Restaurant Break-Even Point: How to Calculate It on Your Own Numbers

What break-even means in a kitchen, and why the shop version does not fit

In a shop the idea is easy to picture: you buy a thing for one price, sell it for another, and the difference pays the rent. A restaurant is harder for one reason — most of what you sell did not exist when you bought it. You buy flour, gas and an hour of a cook's time; you sell a plate. The cost of that plate is assembled inside your own building, out of items that behave differently when the room fills up.

That is the whole difficulty. Break-even arithmetic is not hard; sorting your own costs into the two buckets the arithmetic needs is hard, and that is where an owner ends up with a number that looks precise and is wrong.

• Your product is perishable and your capacity is time-boxed. An unsold dress is inventory. An unsold seat at 20:30 on Friday is gone, and so is the fixed cost that seat was supposed to carry.
• Labour sits on both sides of the line. A salaried head chef is a fixed cost. A student on an hourly contract called in for Saturday is a variable cost. Most restaurants run both, in a mix that changes weekly.
• Your sales are not one product. A plate, a bottle of wine and a delivery order carry very different variable costs, so a threshold calculated on last month's mix quietly assumes next month's mix will match.

None of this changes the formula. All of it changes what you are allowed to put into it.

The arithmetic: fixed costs divided by what each unit of sales leaves behind

The whole calculation is one division. Everything before it is bookkeeping and everything after it is interpretation.

Break-even sales = Fixed costs ÷ Contribution margin ratio

• Break-even sales — net sales for the period at which profit is exactly zero, PLN;
• Fixed costs — total costs for that same period that do not move with the number of covers, PLN;
• Contribution margin ratio — the share of each 1 PLN of net sales left after variable costs, a decimal between 0 and 1.

The period matters as much as the numbers. Fixed costs and sales must cover the same calendar window — a month of rent against a month of sales, never a month of rent against an average week. Mixing periods is the commonest way this calculation fails, and it fails silently, because the output is still a plausible number.

The same arithmetic, step by step

The figures below are round numbers chosen only to show where each one goes. They are not a benchmark, not an average and not a real restaurant — put your own in and the steps are identical.

Take a month with net sales of 200 000 PLN, variable costs of 120 000 PLN and fixed costs of 60 000 PLN.

1. Contribution margin: 200 000 − 120 000 = 80 000 PLN.
2. Contribution margin ratio: 80 000 ÷ 200 000 = 0.40.
3. Break-even sales: 60 000 ÷ 0.40 = 150 000 PLN.

Read the middle line out loud, because the rest of the page depends on it: every 1 PLN through the till leaves 40 groszy behind to pay for rent, so you need enough sales for those 40-grosz slices to add up to 60 000. That is 150 000 PLN, and there is nothing else in the calculation.

Contribution margin: the number that decides how fast you climb out

Contribution margin — the money left from net sales after variable costs, available to cover fixed costs and, once they are covered, to become profit. Expressed as a share of net sales it is the contribution margin ratio; expressed per guest it is the contribution margin per cover.

Contribution margin ratio = (Net sales − Variable costs) ÷ Net sales

• Net sales — sales for the period without VAT, without discounts granted and without voided or comped bills, PLN;
• Variable costs — costs for the same period that move with the number of covers, PLN.

Contribution margin per cover = Average check − Variable cost per cover

• Average check — net sales for the period divided by covers served, PLN per cover;
• Variable cost per cover — variable costs for the period divided by the same covers, PLN per cover.

It is not gross margin. Gross margin usually stops at the cost of food and drink. Contribution margin keeps going and removes everything else that moves with volume — packaging, the platform's commission on a delivery order, the card fee, the hourly staff called in because the room was busy. A restaurant with a comfortable gross margin and a heavy delivery mix can have a contribution margin ratio far below what its food cost suggests, and the threshold shows that immediately where the gross margin hid it.

It is a blend, not a property. You do not have "a" contribution margin ratio. You have one per dish, one per channel, one per daypart, and the figure in your formula is the weighted average of all of them at last period's mix. Push a cheaper lunch menu, take more delivery, sell more wine, and the ratio moves without a single price changing. That is why a restaurant's break-even point drifts in months when nobody touched a price or a supplier.

