Skip to content
PayoutLedger

Which Delivery Platform Actually Costs You Least? A Quarterly Review

Most restaurants choose their delivery platform mix once — usually at the point of signing up, usually on the basis of which one seemed biggest — and then never revisit it. Rates change, order mixes shift, refund rates drift, and a decision made two years ago quietly stops being the right one.

This is a guide to reviewing that decision properly, on a quarterly cadence, using numbers you already have.

Why headline commission is the wrong comparison

The instinct is to line the platforms up by their commission rate and take the lowest. Three things make that misleading.

Only two of the three publish rates. Uber Eats sets out its UK merchant pricing openly — 30% for orders its couriers deliver, 13% for self-delivery, 13% for pick-up, all stated as excluding VAT, plus a £650 activation fee. Just Eat publishes two headline rates on its partner signup page — 14% where you deliver, 30% where Just Eat delivers — with VAT added on top in its own worked examples. Deliveroo's merchant site publishes no rates at all. So a three-way table of published rates cannot be built from primary sources: two columns are first-party, the third is somebody's report.

And even for the two that publish, the figure is the standard signup rate rather than necessarily your contracted one — which is why the comparison below is built from your own settled payouts, not from rate cards.

Commission is not the only deduction. VAT on the commission, refunds, adjustments and promotional costs all reduce the deposit. A platform with a lower commission rate and a higher refund rate can easily be the more expensive one. Volume is not neutral. A platform charging more per order but bringing three times the orders may still contribute more cash overall. Cost per order and total contribution are different questions and the answer often differs between them.

The two numbers that actually matter

Effective rate

The share of gross order value you never receive:

``` effective rate = (gross order value − net bank deposit) ÷ gross order value ```

This captures everything — commission, VAT on commission, refunds, adjustments — in one figure. It is the only number that lets you compare platforms on equal terms, because it does not depend on how each platform chooses to present its deductions.

Take it from settled periods over a full quarter rather than one week. A single week is distorted by whatever happened to refund that week.

Contribution per order

Effective rate tells you what a platform costs. It does not tell you what it is worth.

``` contribution per order = (average order value × (1 − effective rate)) − food and packaging cost per order ```

This is what each order actually adds after the platform has taken its share and you have paid for the food. A platform with a punishing effective rate and a high average order value can out-contribute a cheaper platform with small baskets.

Our free payout reconciliation checker will check whether a given period's deductions explain the deposit to within 0.5% of gross sales (its tolerance), and the commission calculator compares the headline rate with an estimated effective cost across all three platforms at once.

A worked comparison

A quarter's figures for a restaurant running all three platforms:

Platform APlatform BPlatform C
Gross order value£18,400£11,200£6,900
Net deposits received£11,960£7,616£4,830
Effective rate35.0%32.0%30.0%
Orders640380290
Average order value£28.75£29.47£23.79
Revenue retained per order£18.69£20.04£16.66
Food + packaging per order£9.50£9.80£8.20
Contribution per order£9.19£10.24£8.46
Total contribution£5,880£3,892£2,452
Read the table three ways and you get three different winners. Cheapest by effective rate: Platform C, at 30%. On rate alone it looks like the one to favour. Best per order: Platform B, at £10.24. Its slightly higher average order value more than offsets its higher rate. Biggest total contributor: Platform A, at £5,880 — comfortably the largest single contributor despite having the worst effective rate on the table, though not quite the other two combined (£6,344).

None of these readings is wrong. They answer different questions. If you are deciding where to push marketing spend, contribution per order matters. If you are deciding whether a platform earns its place at all, total contribution matters. If you are negotiating, the effective rate is the number to take into the conversation.

What you should not do is drop Platform A because 35% looks bad.

Running the review

Quarterly is the right cadence. Monthly is noise; annually is too slow to catch a drift. A quarter gives you twelve or thirteen payout periods per platform — enough to average out an unusual week.

The routine, once your reconciliation is up to date:

  1. Total gross order value and net deposits per platform for the quarter.
  2. Calculate the effective rate for each.
  3. Compare against last quarter's figure. Movement is the signal — a platform whose effective rate has climbed two points has changed something, or your refund rate has.
  4. Pull average order value and order count per platform.
  5. Work out contribution per order and total contribution.
  6. Decide what, if anything, changes: marketing emphasis, self-delivery on the platforms that offer a lower rate for it, menu pricing by platform, or a rate conversation with an account manager.
Step 3 is the one that earns its keep. Absolute numbers are interesting; the trend is actionable.

What to do with the answer

If one platform's effective rate has drifted upward, find out why before acting. Check refunds first — see our guide to delivery platform refund disputes — then check whether a promotion is still running that you meant to end. If self-delivery is available at a materially lower rate, work out your genuine cost per drop before switching. Uber Eats' published gap between 30% and 13% is large enough to be worth the arithmetic, but only if your drivers have real capacity. If a platform is not contributing meaningfully, the question is whether its orders would migrate to your other channels or simply disappear. Usually some of both — and the answer determines whether dropping it recovers cost or just loses revenue. If your effective rates look reasonable but cash is still tight, the problem may be timing rather than cost. Our guide to delivery platform payout schedules covers the working-capital side.

This only works if the reconciliation is done

Every number above comes from reconciled payouts. If your books record deposits rather than gross sales, you cannot calculate an effective rate at all — the gross figure simply is not there.

That is the prerequisite, and it is where most of the work is. Our guides to reconciling Deliveroo, Uber Eats and Just Eat payouts cover each platform's format, and the multi-platform guide covers running all three together.

Key takeaways

  • Do not compare platforms on headline commission. Uber Eats and Just Eat both publish UK rates; Deliveroo does not, so any Deliveroo figure you find is second-hand. And a published rate is still not your rate.
  • Effective rate — `(gross − net) ÷ gross` — is the only like-for-like cost comparison.
  • Contribution per order answers the different and more important question of what a platform is worth.
  • Cheapest rate, best per order and biggest total contributor are frequently three different platforms.
  • Review quarterly and watch the movement, not just the level.

Every figure in this review comes out of reconciled payouts — which, done by hand across three platforms and thirteen weeks, is where the quarter goes. That is the gap PayoutLedger is being built to close: upload each platform's report, get VAT-correct journals, and reconcile every payout to the bank deposit. Join the waitlist for early access.

Sources

Stop reconciling manually

PayoutLedger is being built to reconcile your Deliveroo, Uber Eats, and Just Eat payouts automatically — VAT-correct and pushed into Xero. Join the waitlist for early access.

We'll email you once when PayoutLedger launches. Ask to be removed any time. Privacy policy