You look at the POS report on a Saturday night: 140 covers, a waitlist, a full bar. Then you look at the bank account on Monday and it's thinner than it was on Friday. This is the single most common story in the restaurant business, and it almost never has a dramatic cause. It has an arithmetic one.
The arithmetic is food cost percentage — the share of each sale that goes back out the door as ingredients. Get it wrong and no amount of full tables saves you. Get it right and a modest neighbourhood place pays its owner a salary. Below is the version I wish someone had walked me through before the first menu went to print: how to calculate the real number, where it leaks, how to find the dishes that are quietly bleeding you, and what a weekly routine looks like that keeps it from drifting back.
Myth 1: "If the place is full, the money will follow"
This is the belief that kills more first restaurants than bad cooking. Revenue and profit are two different reports, and busyness actively hides cost problems because nobody has time to look.
Take a plausible example. A 60-seat bistro does 4,000 covers a month at an average check of $28. That's $112,000 in revenue. Sounds healthy. Now suppose the actual food cost is 38% instead of the 30% the owner assumed when pricing the menu. The gap is 8 points — about $8,960 a month, or roughly $107,000 a year, quietly disappearing. Nobody stole it. It went out as over-portioned pasta, a case of produce binned on Sunday, and a supplier invoice nobody renegotiated since opening.
The misconception comes from a reasonable place: in most retail businesses, footfall is the hard part. In food, footfall is the easy part — people are hungry every day. The hard part is restaurant profit margin, which is what's left after food, labour, rent, and the long tail of everything else.
A rough shape of where a dollar of revenue goes in a full-service restaurant:
| Cost bucket | Typical share of revenue |
|---|---|
| Food and beverage cost | 28–35% |
| Labour (incl. payroll taxes) | 25–35% |
| Rent and occupancy | 6–10% |
| Utilities, marketing, supplies, repairs | 8–12% |
| What's left as operating profit | often 3–8%, sometimes less |
Read that bottom row again. A restaurant running at 38% food cost instead of 30% isn't trimming profit — it's erasing it entirely and eating into the owner's reserves. That's the mechanism behind "we're always packed and always broke." More on the topic: Prompts That Read a P&L Like a CFO.
Myth 2: "Food cost percentage is just ingredient cost divided by the menu price"
That's the theoretical number, and it's the one every spreadsheet starts with. It is also the one that lies to you, because it assumes perfect portioning, zero waste, and prices from the day you wrote the menu.
The number that matters is actual food cost percentage, and the only honest way to get it is to divide what you actually spent on food over a period by what you actually sold in that same period:
Actual food cost % = (opening inventory + purchases − closing inventory) ÷ net food sales × 100
Do it weekly, not monthly. Monthly is too slow to catch anything — by the time the statement lands, the damage is four weeks old.
Here's how it looks in practice. A kitchen starts the week with $9,400 of inventory, buys $14,200 during the week, and counts $8,100 at the end. Food consumed: 9,400 + 14,200 − 8,100 = $15,500. If net food sales for the week were $46,000, actual food cost is 15,500 ÷ 46,000 = 33.7%.
Now the interesting part: compare that to the theoretical cost of everything you sold, calculated from recipes and POS mix. If theoretical says 30% and actual says 33.7%, you have a 3.7-point gap — about $1,700 that week that no recipe accounts for. That gap is the whole game. It is never a rounding error; it is always a specific, findable behaviour.
Where the four leaks actually live
Once you have the gap, you go hunting in exactly four places. In my experience they show up in a fairly predictable order.
Portioning. A cook who free-pours oil, adds "a bit more" protein to avoid a complaint, or eyeballs the cheese. One extra 20g of protein on a $24 dish at $12/kg is $0.24 — trivial per plate, $960 a month across 4,000 covers. This is why a kitchen scale on the pass is not an insult to the chef; it's the cheapest cost-control tool in the building.
Waste and spoilage. Over-ordering fresh produce for a menu that sells unevenly, prep done for a forecast that didn't happen, staff meals and mistakes that never get logged. The tell is a walk-in that's full on Monday and half-rotten on Thursday. If you have ever thrown out a case of herbs you bought on spec for a garnish that sold twice, you know the feeling.
Supplier pricing drift. Invoices creep. A case price goes up 6% and nobody notices because the order is placed the same way it's been placed for two years. This is where a simple quarterly re-quote across your top ten items pays for itself in an afternoon — but only if you have current spec sheets to compare against.
Menu mix. This one is invisible without data. A dish can be perfectly portioned, made from well-priced ingredients, and still destroy your margin because it sells in volume at a price that never covered its cost. Which brings us to the part most owners skip.
