The UK Delivery Profitability Playbook
How UK takeaways and multi-site operators can turn delivery from a volume business into a margin business in 2026.
1. The problem, in one page
Delivery is now the largest source of new revenue in UK quick-service and fast-casual restaurants — and the largest source of margin leakage.
Between 2019 and 2025 the share of UK takeaway revenue flowing through third-party aggregators (Uber Eats, Deliveroo and Just Eat) more than doubled. For a growing number of independent and multi-site operators, aggregators now account for more than half of weekly orders. But the economics have not kept pace with the volume.
Three forces are squeezing the delivery P&L at the same time:
- Labour costs have risen sharply. The National Living Wage moved to £12.71 per hour in April 2026. Industry analysis puts the cost to the sector of recent National Insurance and wage changes at around £1.4bn in additional payroll. 88% of hospitality operators cite labour cost as their number-one financial concern.
- Aggregator commissions have not fallen. Headline commissions of 25–35% on marketplace orders, plus per-order card fees, plus consumer-side service fees that depress basket size, have remained largely unchanged. Meanwhile aggregators have added layers: sponsored placement, promotions funded by the restaurant, and stricter ranking rules that effectively tax any operator who does not discount.
- Food and packaging inflation has embedded itself. Even as headline CPI has cooled, the cost base on which most takeaways operate — chicken, cooking oil, cheese, dairy, paper packaging — has not reset to 2019 levels.
The result is a sector where top-line delivery growth is disguising margin compression. Restaurant insolvencies in the year to February 2025 passed 3,400. And in a recent UK Hospitality / Lloyds survey, only 28% of operators reported that technology investments had produced a measurable profitability gain — the rest had either broken even or lost money on their tech spend.
Our view. This playbook is for the operators who want to sit in that 28%. It is a practical, numbers-first guide to turning delivery from a volume business into a margin business.
2. The unit economics of a delivery order
Before you can fix delivery profitability you have to see it. Most operators know their food cost percentage and their weekly labour bill, but very few can tell you what is actually left on a typical delivery order after everything has been paid.
Here is what that looks like, in pounds and pence, for a £27 food basket ordered through a third-party aggregator in the UK in 2026. Every number in this table is within the observed range we see across hundreds of live Andromeda sites. Your numbers will vary — the purpose is to make the shape of the P&L visible.
A quick note on VAT. The figures below assume a VAT-registered business on the standard scheme selling hot takeaway and delivery food at 20% VAT — the common UK setup. VAT is collected from the customer on the gross "customer pays" line and paid to HMRC. Supplier charges (aggregator commission, card processing, Andromeda platform fee, Uber Direct) are shown as their net P&L cost — VAT charged by these suppliers is fully reclaimable on a standard-scheme return, so the P&L impact is simply the net fee. Operators below the £90,000 VAT threshold can ignore the VAT row on the customer side — the shape of the P&L shifts, but the direct-vs-aggregator margin gap is similar, often wider. The "% of net food" column uses the £22.50 net food sale as the reference base in both tables so costs compare like-for-like across channels.
A typical £27 aggregator basket in 2026
| Line | £ | % of net food sale | Notes |
|---|---|---|---|
| Gross food basket (customer pays) | 27.00 | — | Customer also pays a £2.99 delivery fee direct to the aggregator — total customer outlay £29.99. That £2.99 does not flow to the restaurant. |
| Less: VAT on food (20%) | (4.50) | — | Paid to HMRC |
| Net food sale (ex-VAT) | 22.50 | 100.0% | The honest revenue base |
| Less: aggregator commission — 30% of £27 gross ticket | (8.10) | 36.0% | Commission is charged on the gross food basket, not on the net-of-VAT figure — this is the net P&L cost |
| Less: card processing — ~2% of £27 gross | (0.54) | 2.4% | Often bundled inside commission; shown separately here |
| Net revenue from aggregator (ex-VAT) | 13.86 | 61.6% | The cash that actually lands in your bank |
| Less: food cost | (6.30) | 28.0% | Typical QSR food cost; higher on pizza, lower on fried chicken and Indian |
| Less: packaging | (0.68) | 3.0% | Boxes, bags, sauce pots, cutlery, labels |
| Less: kitchen and pack labour | (3.38) | 15.0% | Pro-rated share of kitchen labour for delivery orders |
| Less: consumables, energy, waste | (0.45) | 2.0% | Small but real |
| Variable contribution | 3.05 | 13.6% | Before any fixed cost |
| Less: allocated fixed costs (rent, rates, insurance, admin) | (3.38) | 15.0% | Per-order allocation on a typical 1,000-order-per-week site |
| Net profit per £27 aggregator order | (0.33) | −1.5% | A small loss per order, once commission is applied correctly to the gross ticket |
A negative number per order is not a typo. It is the shape of a marketplace-heavy delivery business at 2026 cost levels once commission is applied to the gross ticket — and for many operators the true picture is worse, once promotions, sponsored placement, and discount-funded ranking moves are factored in.
