Every restaurateur knows their food cost to the decimal. Almost none of them know whether that number is any good.
Ask an operator what their food cost is and you get an exact answer: 31.4 per cent. Ask whether that is good and the room goes quiet. What usually follows is "it was 32 last year". That is not an answer — that is the same number, one year older.
A percentage without a band is trivia. 31% food cost is excellent for a fine-dining kitchen, middling for a bistro and alarming for a pizzeria. The same 38% labour cost is normal in Copenhagen and fatal in Sofia. Without the range for your segment and your country, all you know is that the number exists, not what it means.
And the figures you find online rarely apply to you. Most benchmarks in circulation are American: different VAT, different employer charges, a different tipping culture, a different relationship between rent and turnover. Putting a US labour norm of 28% next to a European payroll is not ambitious — it is simply wrong.
This guide puts nine numbers next to the band they belong in across Europe, per segment and adjusted for the wage and price level of your country. At the bottom you enter your own figures and see, line by line, how much money per year sits between you and the band.
Why "better than last year" tells you nothing
Comparing yourself to last year feels like progress and rarely is. A business that gains one point of food cost a year while sitting six points above the market is not catching up — it is losing more slowly. Only once you know where the band lies do you know whether you are closing a gap or drifting away from it.
There is a second reason. An internal number moves for all sorts of reasons: an expensive winter, a new supplier, a summer with a terrace. The band does not move. That makes your distance from the band a far more stable measure than the change against your own past — and it is the only figure that helps when you have to decide which problem to fix first.
One caveat that matters: a band is a band, not a law. Outside it does not mean your business is broken, and inside it does not mean you are doing well. It means there is a difference here that needs explaining. Sometimes the explanation is excellent — a chef who makes everything from scratch buys more expensive raw product and runs a higher food cost than the neighbour working from prepped components. That is a choice, not a mistake. The numbers below tell you where to look; you decide whether what you find is a problem.
The ultimate guide The ultimate guide to restaurant finance From food cost to cash flow: every number that steers your business, in one guide. Open the guideThe 9 numbers, and the band they belong in
Every percentage below uses turnover excluding VAT, on both sides of the division. That sounds obvious and it is the single most common error in the trade: put your purchase invoices ex VAT against your till turnover including VAT and you flatter your food cost by four to six points, then believe it for years.
1. Food cost: what you buy, divided by what you sell
The classic. All food and beverage purchases in a period, divided by turnover in the same period, both excluding VAT. Measure consumption, not ordering: what is sitting in your walk-in on 31 December is not a cost yet. Without opening and closing stock you are measuring your purchasing habits rather than your usage.
The band varies sharply by concept: roughly 28–34% for a bistro, 30–38% for fine dining, 22–30% for a café and 26–33% for a fast format. Split food and drink and you will often find one of the two is carrying the average — drink belongs somewhere between 18 and 25%, and a beverage margin that slips is usually a pouring problem, not a purchasing problem.
The trap: food cost cannot be read on its own. Buying prepped components trades labour cost for purchase cost. A kitchen can sit comfortably at 36% food cost and still be healthier than the one next door at 29%. That is why this number never stands alone — always read it with point 3. To see where the cents go dish by dish, the recipe costing tool does the arithmetic.
2. Labour cost: everything staff costs, not what the payslip says
The number most often flattered. Labour cost is not net pay and not even gross pay: it is gross plus employer charges, holiday pay, year-end bonuses, meal allowances, insurance, casual staff, students and agency cover. In many countries the real cost runs 25 to 45 per cent above gross.
And then the point most owner-operators trip over: your own wage belongs in there, at market rate. If you work sixty hours a week without paying yourself, you are subsidising your own business and measuring a fiction. Use what you would have to pay a replacement to cover your hours. If nothing is left after that, you have learned something important you did not know before.
The band moves with the country. A bistro in Denmark running at 38% labour is perfectly fine; the same 38% in Romania points to a structural problem, because both wages and menu prices sit lower there. The calculator below shifts this band automatically with the wage level of your market. To tighten the rotas themselves, use the staff schedule maker.
3. Prime cost: the only number you can compare with anyone
Food cost plus labour cost together, as a percentage of turnover. This is the most important figure in the whole article, for one reason: it absorbs the trade-off between the two. Making everything in-house lowers your purchasing and raises your wages; prepped components do the reverse. Prime cost stays comparable under either choice, which makes it the only percentage a pizzeria can meaningfully hold up against a fine-dining kitchen.