Two neighbouring figures live on their own pages and are deliberately not re-derived here: food and labour cost together as a share of sales is prime cost, and what survives at the very bottom, including lines that never appear in this formula, is net profit margin.

Fixed, variable, and the costs that only pretend to be fixed

Fixed cost — a cost that does not change with the number of covers within the period: rent, insurance, base salaries, licences, subscriptions.

Variable cost — a cost that moves with each cover: ingredients, packaging, platform commission, card fees, hourly and agency labour.

Semi-variable cost — a cost with both parts, a base charge plus usage: energy, a POS plan with per-order pricing, laundry with a minimum collection.

The third category is where the calculation is usually lost. Energy is the clearest case: the walk-in runs whether or not anyone books a table, and the combi oven runs harder when the room is full. Put the whole bill in fixed costs and your threshold is too high; put it all in variable and your contribution margin ratio is too low. Both errors are invisible, because the result is still a number.

Splitting a semi-variable cost without guessing

You do not need a submeter. You need two periods already on file — your busiest and your quietest — each with its cover count and its bill.

Variable rate = (Cost in the busy period − Cost in the quiet period) ÷ (Covers in the busy period − Covers in the quiet period)

Fixed part = Cost in the busy period − Variable rate × Covers in the busy period

• Cost — the same bill in both periods, PLN;
• Covers — guests served in each period, count;
• Variable rate — PLN of that cost per cover;
• Fixed part — the part of the bill that would appear at zero covers, PLN.

With round demonstration figures: an energy bill of 9 200 PLN in a month with 2 400 covers, and 7 400 PLN in a month with 1 500 covers. Variable rate = (9 200 − 7 400) ÷ (2 400 − 1 500) = 1 800 ÷ 900 = 2.00 PLN per cover. Fixed part = 9 200 − 2.00 × 2 400 = 4 400 PLN. Check against the quiet month: 7 400 − 2.00 × 1 500 = 4 400 PLN — the same figure, which is what tells you the split is arithmetically consistent rather than merely plausible.

One honest limitation: the method assumes the two periods differ mainly in volume. February and July also differ in weather, and the split will quietly load heating into the variable rate. Pick two periods in the same season where you can.

The cost that belongs in neither column

If you work in your own restaurant without paying yourself a wage, your labour is not in the fixed costs — so the threshold you calculate is the level at which the business breaks even while your time is free. That is a legitimate number, but not the one that answers whether the restaurant is worth running. Calculate it twice: once as the books show it, once with a market salary for your own role added to fixed costs. The gap is what your own labour currently subsidises.

Net sales in Poland: what comes off the till total before the arithmetic starts

This section and the two below it use Polish rules and Polish official statistics. Elsewhere the method transfers and the numbers do not.

The formula says net sales, and in Poland that word does real work. The national statistics office publishes revenue from catering activity the other way round — including VAT, stated explicitly in its definition of the indicator (GUS, report on the internal market in 2024, published 03.11.2025). Drop a headline figure from that publication into this formula unchanged and your contribution margin ratio is overstated by roughly the tax you never got to keep.

Three things come off the gross till total before a number may enter the formula:

1. VAT. The usual shorthand "catering is 8 % in Poland" is not quite what the law says, and the difference matters here. Article 41(2) of the Polish VAT Act sets the reduced rate and explicitly excludes PKWiU 56 catering from it; article 41(12f) then applies that same rate back to services related to the provision of food (PKWiU 56), with drinks carved out; and article 146ef(1)(2) sets that rate at 8 % for the current period, against a standard rate of 23 % (Polish VAT Act, Dz.U. 2004 nr 54 poz. 535, consolidated text as at 20.07.2026). The practical consequence: a plate and a drink can leave the same table under different rates, so your effective VAT share follows your own mix, not a single percentage.
2. Discounts actually granted — the money that never arrived, not the list price with a note beside it.
3. Voided, comped and written-off bills. These are not sales. Leaving them in inflates sales and, because their variable costs were genuinely incurred, hides the loss.

Break-even in money, in covers and in days

One threshold, three units, and they are not equally usable: money per month is auditable, covers per shift is actionable, days into the month is the version people repeat.

Break-even covers = Fixed costs ÷ (Average check − Variable cost per cover)

• Fixed costs — fixed costs for the period, PLN;
• Average check — net sales per cover, PLN per cover;
• Variable cost per cover — variable costs per cover, PLN per cover.