Three observations are worth highlighting.
First, aggregator commission is almost always the largest single cost line. On a £27 marketplace basket it is larger than food cost. It is larger than labour. In most P&Ls it is the single biggest lever you have, and most operators are under-optimising it.
Second, the remaining contribution is very sensitive to basket size. If the same customer orders £19 instead of £27, you do not lose £8 of contribution — you lose almost all of it, because fixed costs per order do not scale down. A 30% drop in basket size can turn a marginal order into a clear loss-maker. (We come back to this — and to a simple upsell model — in §2.1.)
Third, most of the levers that move this P&L meaningfully are levers the operator controls themselves. They are not things aggregators or delivery partners will fix for you. The rest of this playbook is about those levers.
The same £27 basket on your own website (direct)
The equivalent numbers through a direct digital ordering channel — your own branded website or app, with payments settled on a single transparent platform fee and delivery fulfilled by a DaaS partner — look very different. Two key structural differences: the £2.99 delivery fee is kept by the restaurant, and the Andromeda platform fee is a single 6% line that includes card processing (there is no separate Stripe line to deduct on top).
| Line | £ | % of net food sale | Notes |
|---|---|---|---|
| Gross food basket (customer pays) | 27.00 | — | Priced at menu — no aggregator uplift needed |
| Gross delivery fee (kept by restaurant) | 2.99 | — | Customer pays the restaurant direct; no middleman |
| Total gross customer spend | 29.99 | — | Identical to the aggregator scenario |
| Less: VAT on total (20%) | (5.00) | — | Paid to HMRC |
| Net sales (ex-VAT): food £22.50 + delivery £2.49 | 24.99 | 111.1% | Delivery adds 11.1 points to the revenue base |
| Less: Andromeda platform fee — 6% of £29.99 | (1.80) | 8.0% | Single transparent charge that includes card processing — no separate Stripe line |
| Less: Uber Direct courier | (3.75) | 16.7% | Per-drop rates, typically £3.50–£5.50 depending on distance and time of day |
| Net revenue after platform and courier (ex-VAT) | 19.44 | 86.4% | Cash actually left to cover food, labour and fixed costs |
| Less: food cost | (6.30) | 28.0% | Same kitchen, same recipe |
| Less: packaging | (0.68) | 3.0% | |
| Less: kitchen and pack labour | (3.38) | 15.0% | |
| Less: consumables, energy, waste | (0.45) | 2.0% | |
| Variable contribution | 8.63 | 38.4% | |
| Less: allocated fixed costs | (3.38) | 15.0% | Same fixed base |
| Net profit per £27 direct order | 5.25 | 23.3% | The direct channel earns £5.25; the aggregator channel loses 33p — a per-order swing of £5.58 |
The direct channel is not magic — you still pay a courier and you still pay a transparent platform fee that covers card processing. But the combined cost of platform plus courier, net of the £2.99 delivery fee the customer pays direct to the restaurant, is around £3.06 per order. That is materially less than aggregator commission plus card processing alone, which on the same basket comes to £8.64. And critically, the customer relationship, the data, and the repeat ordering flywheel all belong to you.
If a site currently does £10,000 per week on aggregators (≈ 370 orders at £27 each), the per-order profit delta between an aggregator order and a direct order is £5.58. Shifting just 20% of that volume — 74 orders per week — to the direct channel is worth an extra £413 per week, or over £21,000 a year, per site, without a single additional customer.
