with it, but your food cost does not move at all.

Same £25 order with a 25% discount you have funded. The customer pays £18.75, the net sale is £15.63, the commission at 30% is £5.63, food cost is still £6.25 because the portion is identical, packaging is still £0.80. Contribution: £2.95, before labour. Take the £2.17 of labour off and there is 78p left on a £25 basket.

Promotions are not automatically wrong. Run as an acquisition cost with a measured second-order rate, they can be perfectly rational. Run permanently because switching them off drops you down the listings, they are a subsidy from your kitchen to somebody else's app.

The three tests worth running before you renew

Contribution, not revenue. Work out the pounds left after commission, food, packaging and the labour that touched the order. Do it per order, for each platform separately, and compare it against a counter order. Almost nobody does this and almost everybody is surprised.

Incrementality. Is delivery bringing in custom you would not otherwise have, or is it moving existing customers from a channel that pays you £13.40 to one that pays you £6.28? A quiet dining room and a busy delivery screen can be the same customers, and that is a downgrade dressed as growth.

Capacity cost. A delivery ticket during a Saturday service does not just cost commission, it costs the table it slowed down. If your kitchen is at capacity, the marginal delivery order is displacing your best-paying work, and the honest answer may be to cap delivery hours rather than optimise them.

What actually fixes it

Price the channel separately, if your agreement permits it — most platforms accommodate different delivery pricing, and it is worth confirming yours does before you change anything. A 15% uplift takes the £25 basket to £28.75, and the contribution goes from £6.28 to £8.28. That is not gouging; it is charging for a service that genuinely costs more to deliver, which is the same principle as pricing for profit anywhere else in the business.

Move the close postcodes to self-delivery where the published rate drops from 30% to 13%, and be honest about your own delivery cost before you do — a driver, insurance and fuel is not free, it is simply a cost you control.

Build the direct channel deliberately. Your own ordering page, a QR code on every single bag, a card in the packaging, a small incentive for ordering direct next time. The economics above are the entire argument: the direct order is worth more than double, and the cost of getting it is a website and a sticker. Card processing fees and what they cost your margin sets out what the payment side of that really costs.

Engineer the menu for the channel. Items that travel badly generate refunds, and refunds on platforms usually come out of your money. Items with thin margins that only worked as add-ons at the counter should not be on the app at all.

And watch the cash timing, not just the margin. Platform payouts run on their own cycle, which is why a kitchen can be busy, profitable on paper and short of money in the same week — the distinction cash flow versus profit exists to explain, and one the delivery channel makes sharper than almost anything else in hospitality.

The kitchens that come out of this well are rarely the ones that quit the apps. They are the ones that stopped treating a platform order and a counter order as the same £25, and started running the two as separate businesses that happen to share an oven. What a café owner's margins actually look like is the wider version of the same picture.

Common questions

How much commission do UK delivery apps charge restaurants?

Uber Eats publishes its UK rates: 30% of the order value where Uber delivers, 13% where you deliver yourself, 13% on pick-up orders, plus a £650 activation fee excluding VAT. Deliveroo and Just Eat quote rates per restaurant rather than publishing one figure, and what you pay depends on the delivery model, the plan or visibility tier, your location and any exclusivity you have agreed. Commission is charged plus VAT, which a VAT-registered business reclaims. The only figure that matters for your own planning is the one in your agreement, including whether the percentage applies to the VAT-inclusive or VAT-exclusive order value.

Can I charge more on the delivery app than in the restaurant?

Most platforms accommodate different pricing for delivery menus, and your own agreement is the place that settles it before you change anything. The case for doing it is straightforward: a platform-delivered order carries commission and packaging that a counter order does not, so identical pricing means you are selling the same dish at a materially lower margin through the more expensive channel. On a £25 basket at 30% commission, a 15% uplift on the app menu lifts contribution from roughly £6.28 to £8.28. The risk is customer perception, so most kitchens uplift moderately across the menu rather than dramatically on headline items.

Is delivery app commission worth it for a small kitchen?

It depends entirely on whether the orders are incremental. If the apps bring customers you would not otherwise have, and your kitchen has spare capacity during those hours, a lower-margin order is still better than an empty oven. If they are moving your existing customers from a channel worth around £13 of contribution on a £25 basket to one worth around £6, the platform is being paid out of margin you already had. Work the numbers per order rather than per month, separately for each platform, and pay particular attention to what happens during your busiest service, when a delivery ticket displaces a table.

Who pays the VAT and who funds the discounts on a delivery order?

You account for VAT on what the customer pays for the food, at 20% for hot takeaway, and the platform charges its commission plus VAT, which you reclaim if you are VAT registered. If your agreement makes the platform the principal rather than your agent the treatment differs, and the agreement states which it is. Promotions are the bigger issue: platform discounts are typically funded by the restaurant, so a 25% discount reduces your sale and your commission but leaves your food cost completely unchanged. On a £25 basket that can take contribution from around £6.28 to under £3 before labour.