Continuing the demonstration figures — fixed costs 60 000 PLN, average check 80 PLN, variable cost per cover 48 PLN:

• contribution margin per cover: 80 − 48 = 32 PLN;
• break-even covers: 60 000 ÷ 32 = 1 875 covers per month;
• across 25 trading days: 1 875 ÷ 25 = 75 covers per day;
• at 200 000 ÷ 25 = 8 000 PLN of net sales a day: 150 000 ÷ 8 000 = 18.75, so the month turns profitable partway through day 19.

Running both versions is the point: 1 875 covers × 80 PLN = 150 000 PLN, exactly the threshold the first formula gave. If your two versions disagree, one input was measured over a different period than the other, and you have found the error before it reached a decision.

Why the Eurostat purchases-to-turnover line cannot be split into fixed and variable

There is no official figure for "the normal fixed-cost share of a Polish restaurant". There is official structural data for the sector, and reading it carefully explains why the first thing does not exist.

The whole table from that dataset — its remaining rows and the warning each of them needs — belongs to the prime cost page, where that same table stands in full. It is not repeated here: one table living in two places starts to disagree with itself.

And that single line is exactly why no fixed-cost benchmark can be extracted from it. It mixes food, drink, rent, utilities, cleaning and outsourced services — it mixes the numerator of your calculation with its denominator. As a measure of how much of the sector's turnover leaves the sector it is fine; as a target for either of your two buckets it is useless, and anyone who converts it into one has produced an invented number with an official-looking citation attached.

A separate line of that same dataset says more about the fixed-variable split than all the rest of it put together: spending on agency workers. Agency labour is the purest variable labour there is — you pay for the shift you called. Across Poland's whole food-service sector it comes to 15.9 mln EUR a year; across the EU-27 the same line is 2 886.0 mln EUR (same dataset sbs_ovw_act, data updated 10.03.2026). Proportionally to its own market's turnover, the European sector spends several times more of it on labour it can switch off. That does not tell you what to do. It tells you why "just cut hours when it is quiet" is easier to write than to execute in a Polish kitchen, and why more of your labour probably belongs in the fixed bucket than you would like.

Break-even and cost-share bands are quoted everywhere in the trade press, in software vendors' blogs and in benchmark guides. Follow them to their source and you generally arrive at another vendor blog. There is no official statistic behind them, and we will not print one as though there were. What is above is what official sources actually say, with dates and caveats. Everything else here is method, so that the number you finish with is yours.

What moves the threshold, and which changes move both halves at once

A break-even point can move for two structurally different reasons: fixed costs changed, or the contribution margin ratio changed. Confusing them means pulling the wrong lever — a rent rise cannot be answered by improving food cost by the same percentage, because the two enter the calculation in different places and with different force.

The minimum-wage row is the one people get wrong, and it is why that table has two columns instead of an arrow. In Poland, from 1 January 2026 the minimum monthly wage is 4 806 PLN and the minimum hourly rate is 31.40 PLN (Council of Ministers regulation of 11 September 2025, Dz.U. 2025 poz. 1242, promulgated 15.09.2025). A rise in that rate lifts the base pay of salaried and guaranteed-hours staff — the numerator, fixed costs — and the cost of every extra hour called in on a busy Saturday — the denominator, through variable cost per cover. Both halves of the division move in the direction that raises the threshold. Model only one and you understate what the change costs you.

That figure is replaced each September by a new regulation for the following year. Reading this later, check the current one before putting it into anything.

Reading a price change without inventing the outcome

If prices rise by a factor and variable cost per cover does not move, the new ratio is (Net sales × (1 + p) − Variable costs) ÷ (Net sales × (1 + p)), where p is the increase as a decimal. The same variable cost is now divided into a bigger number, so the ratio rises and the threshold falls — before any change in guest behaviour.

What arithmetic cannot supply is what happens to covers afterwards, and that is the whole question. Model the rise you are considering against the fall in covers that would exactly cancel it, and you at least know how much volume you are betting. Keeping both sides of that trade-off in a running model is what a what-if service is for; keeping the cost inputs behind it current is what a finance service is for.

Margin of safety: how far sales can fall before you reach the line

Break-even says where the line is. It does not say how close you are standing to it, and that is the figure that actually predicts trouble.