That is the headline of this playbook: the single largest profitability lever in UK delivery today is channel mix. But it is not the only one. The next two sub-sections show what happens when you add an upsell and a second order a year on top.
Want the numbers to hand?
Download the full playbook as a print-ready PDF — every table, every checklist, every worked example.
2.1 Model 1 — what if you drove £1.50 more through upsells?
An upsell is any order that grows by a small increment after the customer has already decided to buy — a garlic bread on a pizza, a dip on a burger, a premium base or a larger size. On a digital channel the operator controls the prompt, so the economics are unusually clean: the food cost of the upsell moves, but fixed costs, labour, packaging and courier fees do not. Almost the entire incremental contribution drops to the bottom line.
| Per-order impact of adding £1.50 AOV via upsell | Aggregator £ | Direct £ | Notes |
|---|---|---|---|
| Incremental gross revenue | +1.50 | +1.50 | A single modifier, side, or premium upgrade |
| Less: VAT (20%) | (0.25) | (0.25) | |
| Less: aggregator commission 30% / Andromeda 6% | (0.45) | (0.09) | Commission scales with basket; Andromeda fee barely moves |
| Less: card processing 2% (agg only) | (0.03) | — | Included in Andromeda's 6% |
| Less: incremental food cost (28% of net upsell) | (0.35) | (0.35) | Only the food moves; all other costs fixed at order level |
| Net profit uplift per upsold order | 0.42 | 0.81 | Direct channel captures almost 2× the uplift |
At site level. A 1,000-order-per-week site with a 40% upsell take rate — realistic for a well-designed online menu — generates 400 upsold orders per week. At £0.81 direct, that is £324 per week — or just under £17,000 per year per site — from a single cross-sell prompt. At £0.42 aggregator, the same take rate delivers about £8,700 per year. The direct channel almost doubles the return on the same merchandising effort.
This is the often-missed second lever. Operators chase volume growth and channel mix, but the cheapest marginal pound on the delivery P&L is the upsell pound — because it travels on fixed infrastructure you have already paid for.
2.2 Model 2 — what if loyalty drove one extra order a year?
The third lever is frequency. Every additional order from an existing customer skips acquisition cost entirely and earns full per-order margin. Andromeda's own data across live UK operators shows a very stark shape:
| Customer type | Orders / year | Direct profit / year | vs baseline |
|---|---|---|---|
| Web-only customer (baseline) | 3.4 | £17.85 | 1.0× |
| Web-only with +1 extra order (loyalty nudge) | 4.4 | £23.10 | 1.3× |
| App-only customer (installed, logged in) | 4.8 | £25.20 | 1.4× web-only |
| Dual-channel customer (web + app) | 12.7 | £66.68 | 3.7× web-only |
Source: aggregated anonymised Andromeda customer data, 2026. "Direct profit" uses £5.25/order from §2.
At site level. Imagine a site with 500 active direct-ordering customers. If a simple loyalty programme drives just one extra order per customer per year, that is 500 × £5.25 = £2,625 of additional profit per year from loyalty alone — more than covering the cost of a well-run scheme. If that same site converts just 100 of those customers into app users, the lift on those 100 alone (from 3.4 to 4.8 orders/year × £5.25) is another £735/year. And if 50 of them become dual-channel (web + app, which is the shape loyalty mechanics tend to produce), that is a lift of 50 × (12.7 − 3.4) × £5.25 = over £2,441/year.
Stacked, the three levers in this section compound into the same number from three different angles: channel mix (£21k+), upsells (£17k+), and loyalty / app activation (£5–10k+), all per site per year. None of them requires a single additional customer through the door. They require a digital stack that actually captures the behaviour — and that is the bridge into §3.
3. The five levers that move the needle
Every delivery profitability workshop we have run with an operator comes back to the same five levers. You do not need to pull all five at once. You do need to know which one is most broken at your site and fix that one first.
01Channel mix (aggregator share vs direct share)
What it is. The split of your delivery volume between marketplace aggregators and direct digital channels (your own website, app, QR codes, repeat ordering).
Why it matters. As the unit economics above show, a direct order earns over £5 of net profit, while the same basket on a marketplace platform is marginally loss-making once commission is applied to the gross ticket — a per-order swing of roughly £5.58. If aggregators are 80% of your delivery mix, you are not just leaving the biggest chunk of margin on the table; on some of those orders you are actively subsidising the marketplace.
