Every winter, restaurants with full dining rooms and profitable books quietly go under. The food was never the problem. The owner watched one number — the bank balance — and trusted a feeling for the rest. Both lie. The question this guide opens with: how can a restaurant make money on paper and still miss the rent in February?
The answer takes six chapters, and none of them require loving spreadsheets. You'll learn to read a P&L the way you read a ticket rail, find the one number that predicts survival better than revenue does, see why cash flow kills healthy businesses, count the exact cover where your evening flips from losing to earning, measure what every seat-hour is worth, and decide which investments deserve your money. Finance, it turns out, is just six recipes. The first one shows where every euro of Saturday night actually went.
The short version
- Read your P&L monthly in percentages, not euros — revenue is always 100%, and every line is a recipe ratio.
- Prime cost (food + drink + labour) is the number: keep it at or under 60% of revenue and the rest of the P&L usually behaves.
- Profit is an opinion, cash is a fact — run a rolling 13-week cash forecast; most restaurant deaths are cash deaths.
- Know your break-even in covers per service — "34 covers on a Tuesday" is a target a whole team can see.
- Measure RevPASH (revenue per available seat hour) to find money hiding in slow turns and empty shoulders.
-
The map
Read your P&L like a recipe: every line is a ratio
A restaurant P&L becomes readable the moment you convert every line to a percentage of revenue: food cost 28–32%, labour 30–35%, occupancy under 10%, everything else 8–12%, leaving a 3–9% net margin. Read monthly in percentages, problems announce themselves.
Euros lie to restaurateurs — a €38,000 month sounds different in July than in January, and costs drift invisibly inside growing revenue. Percentages don't lie. The discipline that changes everything is embarrassingly simple: every month, divide each cost line by revenue, and compare against both last month and the benchmarks below.
The fine-dining P&L, as ratios of revenue Line Healthy range Where it's managed Revenue 100% Reservations & marketing Food & beverage cost 28–32% Menu engineering Labour (incl. you) 30–35% Scheduling & retention Occupancy (rent, utilities) 6–10% Lease negotiation, energy Operating costs (the rest) 8–12% Subscriptions, maintenance, fees Net margin 3–9% Everything above, compounding Two habits make the reading honest. First, pay yourself a real salary inside labour — a restaurant that's only profitable when the owner works free isn't profitable. Second, build the reading from a real budget: forecast each line ahead of the year, then compare actuals monthly. The budget isn't a cage; it's the recipe card for the business itself.
The fifteen-minute monthly ritual
Same coffee, same morning, every month: print the P&L, write the five percentages in the margin, circle anything that moved more than a point, and ask why aloud. That ritual — not an accounting degree — is what financially literate ownership looks like.
Do this tonightTake last month's P&L and write the percentage next to every line (each ÷ revenue). Circle the one furthest from the table above. That circle is your next month's project — and probably worth more than a record Saturday.
-
The number
Prime cost: the one number that predicts survival
Prime cost is food and beverage cost plus total labour, expressed as a share of revenue. At or below 60% a full-service restaurant can usually thrive; at 65% it survives; above 70% it is dying in slow motion regardless of how busy it looks. Track it weekly, not monthly.
If you only ever track one number, make it this one. Prime cost combines the two costs you can actually manage week to week — what you buy and who you schedule — and it moves fast enough to act on. Rent is a yearly negotiation; prime cost is a Tuesday decision.
Why weekly beats monthly
A monthly prime cost of 63% tells you that something went wrong, on average, weeks ago. A weekly reading tells you which week — the over-ordered protein, the over-rostered quiet stretch — while the cause is still in the room. The calculation takes ten minutes once the routine exists: this week's purchases (from invoices) plus this week's labour (from the rota), divided by this week's revenue.
Reading your prime cost Prime cost Verdict The move Under 55% Exceptional — check you're not under-investing in quality or people Consider raising quality, not just margin 55–60% Healthy fine dining Hold the line; tune with the seasons 60–65% Surviving, not compounding One point from food (menu engineering), one from labour (forecast rostering) Over 65% Structural problem Re-engineer the menu and the roster this month, not this quarter The two levers have their own guides: the food side lives in menu engineering (costing, waste, pricing), the labour side in staffing (forecast-driven rosters). Suppliers are the quiet third lever: re-quoting your top ten ingredients twice a year, as covered in negotiating with suppliers, routinely claws back a full point.