Margin of safety % = (Current sales − Break-even sales) ÷ Current sales × 100

• Current sales — actual net sales for the period, PLN;
• Break-even sales — the threshold calculated above for the same period, PLN.

Sales could fall by a quarter before that month broke even.

There is no official figure for a "healthy" margin of safety and this page will not invent one. What is worth knowing is what shrinks yours, usually without an announcement: a fixed cost that rose against flat sales; a mix shift towards lower-contribution channels, which moves the line without any cost changing at all; a quiet erosion of average check through discounting, portion creep or a promotion that outlived its campaign; and seasonality, which does not so much shrink the margin as reveal that the annual figure was never true for any single month.

Seasonality: why an annual break-even lies and a monthly one does not

Run the calculation on twelve months at once and you get an average describing no month you actually trade. A restaurant with a strong summer and a thin February can sit comfortably above its annual threshold and below its February one — and the rent is paid out of February's takings, not out of the year's average.

The failure is arithmetic, not intuition. Fixed costs are close to constant across the year; sales and mix are not. Dividing a twelve-month fixed-cost total by a twelve-month blended contribution margin ratio produces a threshold too low for the quiet months and too high for the peak ones, in both cases by an amount the annual figure hides.

So calculate it month by month, each with its own contribution margin ratio, and read three things:

1. Which months sit below their own threshold. Not "how many" — that answer is whatever your data says, and anyone who states the count before seeing your data is guessing.
2. How much those months must be carried by the rest. Their shortfalls, added up, are the surplus the good months have to generate before the year makes anything.
3. Whether the shape is the calendar's or yours. A cost you could have moved out of a quiet month — a maintenance contract, a marketing burst, a holiday rota — is a decision, not a season.

Seasonal fixed costs deserve a note of their own. A summer terrace carries rent, licences and equipment that only earn for part of the year, and charging them entirely to the months they were paid in makes those months look worse than the decision was. Spread them across the months they serve, and write in your own notes that you have done so.

Doing this by hand every month is where most owners stop doing it at all. A forecast service keeps next month's expected covers and mix current instead of rebuilt from scratch, and an analytics service keeps the actuals behind it in one place. Putting the threshold itself on a screen someone reads every evening is a separate subject, covered in the restaurant KPI guide.

A restaurant that has not opened yet: what the formula cannot see

The formula needs a contribution margin ratio, and that needs sales. Before opening you have neither, so what you are doing is building a plan and testing it — legitimate, as long as you say out loud which numbers are assumptions.

• Fixed costs are the most knowable thing about a restaurant that does not exist yet. Rent is in the lease, insurance is quoted, base salaries are decided, subscriptions are priced. Add them honestly, including the ones that feel too small to bother with.
• Variable cost per cover comes from recipe costing on your own menu at your own supplier prices, plus packaging, plus the card and platform percentages you have actually been quoted.
• Average check comes from your own menu and an assumed mix. Write the assumed mix down: it is the assumption most likely to be wrong and most likely to be forgotten.

Two things sit outside the formula entirely, which is why a pre-opening threshold is always optimistic if you stop at the division. Capital expenditure is not a fixed cost — fit-out, equipment and deposits do not appear in the monthly fixed-cost line, only their depreciation does, and depreciation is smaller than the cash you spent. Accounting break-even is not the same as having paid back what you put in. And the ramp is real and is not in the arithmetic: covers and variable cost per cover are not at their steady state in the first weeks. Whether you plan for that period is a question about working capital, and this page will not tell you how long it lasts, because there is no honest general answer.

The fixed-cost line is usually underestimated in the small recurring items — booking, POS, payroll, telephony, accounting. What that systems layer costs is worked through in what process automation costs a small business, and what actually sits in a restaurant's own stack in restaurant automation: reservations, suppliers, reviews.

Where the two inputs actually come from in a working restaurant

The calculation fails in practice for a boring reason: nobody owns the two numbers. Sales live in the POS, purchase invoices live with the accountant, the rota lives in a messaging app, and the platform commission arrives in a statement after the month is over. Assembling all four once is an afternoon. Assembling them every month is a job, and jobs belonging to nobody stop happening around month three.