The band is the sum of the two bands above — which is exactly how the chart below builds it, rather than as a separate figure that eventually starts contradicting its own components. For a bistro that works out at roughly 58–70%. Sit above it and every other line in your accounts is fighting a losing battle: whatever survives prime cost still has to carry rent, energy, insurance, maintenance, depreciation and your own return.
One point of prime cost on €480,000 of turnover is about €4,800 a year. That is why this figure carries the heaviest weight in the score below. For the full treatment, see the separate article on prime cost.
One bistro, nine lines, and where it really goes wrong
A 55-seat bistro, 10 services a week, 290 covers, €480,000 turnover. The green bar is the band, the tick is this business.
Seven of the nine lines sit inside the band or right up against it — this does not feel like a business in trouble, and it does not feel like one to the owner either. Yet 2.1% is all that remains and turnover may fall just 3.9% before it turns into a loss. The cause is not in the two lines at the bottom: those are the consequence. It is in the 0.8 points of labour cost and the 0.2 points of prime cost above them, plus an average spend hanging just under the band. Three small deviations, together worth the difference between a 2% and a 6% net margin. That is why you read the nine together.
4. Occupancy cost: the one cost you cannot work away
Rent plus every fixed charge attached to the building — property tax, service charges, buildings insurance — divided by turnover. The band sits roughly between 6 and 10%. Above 12% it becomes structural: you cannot economise your way out of an expensive lease, you can only outgrow it or move out of it.
This is the figure that quietly sinks most acquisitions. A site at 14% occupancy means almost a month and a half of turnover a year goes to the landlord before a single gram of product has been bought. How much you can still do about it depends on the years left on the lease — in a takeover that is, along with the price, the most important question to ask, and the valuation tool factors it in.
If you are too high and cannot move, the only way out is the denominator: more turnover from the same square metres. Which brings you to points 6 through 8.
5. Net margin: what actually remains, with your own wage in it
Turnover minus everything, divided by turnover — after you have paid yourself a market wage. For an owner-operated restaurant the band sits roughly between 3 and 8%. That is lower than most people assume, and it explains why a business that "does well" still never builds a buffer.
Why those few points matter so much: this is the money with which your business pays for its own future. A combi oven dies, a walk-in gives up on a Saturday, the terrace needs replacing, the dining room needs repainting every eight years. At 1% net margin every one of those moments comes out of a loan. At 6% it comes out of the till. That difference decides whether you are an entrepreneur or a firefighter — see also why the walk-in always dies on a Saturday.
Watch what sits inside "everything": depreciation belongs in there, your personal drawings do not. Confusing those two is why paper profit and bank balance so often look nothing alike.
6. Average spend per guest
Turnover divided by covers. The simplest number on the list and by far the fastest lever, because it works without one extra guest, one extra opening hour or one extra member of staff.
Run the arithmetic: 400 covers a week is 20,800 a year. Two euros more per guest — an aperitif suggested more often, a coffee with something beside it, a bottle instead of two glasses — is €41,600 of extra turnover a year, most of which drops straight through to your margin because your fixed costs do not move. There is no other line in this article where so little effort returns so much.
If you sit below the band, look at your menu before you look at your team. Menu engineering and rewriting descriptions usually achieve more than a selling course; the menu price calculator shows what a price change does to your margin.
Where €100 of turnover goes
Two bistros with the same turnover, the same rent and the same other costs. Only their prime cost differs.
Busy, feels stretched, keeps nothing.
Same room, same rent, ten times as much left.
Nine points of prime cost between them. On €480,000 of turnover that is €43,200 a year — enough to be the difference between a business that refits its own kitchen and one that has to go to the bank for it.
7. RevPASH: revenue per available seat-hour
The metric hotels have used for thirty years and restaurants barely know. You divide turnover by seats multiplied by the hours you are open. The answer is in euros per seat per hour, and it is the only number that can catch "full but earning nothing".
A dining room packed on a Thursday lunchtime with €14 tickets earns less per seat-hour than the same room half full on a Friday night at €48 a head. Occupancy alone therefore says nothing: you can be sold out and still sit below your band. RevPASH combines the two, which makes it the number you use to decide which services to extend and which to drop.
What to do with it: opening hours that sit structurally below the band cost you wages and energy without earning them back. Model what dropping or moving a service is worth with the revenue simulator, and read the detail in the article on RevPASH.