How to move it.
- Make sure every aggregator bag that leaves your kitchen has a voucher or flyer for your direct site, with a first-order incentive that is cheaper than the commission you would have paid on a repeat aggregator order.
- Run paid social tightly geo-fenced around your postcode, pointing to your own site, not the aggregator app.
- Use your POS to capture phone numbers on walk-in and collection orders, then send an SMS campaign offering a £3 discount to the first 500 customers who order direct.
- Put a QR code on every pizza box, every chicken bag, every sushi tray — landing on your direct ordering site.
- Do not try to remove yourself from aggregators. They are a discovery channel. But stop treating them as your primary order-taking channel.
Realistic impact. For independents we typically see direct share move from 10–15% to 30–40% inside twelve months when the levers above are applied consistently. For small multi-site operators the number can reach 50–60%.
02Basket engineering
What it is. The set of tactical moves that increase the average order value: upsells at checkout, modifier defaults, deal structure, minimum order thresholds, menu layout.
Why it matters. Fixed costs per order do not scale with basket size. A 10% increase in average order value drops straight through to margin — there is almost no corresponding cost increase.
How to move it.
- Cross-sell at the right moment. A "forgot your drink?" modal shown after the customer has added their mains, but before they pay, converts at 15–25%. Shown at the wrong point it annoys the customer and converts at 2%.
- Make modifiers default-on where they make sense. If 80% of customers order chips with a burger, default chips into the burger builder and let people opt out. Default behaviour wins.
- Raise your minimum order value by £2. On most UK takeaway sites the minimum is set at the level the owner would accept for a last customer on a Tuesday night, not the level that would be profitable. Test a £1–2 rise; drop-off is usually under 3%.
- Build two-for-£X deals that lift the basket without heavy discounting. Two pizzas at £19.50 is materially more profitable than one pizza at £12.50.
- Engineer the menu visually. Put the high-margin items top-left on the digital menu. Add photos to the high-margin items. Bury low-margin side dishes.
Realistic impact. A well-run basket engineering programme moves average order value by 8–15% in 60 days.
03Commission elasticity
What it is. The amount of commission you are actually paying on aggregator orders, versus the amount you have to pay.
Why it matters. Most operators accept the commission tier they were signed up on three years ago. Aggregators have discount tiers, self-delivery tiers, and special programmes that can shift your effective commission rate by 3–6 percentage points — worth thousands per site per year.
How to move it.
- Move to self-delivery tiers on Uber Eats and Deliveroo where you can fulfil yourself. The commission step-down is typically 8–10 percentage points. You carry the fulfilment cost, but on short-distance deliveries you are usually better off.
- Use your POS and ordering system to raise aggregator-channel menu prices. A 10–15% uplift is accepted industry practice and fully funded by the consumer. Every pound of uplift flows through to your margin.
- Renegotiate annually. Aggregator account managers are measured on restaurant retention. Ask for a commission review every twelve months. The worst they can do is say no.
- Turn off under-performing aggregators rather than keeping them on "in case." Each platform you are on adds menu maintenance overhead, order interception risk, and ranking volatility. If Uber Eats is 5% of your volume and Deliveroo is 40%, the argument for paying to stay on Uber Eats is weak.
Realistic impact. Effective commission rate reductions of 3–6 percentage points are common. On a site doing £5,000 per week in aggregator orders, that is £7,800–£15,600 per year.
04Labour productivity (minutes per order)
What it is. The amount of labour time, in minutes, that it takes to produce and despatch a typical order.
Why it matters. Labour is now 28–35% of revenue for most UK takeaways. A 10% productivity improvement is worth more than most operators' full delivery margin.
How to move it.
- Take ticket handling off paper. A kitchen display system (KDS) that receives orders directly from all channels — POS, website, aggregators, QR — removes the most common waste in a takeaway kitchen: re-keying, losing tickets, double-making.
- Use the POS to route orders to the right station automatically. A pizza station that only sees pizza tickets is faster than one that has to filter wings, salads, and drinks out of every printout.
- Measure rack time and make time. The time an order sits on the pass waiting to be packed is labour you have paid for and not captured. Reduce the median by two minutes and you can often cut one hour of labour per service.