Do this tonightCompute last week's prime cost from invoices, rota and revenue — one number, ten minutes. Put it on the same whiteboard as your occupancy. Those two numbers together are 80% of restaurant management.
The Saturday that loses money
Run prime cost per service once and you'll find it: a packed Saturday that earns less than a calm Thursday. Heavy tasting-menu comps, an extra runner "because it's Saturday", premium proteins prepped for walk-ins who didn't come — busy and profitable are different axes. The houses that know their per-service prime cost schedule and prep to the booking curve, and their quiet Thursday quietly out-earns the neighbour's loud Saturday.
-
Oxygen
Cash flow: why profitable restaurants still die
Restaurants fail from cash gaps, not just losses: VAT quarters, supplier terms, December's deposits hiding January's drought. The defence is a rolling 13-week cash forecast, a tax account that's never touched, and one month of fixed costs as a buffer — boring, and life-saving.
Profit is an opinion produced once a month; cash is the fact that pays Friday's wages. The restaurant graveyard is full of houses that were profitable on paper and dead at the bank — killed by a VAT quarter landing in the same week as the insurance annual and a slow February. None of those events were surprises; all of them were unscheduled.
The 13-week radar
One spreadsheet, thirteen columns, updated every Monday in ten minutes: expected money in (booking-led revenue forecast, events, gift card sales), expected money out (wages, rent, suppliers, the VAT quarter, the insurance annual), running balance at the bottom. The radar's only job is to make a week-22 problem visible in week 9, while the fixes are still cheap — shifting a supplier payment, pushing an event, pre-selling a wine dinner. The full method is in managing restaurant cash flow.
Restaurant-specific cash moves
- The untouchable tax account: a fixed percentage of every week's revenue moves automatically to a separate account for VAT and payroll taxes. The single most effective habit in this entire guide.
- Deposits and prepaid menus (chapter 2 of the reservations guide) turn future bookings into present cash — and zero no-shows.
- Gift cards are an interest-free loan from December to your January–February trough. Sell them deliberately.
- Supplier terms are negotiable — moving your two biggest suppliers from 14 to 30 days adds half a month of breathing room permanently.
Do this tonightOpen a savings account named TAX, and set an automatic weekly transfer of your VAT-plus-payroll percentage. Twenty minutes of admin tonight removes the most common near-death experience in this industry.
-
The crossing
Break-even: the covers where you start earning
Break-even in covers = fixed monthly costs ÷ contribution per cover (average ticket minus its variable cost). Expressed per service — "34 covers on a Tuesday" — it converts the whole P&L into a target the entire team can see, count and beat in real time.
Somewhere tonight there is a cover number — maybe 31, maybe 47 — where your restaurant stops paying for the rent and starts paying you. Most owners have never computed it, which means every service runs without a scoreboard. The break-even analysis takes twenty minutes and changes how the whole team sees a Tuesday.
The recipe
- Fixed costs per month: rent, salaried wages, insurance, subscriptions — everything that arrives whether or not a single guest does.
- Contribution per cover: average ticket minus its variable cost (ingredients of that ticket, roughly your food-cost % — plus hourly labour if you scale staff per service).
- Break-even covers = fixed ÷ contribution. Divide across your services and you get the per-night scoreboard.
Worked example: €31,000 fixed monthly, €95 average ticket, 30% variable → €66.50 contribution per cover → 466 covers a month, or roughly 19 per service across 24 services. Suddenly the half-empty Wednesday at 16 covers isn't "a bit quiet" — it's three covers from break-even, and the waitlist nudge from chapter 3 of the reservations guide is worth exactly €199.50.
What break-even teaches pricing
Re-run the formula with a €4 higher average ticket (one aperitif, chapter 4 of the menu guide): break-even falls by ~28 covers a month. Run it with 2% lower food cost: similar. Break-even is where every other guide's work becomes visible as fewer covers needed to be safe — which is why it belongs on the office wall, recalculated every season.
