Prime cost for UK cafés: food and labour combined
You know your food cost. You could probably tell me what the wage bill runs at in a normal week, give or take fifty quid. What almost nobody carries in their head is the two added together, which is a shame, because food and labour combined is the number that decides whether the café can work at all.
That combined figure has a name. Prime cost. Standard in American restaurants, barely used over here.
Here's the catch. Search "prime cost restaurant" and page one is US vendors telling you 60% is healthy. Run that against UK numbers and you'll conclude you're fine when you aren't.
What prime cost is, and why the combined number beats either half
Prime cost = cost of sales + total labour, as a percentage of turnover excluding VAT.
Both halves are broader than they sound, which is what the corrections below are for.
It beats either half because the two trade against each other. Bring in bought-in cakes and your labour drops while your cost of sales climbs. Bake everything in house and the reverse happens. Both legitimate decisions, and neither visible if you only ever look at food cost percentage or only ever look at wages.
Prime cost is the referee. It doesn't care which side of the line the work sits on, only what the two cost together, and on the numbers below that's around seven pounds in every ten of ex-VAT turnover.
The prime cost formula, with three UK corrections
The formula is simple. Getting the inputs right is where UK operators go wrong.
Correction one: divide by turnover net of VAT
Your till shows gross takings. VAT is inside that number and it was never yours. Divide costs by gross takings and prime cost comes out flattering by roughly a sixth.
Take gross takings of £8,000, all standard rated. Net of 20% VAT that's £6,667. Costs of £4,926 read as 62% against £8,000, and 74% against £6,667. Same week, same café, twelve points apart.
Two things to watch:
- Not everything is standard rated. Most cold takeaway food is zero rated, though confectionery, crisps, soft drinks and ice cream stay at 20% even cold and to go. With a real sandwich and cake-to-go trade, take the net figure off your VAT return or your EPOS reporting rather than dividing the lot by 1.2.
- The temporary 5% rate on children's meals ended on 1 September 2026. Those sales are standard rated at 20% again, so a spreadsheet built over the summer is now wrong.
If you aren't VAT registered your takings are already net, so skip this correction.
Correction two: labour means the full cost of employing people
Gross pay is what lands in their bank accounts. It is not what they cost you. All of this belongs in the labour half:
- Gross pay for hours actually worked, not hours rota'd
- Employer National Insurance, 15% on earnings above £5,000 a year per employee
- Employer pension contributions, a minimum of 3% of qualifying earnings under auto-enrolment
- Holiday pay accrued, which is earned in the week it's worked even though it's paid later
- Your own wage
That last line is the one that quietly ruins the number. Work forty hours behind the counter, take drawings when the bank looks healthy, and your prime cost is missing a full-time salary. There's a longer argument about what a café owner should realistically pay themselves; the rule here is that if the job would need paying for when you're not there, it's a cost, drawn or not.
For a mostly hourly-paid team that uplift lands around a quarter, in line with the fully loaded hourly rates in our wage percentage benchmarks.
Correction three: cost of sales means everything you bought to sell
Not just food. The cost of sales half is:
- Food and ingredients
- Drink, including your coffee, milk and syrups
- Packaging, cups, lids, boxes and napkins
- Everything you binned
Waste is the one people leave out, and that's the whole problem. The trays binned on Tuesday were bought and paid for, and they belong in cost of sales exactly as much as the ones you sold. Take the figure from supplier invoices rather than theoretical recipe costs and it's already in there.
A worked week: food and labour combined for a café taking £8,000
A generic all-day café with a decent brunch trade, five on the payroll on £350 each plus the owner front of house. Illustrative numbers. Labour first, because that's the half people get wrong.
| Labour line | Weekly |
|---|---|
| Gross pay for hours worked | £1,750 |
| Employer NI at 15% above £5,000 a year per employee | £190 |
| Employer pension at 3% of qualifying earnings | £35 |
| Holiday pay accrued at 12.07% | £211 |
| Owner's own wage, 40 hours at £15 | £600 |
| Total labour | £2,786 |
Note the shape of that. Gross pay £1,750, true cost £2,786. £436 of that is employment cost sitting on top of gross pay, and another £600 is the wage you are not paying yourself.
Employer NI there is gross. At roughly £9,900 a year the bill sits under the £10,500 Employment Allowance, so if this café qualifies the cash cost is nil. Prime cost carries the gross figure so the cost of an hour stays honest, and the allowance is taken as a separate credit.
| Line | Amount | % of net turnover |
|---|---|---|
| Gross takings, VAT inclusive | £8,000 | |
| Turnover net of VAT | £6,667 | 100% |
| Cost of sales, including waste | £2,140 | 32.1% |
| Labour, fully loaded | £2,786 | 41.8% |
| Prime cost | £4,926 | 73.9% |
73.9%. Now the same week with the two common mistakes in it.
- Divide by gross takings instead of net: 61.6%
- Leave the owner's wage out and use net turnover: 64.9%
- Do both: 54.1%
Fifty-four per cent looks superb. Seventy-four per cent looks tight. Same café, same week, same money. The only thing that changed is whether the arithmetic was honest.
What is a good prime cost percentage for a UK café?
Straight answer: there is no published UK prime cost benchmark. Not from a trade body, not from a bank, not from anyone I could point you at. If a page hands you one with a confident source, be suspicious.