- Schedule to demand, not to tradition. If your POS and ordering system show that you take 45% of your week's orders in the Friday and Saturday evening peak, staff accordingly.
Realistic impact. 10–20% labour productivity gains are routine when a modern POS and KDS replace paper-based operations.
05Pricing discipline
What it is. The prices on your menu, by channel, reviewed and adjusted deliberately rather than set once and forgotten.
Why it matters. Most UK takeaway menus are priced by copying what the shop down the road charges. Very few are priced for target margin.
How to move it.
- Price by channel. Your in-store menu, your own website, and your aggregator listings should not be on the same prices. The customer knows, and the aggregator already charges them more once service fees are added. Use your POS and ordering platform to hold different price lists cleanly.
- Review prices every six months. Commodity costs do not stand still. Neither should your menu.
- Apply selective uplift, not across-the-board rises. A 10% rise on the three highest-volume items lifts revenue noticeably. A 2% rise on everything does not.
- Stop discounting the same items every week. Rotating which items are discounted prevents customer habit-forming on your lowest-margin SKUs.
Realistic impact. Disciplined pricing reviews typically add 1–2 percentage points to net margin within a single quarter.
4. The profitability checklist
Use this checklist at your next management meeting. Each question is designed so that an operator can answer honestly in under thirty seconds. Count your "no" answers. More than six "no"s and you have a six-figure opportunity hiding in the business.
Channel mix
- I know what percentage of my delivery orders come from each aggregator vs my own direct channels.
- Every aggregator bag leaves my kitchen with a promotion for my direct site.
- I have run at least one SMS or email campaign to existing customers in the last 90 days.
- My direct site is mobile-ready and loads in under three seconds on 4G.
Basket engineering
- I know my average order value by channel.
- My minimum order value has been reviewed in the last six months.
- I have upsell prompts configured in my ordering flow.
- I can list my three highest-margin items without looking them up.
Commission elasticity
- I know my effective commission rate by aggregator.
- I have asked for a commission review in the last twelve months.
- My aggregator menu prices are uplifted versus my in-store prices.
- I have evaluated moving to self-delivery on at least one aggregator.
Labour productivity
- My kitchen operates from a digital display, not paper tickets.
- I know my median make-time and rack-time.
- I schedule staff based on actual ordered demand, not tradition.
- I have reviewed wage cost percentage in the last month.
Pricing discipline
- My menu prices have been reviewed in the last six months.
- I price deliberately differently across in-store, direct, and aggregator channels.
- I know my food cost percentage for the last completed trading week.
- My three highest-volume items have been repriced in the last twelve months.
Tally. 16–20 ticks: you are already in the top quartile of UK operators. Keep the discipline. 10–15 ticks: your site is profitable but leaving meaningful money on the table. 0–9 ticks: do not wait. This is where the top-line growth of the last two years has been disguising bottom-line drift.
5. Worked examples from real Andromeda operators
These three cases are anonymised from real operators on the Andromeda platform. Numbers are rounded; behaviours and sequencing are exactly as they happened.
Case A — An independent pizzeria in the North West
The starting point. A single-site operator, trading for six years, doing roughly £11,000 per week in total orders. 75% of orders through Uber Eats and Deliveroo. Net margin in the last full accounting year was reported at 4%.
What changed. Over eight months the operator rolled out a direct ordering website through Andromeda, moved kitchen operations onto a kitchen display, and started a basic SMS retention programme.
Levers pulled. Channel mix (primary), basket engineering (secondary), labour productivity (secondary).
Measured outcomes. Direct share moved from 12% to 38% of delivery orders. Average order value on the direct channel settled at £23.40 versus £19.90 on aggregators — a 17% uplift driven by better upsell logic and a higher minimum order value. Labour cost percentage fell by 2.1 points as the KDS replaced the printer station. Net margin for the most recent full quarter was 9%.
Case B — A five-site Indian takeaway operator in Greater London
The starting point. A growing small chain with five sites. 60% of revenue through aggregators. Centralised supplier buying, but no visibility of per-site profitability or channel mix.