• Fix the boundaries of net sales in writing. Decide once how discounts, comps, staff meals and delivery gross-ups are treated, write it down, and do not renegotiate it mid-year. A threshold built on shifting definitions is not comparable to last month's.
• Timestamp the cost side to the same window. An invoice dated the 2nd for goods delivered on the 29th belongs to the previous month's variable costs. Cash-basis bookkeeping will file it wrongly, and at restaurant margins that misplacement is larger than the profit it distorts.
• Recalculate when the structure changes, not on a schedule. A new lease, a new channel, a wage change, a renegotiated supplier — each makes last month's ratio obsolete immediately, and none of them wait for the quarter to end.

Two of the largest moving parts have pages of their own: what is actually left to you after a delivery platform takes its share, in online orders for restaurants; and events and catering, a channel with a completely different variable cost structure, which is why blending it into one ratio distorts both, in event catering: from phone call to offer. With more than one site the trap is averaging — two locations with different rents and different mixes have two thresholds and one meaningless average, the subject of managing a restaurant chain from one screen. And which numbers an owner will actually look at rather than admire is in reporting automation.

Frequently asked questions

How do you calculate the break-even point of a restaurant?

Divide total fixed costs for a period by the contribution margin ratio for that same period. The contribution margin ratio is net sales minus variable costs, divided by net sales. Both figures must cover the same calendar window and the same definition of net sales — without VAT, without discounts granted and without voided or comped bills. The result is the level of net sales at which profit is exactly zero.

Which restaurant costs count as fixed?

Costs that do not change with the number of guests within the period: rent and service charge, insurance, licences and permits, base salaries and salaried management, software and telephony subscriptions, accounting fees, and depreciation of equipment. The test is not whether the amount is large or predictable, but whether it would still appear if you served nobody at all.

Is labour a fixed or a variable cost in a restaurant?

Both, and the split is yours to determine rather than a general rule. Base salaries and contractually guaranteed hours behave as fixed costs; hourly staff called in for demand, overtime and agency shifts behave as variable costs. Split payroll along that line rather than by job title, because the same role can sit on either side depending on the contract behind it.

How many covers a day do I need to break even?

Divide fixed costs by the contribution margin per cover — average check minus variable cost per cover — to get break-even covers for the period, then divide by your trading days in that period. Cross-check it: break-even covers multiplied by average check should equal the threshold in money from the first formula. If the two disagree, one input was measured over a different period than the other.

How does raising menu prices move the break-even point?

Through the contribution margin, not through fixed costs. If variable cost per cover stays the same while the average check rises, the ratio rises and the threshold falls. The arithmetic stops there — what happens to your covers afterwards is not something a formula supplies, so model the rise against the fall in covers that would exactly cancel it, and you can see how much volume the decision is betting.

How do I read my own margin of safety, and what makes it shrink?

Take current net sales, subtract break-even sales for the same period, divide by current net sales and multiply by a hundred. That is the percentage sales could fall before you reach the line. It shrinks when a fixed cost rises against flat sales, when the mix moves towards lower-contribution channels, and when average check erodes through discounting or portion creep. In quiet months it appears to shrink, when in fact the annual figure was never true for any single month.

Why does an annual break-even figure mislead a seasonal restaurant?

Because fixed costs are close to constant across the year while sales and mix are not. A twelve-month total divided by a twelve-month blended ratio produces one threshold too low for the quiet months and too high for the peak ones. Rent is paid out of February's takings, not out of the year's average, so calculate the threshold month by month, each with its own contribution margin ratio.

What is the difference between contribution margin and gross margin?

Gross margin normally stops at the cost of food and beverage. Contribution margin removes everything that moves with volume — packaging, delivery platform commission, card fees, hourly labour called in for demand. A restaurant with a healthy gross margin and a heavy delivery mix can have a much lower contribution margin, and only the second of the two tells you where the break-even point actually sits.

Work out your own threshold on your own two numbers, then decide what it is telling you to change. The rest of the figures a restaurant owner is expected to be able to defend live on the restaurant guides hub, and this one is the shortest of them. If you would rather not rebuild all of it every quarter, assemble the fixed-cost list and the contribution margin ratio once, then change only what actually moved: the rent after a lease amendment, the wage after a rise, the commission after a new platform contract. The threshold then recalculates in a minute instead of a day, and a dashboard is where that one number belongs in plain sight.