What I can do is derive one and show the working, because we already publish both halves. The cost of sales side blends our gross profit margin benchmarks by category by format, weighted for the drinks-to-food mix each one runs, so a coffee-led shop sits near the drinks end and a brunch room near the hot food end. The labour side is our wage percentage benchmarks unchanged.
| Café type | Cost of sales | Labour | Prime cost, derived |
|---|---|---|---|
| Coffee-led, takeaway dominant | 22-28% | 25-30% | 47-58% |
| Bakery or counter-led | 27-33% | 28-33% | 55-66% |
| Brunch or full-service | 30-36% | 32-38% | 62-74% |
| Multi-site, 3 or more | 28-35% | 28-34% | 56-69% |
The worked example lands at 73.9%, right at the top of the full-service band. Two other things fall out of it.
The format matters more than the management. A coffee-led kiosk and a brunch room can both be brilliantly run and sit fifteen points apart. Find your row before you judge yourself.
A drinks-led café does not split 30/30. The American convention assumes roughly a third food, a third labour. Sell a lot of coffee and your cost of sales drops into the twenties while labour stays put, so two cafés at 58% can have completely different problems.
Why the US 60% benchmark misleads
The "55 to 65% is healthy, over 70% is a warning" line that US restaurant software guides repeat isn't wrong where it comes from. It just doesn't survive the journey. Employer costs, tipping conventions and wage floors don't map across, so the labour half is measuring something else, and those numbers come from food-led restaurants rather than a café where drinks are a large share of what goes over the counter.
Use the derived table and your own trend, not a number imported from a different market.
Prime cost plus occupancy is the real ceiling
Prime cost alone doesn't tell you whether the business works, because the biggest cost you can't reduce is missing from it. Add occupancy cost, which is rent plus rates plus service charge, plus any insurance recharge or turnover top-up over the same ex-VAT turnover and the picture closes.
That post puts everything that is neither prime cost nor occupancy at around 13% of ex-VAT turnover: energy, card fees, repairs, marketing, software, accountancy and depreciation. Work back from a five per cent net margin and you get a working ceiling:
Prime cost + occupancy needs to come in under about 82% of ex-VAT turnover.
Run the example café through it. Prime cost 73.9%, occupancy 10%, other overheads 13%, leaving 3.1% net. Over the line, and nothing like the comfortable business the 54.1% figure implied.
Move that same café into a bigger room at 14% occupancy and there is nothing left at all. Same coffee, same team, same prime cost. A 66% prime cost is excellent on a 6% rent and right on the edge at 16%, where every point you claw back is the only headroom you have. That's why the two numbers only mean anything together.
Tracking it weekly, in ten minutes
Monthly is too slow. By the time the management accounts land you've traded four more weeks on the same mistake. Prime cost is a weekly number or it's decoration. Once a week, same day, same sheet:
- Net turnover. The ex-VAT sales figure from your EPOS. Not the banking, not the gross takings.
- Cost of sales. Total the week's supplier invoices for food, drink and packaging. Deliveries won't match consumption in any single week, so watch the four-week average.
- Labour. Actual hours worked times rates, from the clock-in report or the timesheet, not the published rota, plus about 25% for NI, pension and holiday, plus your own wage as a fixed line. Check that 25% against a real payroll run once a quarter.
- Add and divide. Write both halves down, not just the total. A prime cost that holds steady while cost of sales climbs and labour falls is a café quietly buying convenience, and you want to notice that on purpose.
Ten minutes, and it slots into the weekly bookkeeping routine that produces these numbers anyway.
Over your line? The order to pull the levers
When the number is too high the instinct is to cut hours, because that's the lever you control on Sunday night. It's usually the wrong one to pull first. Work cheapest-to-the-customer first.
1. Find out which half moved. If cost of sales climbed, cutting hours does nothing but make the room worse.
2. Mix. A few points shifted from lower-margin plates to higher-margin drinks moves prime cost without touching a price or a rota.
3. Recost, then check portions and waste. If you last costed the brownie in February, that margin is fiction. Weigh what goes on the plate against the recipe, then look at what you bin at close.
4. Reshape the rota, don't just cut it. Build from the demand curve, not last week's pattern. Cutting evenly takes service off your busiest hours and makes the ratio worse.
5. Price, last and deliberately. The fastest lever and the one customers notice, so spend the other four first.
Where Brikly fits
CostingBrik keeps the cost of sales half current, reading your supplier invoices and updating every recipe using an ingredient when its price moves. That's the half that drifts silently. To sanity check one dish first, the free recipe costing calculator does a single recipe with no signup.
It won't produce your prime cost for you. What it removes is the part where you don't trust the cost of sales figure because you know it's six months old.
Your food cost is a good number. Your wage percentage is a good number. Neither tells you whether the café works, because each can be improved by making the other worse.
Add them, divide by turnover net of VAT, include your own wage, then add your occupancy cost. Do it this Sunday for last week, then every Sunday after. The first one will probably be uncomfortable. The twelfth will be the most useful number in the business.
Ed O'Brien has run Hunters Cake Company for 17 years across cafés in Witney, Burford, and a bakery in Carterton, Oxfordshire. He's building Brikly - modular tools that give independent café owners the same data the big chains have, without the big chain price tag.