Do this tonightCompute your break-even covers per service with this chapter's recipe — twenty minutes, three numbers you already have. Then tell the team tomorrow's number at the briefing and watch how differently a "quiet night" gets played.
-
The metric
RevPASH: the metric that sees what occupancy hides
RevPASH — revenue per available seat hour — divides revenue by seats × open hours, pricing every seat-hour you own. It exposes what occupancy hides: slow turns, weak shoulders, underpriced peaks. Fine-dining rooms typically run €15–40; the trend matters more than the level.
Occupancy says the room was full. RevPASH asks the sharper question: full of what? A table of two lingering three hours over one bottle and a table of four through a tasting menu in two both count as "occupied" — they are not the same business. Borrowed from hotel revenue management, RevPASH is the restaurant's truest productivity metric because its denominator is the only thing you truly sell: seat-hours.
Using it without a spreadsheet PhD
Revenue ÷ (seats × opening hours), per service. A 50-seat room, open 4 hours, taking €3,800 on Friday dinner: RevPASH €19. The power move is comparing your own services against each other:
What RevPASH gaps mean Pattern Diagnosis Lever High occupancy, low RevPASH Slow turns or soft average ticket Turn-time craft, aperitif & pairing moments Strong 20:00, dead 18:00 Shoulder hours unsold Early-evening products: pre-theatre menu, counter seats Friday ≫ Tuesday (3×+) Demand concentration Events & private dining on the quiet side (reservations guide, ch. 5) Flat everywhere Pricing too timid at peak Premium peak experiences; the menu guide's anchoring Your analytics dashboard can compute it per service automatically; review it monthly next to prime cost. One number for what comes in per seat-hour, one for what goes out — together they are the cockpit.
Do this tonightCompute RevPASH for your best and worst service last week. Write both numbers down and the ratio between them. If it's above 3×, chapter 5 of the reservations guide is your highest-paying reading this month.
-
Compounding
Invest like an owner: every euro must earn its keep
Restaurant investments — terrace, renovation, equipment, software — deserve the same costing as a dish: payback period (investment ÷ monthly gain) and a simple annual ROI. Under 18 months payback is strong; over 36 needs a strategic, not financial, justification.
The first five chapters defend money; this one multiplies it. Restaurants bleed capital on enthusiasm — the €40,000 renovation that "felt right", the combi oven used at half capacity — and starve the boring investments that compound. The cure is one envelope-sized calculation before every yes: the same ROI thinking you now apply to dishes.
The envelope method
- Payback period = investment ÷ extra monthly contribution. A €12,000 terrace adding 90 covers a month at €25 contribution pays back in just over five months — screaming yes.
- Count the costs honestly: the oven's price includes installation, training and the service it disrupts; the terrace includes furniture, permits, and winter storage.
- Count the gains conservatively: use 70% of your optimistic estimate. If it still clears 18 months, proceed.
Where the boring ROI hides
The highest-return investments in this industry are rarely visible to guests: a reservation system that recovers no-shows (often a payback measured in weeks), energy-efficient refrigeration eating a utilities line, training that drops turnover a notch (chapter 5 of the staffing guide priced that), the automation layer that returns ten staff-hours a week. Glamour ages; compounding doesn't.
And when the investment is growth itself — a second room, a bigger lease — the rule hardens: model it on chapter 4's break-even and chapter 3's cash radar first. Growth that outruns cash is how good restaurants die ambitious.
Do this tonightList your last three significant investments and compute their actual payback with real numbers. No judgement — calibration. Your next investment decision just got smarter than your last three.
The cheapest capital in hospitality
It isn't a bank loan — it's pre-sold demand. A winemaker dinner sold out six weeks ahead, deposits on December's groups, gift cards bought in week 50 and redeemed in week 7: all of it is guests financing your cash flow at 0% interest, with zero no-show risk attached. Houses that systematically pre-sell 10–15% of next quarter's revenue rarely need their overdraft — the guest list is the credit line.
How financially fit is your restaurant?
Eight yes/no checks — the same ones a good advisor would ask in the first meeting. Your browser saves the score for when you return.