In this article
Your food cost is three points higher than last year. Your prices haven't dropped, your recipes haven't changed, and nothing has gone missing that looks like theft. There's just a scoop that got bigger than the recipe card says — and nobody ever decided that it should.
Portion creep isn't an accounting problem — it's a physical problem that looks like one. A cook ladles the sauce, doesn't weigh it, and next time serves a fraction more — not out of carelessness, but because a generous hand feels like good service. It happens on every plate, every service, and the recipe card doesn't change a word of it.
The treacherous part is the direction. A portion that shrinks gets noticed instantly: the guest sees it, the cook feels bad about it, and it gets corrected on the spot. A portion that grows is never noticed — the guest is happy, the cook is proud, and nothing signals that anything needs fixing. The recipe card is a floor almost nobody actively defends, not a ceiling anyone bumps into.
And it's invisible to exactly the tools that catch other leaks. A stock count compares what was bought against what was used, and those two numbers keep matching perfectly: the extra food wasn't stolen, it simply left through the pass instead of the back door. Your stock reconciles, your suppliers are honest, and somehow a little less comes through at the end of every month than the numbers promise.
This guide walks through the seven mechanisms behind portion creep, with the number attached to each one. At the bottom, plug your own five dishes into the calculator — recipe-spec cost, measured cost and weekly covers — and see in one screen which dish is bleeding and how much it costs you a year.
Why your food cost is rising with nothing to blame it on
Every costing tool on this site — recipe costing, menu engineering, your food cost percentage — prices the card portion: what the recipe spec says should be on the plate. None of them ever measures what's actually on the plate. That isn't a gap in those tools; it's exactly why this article exists. A recipe costing is only as accurate as the portion a cook actually scoops, and almost nobody checks that after the first week.
Portion creep is also neither waste nor theft, and that distinction matters because the fix is different for each. Waste is food thrown away — trim, spoilage, a wrong plate sent back. Theft is food that leaves the building with no ticket to show for it. Portion creep is food used correctly, eaten by a satisfied guest, and simply more than the menu price ever assumed. There's no culprit — there's a habit.
That's also what makes it the hardest leak to find. Physical spoilage shows up in the walk-in. Theft shows up as a stock variance. Portion creep shows up nowhere — until you actually weigh a plate against the recipe card, and that's precisely the step almost no kitchen ever takes.
The ultimate guide Menu Engineering & Drinks: 6 Steps to More Margin From menu layout to food cost: everything your menu can earn you, in one guide. Open the guideThe 7 mechanisms behind a growing plate
They're listed in the order they actually stack up in a kitchen: first the human hand, then the people around that hand, and only after that the numbers that make it visible.
1. The invisible drift of an eyeballed scoop
Eyeballing a portion isn't an exact science, and it doesn't need to be — a little plate-to-plate variation is human and normal. The problem is the direction of that variation. Studies of portion behaviour in professional kitchens keep finding the same pattern: without recalibration, a scoop, a ladle or a portioning spoon drifts steadily toward the generous side, never the tight side. A cook who's unsure would rather serve a little too much than too little — nobody ever complains about a plate that looks too full.
That's exactly why the recipe card acts as a floor rather than a ceiling. It sets what a portion should be at minimum to look right, and once that minimum is met, there's no signal telling anyone to scale back. Every next serving nudges the new minimum a fraction higher, and after a few months the card portion is a memory rather than an instruction.
This isn't sloppiness or a bad cook — it happens in well-run kitchens just as much as in poorly run ones. It's a human reflex that only gets checked by a scale, a fill line, or someone actively comparing against the card.
2. The "one more for a regular" effect
A familiar guest gets a slightly bigger portion, an extra spoonful of sauce, a friendly gesture from a cook who knows them. On its own that's harmless — it's exactly the kind of attention that sets an independent restaurant apart from a chain. The problem is what happens next: that slightly bigger plate becomes the reference image for the next guest, and the one after that, until it isn't an exception any more — it's the norm.
Staff don't plate to the card, they plate to the last plate they saw leave the pass. A cook who watches one generous plate go out repeats that size for the rest of the shift — not as a protest against the recipe card, but because that's now become the reference point. One kind gesture for one guest quietly becomes the new average for everyone.
The fix isn't being less generous with regulars. It's making the extra visibly an extra — a side garnish added on the plate, a compliment from the kitchen — rather than a bigger base portion that resets what "normal" looks like.
3. New hires learn from the plate, not the card
A new cook rarely learns a dish by weighing the recipe card. They learn by watching how the experienced cook plates it, and copy that — drift and all. Anyone who takes the busiest plate of the night as their reference then adds a fraction more on top: nobody wants to serve the smallest portion in the kitchen.
Staff turnover compounds this. Each new intake copies the previous one's plate, plus a small margin, the same way a message distorts as it passes from ear to ear in a game of telephone. A kitchen with a lot of turnover — which is most kitchens — never spontaneously recalibrates back to the card; it only drifts further away.
That makes onboarding the cheapest moment to stop it. A new hire who learns to plate with the card in hand from day one — by weighing, not by watching — starts at the right reference point instead of wherever the last intake happened to land.
4. One dish, measured three ways
Take your best-selling main course and measure it three times: what the recipe card says, what normal human variance still allows, and what an actual audit of five plates measures. Below is exactly that, with the cost per year attached.
The middle column isn't a mistake — a scoop that runs a few percent above the card is human variance, not a problem worth chasing. The red section on the third column is what sits above that, and that's where the money is.