Myth 3: "Menu pricing is cost plus a comfortable markup"
Cost-plus is where menu pricing starts, not where it ends. The problem is that it prices dishes in isolation and ignores the thing that actually determines your margin: which dishes sell.
This is what menu engineering is for, and it's less mysterious than the name suggests. For every dish you plot two numbers: contribution margin (price minus food cost, in dollars) and popularity (share of orders versus the average). That gives you four buckets:
- Stars — high margin, high popularity. Protect them; don't touch the recipe or the price.
- Plowhorses — popular but low margin. These are your volume dishes with a pricing or portion problem. Small price moves or a cheaper garnish often fix them without hurting sales.
- Puzzles — high margin, low popularity. Usually a description, placement, or naming problem, not a food problem.
- Dogs — low margin, low popularity. These are the ones to cut, and cutting them is the fastest cost win on the menu.
Here's the concrete case. A pasta dish sells for $19 with a food cost of $7.60 — a 40% food cost and an $11.40 margin. It's the third-best seller. A steak sells for $34 with a food cost of $12.20 — 36% and a $21.80 margin — but only sells a handful a week. The pasta looks "fine" because it's popular, but every one of those sales earns $10 less than the steak. If you can move 15 covers a week from the pasta to the steak with a server recommendation, that's $150 a week, $7,800 a year, from a sentence at the table.
That's what menu engineering actually does: it stops you from pricing on gut feel and starts you pricing on the mix you're actually selling. If you want the deeper version of how a menu is built from the kitchen side — technique, sauces, dough, the craft layer underneath the numbers — the related breakdown on Professional Culinary Arts covers that half of the job. More on the topic: Coffee Craft.
Myth 4: "You fix food cost once, at opening"
Food cost is not a setting. It's a habit, and habits decay. Prices move, staff turn over, a new dish gets added without a recipe card, a supplier changes a pack size. Within a quarter, an untended kitchen drifts two to four points — which is exactly the margin you were counting on.
The fix is a weekly routine, short enough that a busy owner will actually do it:
- Sunday or Monday: count inventory. Same time, same sheet, same person where possible. Consistency matters more than precision here.
- Pull purchases for the week from supplier invoices.
- Calculate actual food cost with the formula above and compare it to theoretical.
- If the gap is over two points, investigate one leak — portioning, waste, pricing, or mix — not all four. Chasing everything at once finds nothing.
- Review the top five and bottom five dishes by margin and make one change: a price, a portion, a description, or a cut.
That's maybe ninety minutes a week. It will not make you a better cook. It will make the difference between a restaurant that survives its second year and one that doesn't.
Where structured learning actually helps — and where it doesn't
The honest trade-off: you can learn all of this from a spreadsheet and a few good books. Plenty of operators have. The cost is time and expensive mistakes — you learn portioning control by watching $2,000 of protein walk out the door, and menu engineering by discovering a loss-making dish six months after it launched.
Self-study works well if you already have a kitchen, a POS that exports clean data, and the discipline to build the weekly habit yourself. It works badly if you're planning to open and have no live numbers to test your thinking against, because there's nobody to tell you that your assumed 30% food cost is optimistic for the concept you've chosen.
A structured course sits in the middle: it gives you the calculations with worked examples and a sequence, so you're not inventing the method from scratch. That's the gap the Cooking & Restaurant Business course is built to close — it combines kitchen technique with the management side (finance, staffing, supplier logistics) and works through real cases of restaurants that succeeded and failed, which is the closest thing to borrowed experience you can get without opening your own place first. The lessons adapt to your level, so if you already run a kitchen you're not sitting through knife skills again.
What it won't do is replace your own numbers. No course can tell you your actual food cost — only your Sunday count can.
Who this is for, and who it isn't
If you run a restaurant and can't reconcile a busy room with a thin bank account, this is your problem to solve and the method above is where to start this week. If you're planning to open, learn the food cost math before you write the menu, because menu pricing is where the concept either works or doesn't — and it's much cheaper to fix on paper.
If you're looking for a creative cooking career and have no interest in the P&L, be honest about that. The craft path and the business path are different jobs, and the people who struggle most are the ones who love the kitchen and resent the spreadsheet. The spreadsheet is not the enemy of good food; it's what lets you keep making it.
Start with one number: your actual food cost percentage for last week. If you can't produce it, that's the first thing to fix — and the course walks through exactly this calculation with worked examples, plus the menu-engineering pass and the weekly routine that keeps it honest. Details and enrolment are on the course page.