What changed. The operator moved all five sites onto a single instance of the Andromeda POS and ordering platform. For the first time, site-level and channel-level P&Ls were visible every morning, not forty days after month-end. The operations team ran a 90-day programme focused on commission elasticity (renegotiating terms on two aggregators and moving to self-delivery on a third) and pricing discipline (a 12% menu uplift on the two aggregator channels, implemented on one day across all five sites).
Levers pulled. Commission elasticity (primary), pricing discipline (primary), channel mix (secondary).
Measured outcomes. Effective aggregator commission rate reduced by 4.2 percentage points across the estate. Aggregator revenue rose 9% despite the menu uplift, because the better prices exceeded the small volume dip. Total gross profit rose 11% in the following quarter.
Case C — A fried chicken chain growing from six to twelve sites
The starting point. Six-site operator with aggressive growth plans. Tight central operations team of three. Worried that adding six more sites on existing tooling would break the operations model.
What changed. The operator used Andromeda's multi-site features to standardise menus, pricing, and promotions across all twelve sites from a single place, including channel-specific price lists and site-specific local deals. Kitchen display, delivery tracking and Uber Direct integration rolled out at the same time.
Levers pulled. Labour productivity (primary), basket engineering (primary), channel mix (secondary).
Measured outcomes. Added six sites without adding central head-count. Labour cost percentage remained flat as sites scaled. Average basket rose by 9% after consistent upsell logic was deployed across all twelve sites. The head of operations reported that the time she spent on "firefighting menu and pricing issues" fell from four days a week to half a day.
The point of these cases is not that every operator will see identical numbers. It is that the five levers are repeatable. The ones who pull them, win.
6. How Andromeda helps
Every operator in the three cases above is running Andromeda. That is not why we wrote this playbook — the five levers work with or without us. But Andromeda is a serious accelerant on each of them, and it is worth being clear about where and how.
- A single platform for every ordering channel. The only UK-built platform that runs POS, your own direct website and app, kiosk, QR order-and-pay, kitchen display, and aggregator integration from one system. Orders from Uber Eats, Deliveroo, Just Eat, your website, your kiosk and your counter all land in the same kitchen in the same format.
- Channel-specific pricing, controlled centrally. Your in-store, direct-web, and aggregator price lists are managed from a single Portal. Uplifts are applied in seconds across one site or an entire estate.
- Built-in direct ordering with low transparent fees. Your own branded website and mobile app, included in the platform, with a small card processing fee and a small platform fee per order. Every customer order data point stays with you.
- Real operational visibility. Site-level and channel-level reporting, delivered daily, not forty days after close. Labour percentage, food cost percentage, average order value, commission leakage — visible the morning after trade.
- Designed for growth. Multi-site operators can scale from one site to a hundred on the same platform, with central menu control, franchise-safe pricing, and standardised reporting. The same platform runs a single independent and a 400-site international brand.
7. About the authors
Ben Portsmouth, Founder & CEO, Andromeda
Ben founded Andromeda in 1998 and has spent more than twenty-five years building restaurant technology for UK operators. He has led the platform's evolution from Windows-based POS in the late 1990s through to today's cloud-native, all-channel platform. He sits on the operational and commercial review of every major account.
The Andromeda team
Andromeda's work is produced by a cross-functional team of product managers, engineers, helpdesk leaders, implementation specialists and customer success managers spanning the UK, Bulgaria, the Philippines and Vietnam. Our customers include independents, multi-site takeaways, and international brands. The analyses in this playbook are built from live operational data across hundreds of sites on the platform.
8. Book a 20-minute profitability review
If you have read this far you are probably already running through the checklist on your own business in your head. The fastest way to turn that into action is to talk to one of our team for twenty minutes.
What you get. A review of your current channel mix, a model of your likely profitability ceiling with the five levers applied, and a short, honest view of whether Andromeda is the right platform for your business.
What we need from you. Your current aggregator mix, your approximate weekly revenue, and your number of sites.
No obligation. No sales script. No thirty-slide deck.
Book your profitability review → Download the PDF ↓This playbook is published by Andromeda UK Services Ltd. All numbers, case studies, and percentages are based on live Andromeda platform data and public UK hospitality sector research as of April 2026. Specific customer references are anonymised. The playbook is intended as general commercial guidance and does not constitute financial, legal, or tax advice.