8. Seat fill: how many of your seats you actually sell
Covers divided by the seat-places you offer — seats multiplied by services. For a bistro the band sits somewhere between 45 and 65%.
One hundred per cent is explicitly not the goal. Above roughly 75% you start losing money rather than making it: you turn walk-ins away, you cannot turn a table when a party runs long, and a single late cancellation can no longer be absorbed. Sitting structurally too high is a sign you have too little capacity or too few services, not that you are doing it right.
Two things quietly eat this number. No-shows, obviously — but also table fit: a party of two on a table for four sells two seats and blocks four. Fail to match your table mix to your party sizes and you lose occupancy that never shows up in the accounts. A floor plan and a no-show policy each tackle half of that problem.
9. Safety margin: how far turnover can fall before you make a loss
The last number, and the only one about survival rather than return. You work out your break-even turnover — the point where you exactly cover your costs — and see what percentage above it your current turnover sits. Below 12% it gets tight; below 5% you are one bad quarter from the red.
Why this is separate from your net margin: two businesses with identical profit can have wildly different safety margins. An operation built on casual and flexible staff scales back quickly when turnover disappoints. One carrying a large permanent team and an expensive lease cannot, and drops through break-even far faster on the same fall in turnover. The calculator below assumes roughly 35% of your wage bill genuinely flexes with turnover — the rest you carry whether anyone walks in or not.
This is the number to hold up against your bank balance before you sign for an investment. The cash-flow planner shows when it gets tight; the break-even analysis explains the calculation.
Put your own numbers beside them
Pick your segment, fill in what you know from your accounts, and the tool lays your nine numbers next to the band for your market. Everything is annual and excludes VAT. Nothing leaves your device — there is no server and nothing is stored.
If you do not know an amount exactly, estimate it. An estimate that is ten per cent out rarely changes the conclusion; what matters is the distance to the band, not the third decimal.
Benchmark your business
Nine numbers, your segment, your market — and the difference in money per year.
Read the result from the bottom up: not the score, but the weakest link. A 71 out of 100 with one line far outside the band is a much clearer instruction than a 71 where all nine sit just outside it — and it is almost always the first case.
Two things are deliberately left out of the cost gap. Prime cost is purchasing plus labour, and both are already in there; net margin and safety margin are the consequence of the rest. Including them would add the same money three times and hand you a figure that exists nowhere.
What to do with it this week, this month and this quarter
Nobody fixes nine numbers at once. This order does work, because each step makes the next one affordable.
This week — measure once, correctly
- Put purchases and turnover both excluding VAT. If you are doing this for the first time, expect your food cost to come out four to six points higher than you thought.
- Count opening and closing stock, or you are measuring orders instead of usage. The stock tool does that count for you.
- Give yourself a market wage inside the labour cost. Without that figure, not one line below it is right.
- Fill in the tool above and write down only your weakest link. Nothing else.
This month — tackle the weakest link
- Food cost outside the band: cost your ten best-selling dishes in the recipe costing tool. In practice two or three dishes are badly out and are carrying the whole average.
- Labour cost outside the band: put your rota next to your turnover per hour instead of next to your instinct. Usually the difference sits in one over-staffed service, not in the whole team.
- Spend per guest outside the band: rewrite the descriptions on your menu and build in two deliberate suggestions. It costs an evening and works from the next service on.
- Occupancy outside the band: dig out your lease and find the years remaining. That decides whether you negotiate or move.
This quarter — rebuild the buffer
- Repeat the measurement after three months using the same method. Only then are you measuring a change rather than your own measurement error.
- Put your safety margin next to your bank balance in the cash-flow planner before you sign anything.
- Only look at the second-weakest link once the first is inside the band — or demonstrably cannot move.
- Lay the nine numbers side by side once a quarter. More often is noise; annually is too late to still do something about it.
A number only becomes useful next to a band
The nine benchmarks above are already in your accounts. What was missing was not more data but the range beside them — and the insight that prime cost and safety margin are the two lines that drive the other seven.
Set the measurement up correctly once, pick your weakest link and leave the rest alone for now. A business that pushes one line into the band every quarter is in a materially different place after a year than one that tries to fix nine numbers at once and moves none of them.
And do not count on it fixing itself once you get busier. More guests on a prime cost above the band is more work for the same result — which is precisely what venue A above has been doing for years.