Notice how small the gap looks on the plate, and how large it becomes on paper once you multiply it by a hundred and thirty portions a week, fifty-two weeks a year. That's exactly why nobody sees it at the table, and everybody feels it on the P&L.
Your signature main course, 130 covers a week. The red section is what sits above normal variance — where the money is.
The gap between the card and what you measure isn't all creep — part of it is the normal variance of a human hand. Only the piece above that variance is the real problem, and what that piece costs is shown above, in euros a year, on this one dish alone.
5. Why every stock count misses this leak
A stock count asks one question: does what was bought match what was used? With portion creep, the answer is always yes. The extra food wasn't lost, spoiled or stolen — it was simply eaten, by a satisfied guest, exactly the way it's supposed to go. The purchase-to-usage ratio looks perfectly healthy, while the cost per plate keeps quietly drifting away from what the menu assumes.
The same applies to your recipe costing. It computes a cost based on what the card says, not on what's actually on the plate — so as long as nobody physically weighs the plate, the costing stays perfectly correct on paper while it's long since stopped being correct in the kitchen.
That's the core of the problem: every instrument an owner normally uses to catch a leak — stock counts, cost calculations, supplier invoices — is built to catch theft and waste, and none of the three ever catches a portion that honestly grew too large.
6. Which dish is actually bleeding
Not every dish drifts equally, and not every drift carries the same weight. A side dish that runs five percent over means little if it only costs fifty cents. A main course that runs ten percent over, and sells a hundred and thirty times a week, is a completely different number.
Below are five dishes from the same menu, each with its own extra cost per year. The bars show immediately where the weight sits — and it's rarely spread evenly across the menu.
In nearly every kitchen that measures this for the first time, one dish carries most of the total drift — often more than half, while it's only one of five or six. That's good news: you don't need to fix five habits at once, you need to find and correct one.
Five dishes from the same menu, each with its own extra cost per year above the recipe card.
One dish carries more than half the total drift here, while it's only one of five. That's the usual pattern: portion creep rarely spreads evenly across a menu. Find the bleeding dish first — the rest can wait until next month.
7. How to catch it without policing every plate
The fix isn't a cook standing at the pass with a scale for every plate — that works for one service and never again. It's a periodic check: once a month, weigh five plates of your key dishes against the recipe card. Ten minutes of work, and it's enough to catch drift before it stacks up.
A photo of the correct plate at the pass works better than a written instruction — staff compare against what they see, not what they once read. Replace it every time your menu changes so it never ages into decoration.
And target training at new hires specifically, not the whole team. They're the fastest-drifting point in any kitchen, precisely because they don't yet have a reference point of their own — teach them to plate with the card in hand, not by watching a colleague, and the next intake won't drift any further than where you stand today.
Put your own five dishes next to it
Enter what your recipe card says, what you actually measured in a five-plate audit, and how many portions you sell a week. The five dishes come pre-filled with the numbers of a 50-seat bistro, so you can see how it reads straight away — overwrite them with your own.
Every dish gets its own drift against its own allowed variance. Below that sit three numbers: your weighted average drift, what portion creep costs you in total per year, and which dish carries most of it.
Portion creep scan
Five dishes, recipe card against measured cost, and the difference in euros a year.
—
The allowed variance per dish is a guideline for normal human plating — a hand-cut starter can swing more than a weighed main course. Everything runs in your own browser; nothing is sent or stored. This screen never suggests raising your menu price: portion creep is a portion problem, not a pricing problem.
Two things worth knowing while reading this. Allowed variance isn't an excuse to overserve — it's a boundary that stops you chasing every plate for nothing. A hand-cut starter naturally swings more than a weighed, portioned main course, and that isn't a problem worth fixing.
And this screen never suggests raising your menu price. Portion creep is a portion problem: the fix is a fill line and a recipe card in a new hire's hand, not a higher price the guest pays for something that was never agreed.
What to do with this this week, this month and this quarter
Stopping all drift at once is impossible. This order works, because every step makes the next one measurable.
This week — measure, don't judge yet
- Pick your three highest-volume dishes and weigh five plates of each against the recipe card.
- Record the weight or cost per plate, not your impression — a scoop that looks generous is sometimes exactly on spec.
- Calculate the percentage variance per dish and put it next to the numbers in the calculator above.
- Don't act yet: measure first, correct after, or you'll mistake a good service for a good habit.
This month — recalibrate what's bleeding
- Put a photo of the correctly plated dish at the pass for whichever dish shows the biggest gap.
- Put a fill line or a mark on the ladle or spoon used most often for that dish.
- Walk every new hire through the recipe card physically, scale in hand — not by having them watch a colleague.
- Repeat the five-plate check after two weeks on the same dish to see whether the recalibration held.
This quarter — make it permanent, never raise prices
- Fold the five-plate check into your regular monthly rhythm, alongside your other food cost checks.
- Refresh the photo at the pass every time your menu changes, so it never ages into decoration.
- Feed your real, measured costs back into your recipe costing, so your menu prices are based on reality instead of the card.
- Don't touch your menu prices to compensate — a portion problem gets fixed at the pass, never on the menu.
The plate is the only instrument that sees this
Portion creep is the leak that survives because every other control tool is looking somewhere else. Your stock count checks for theft. Your recipe costing checks the card. Your supplier invoices check the price. None of the three ever weighs the plate that actually leaves the pass — and that's exactly where the money is.
The good news is the fix takes no more discipline than ten minutes a month. Weigh five plates, put a photo at the pass, and hand every new hire the recipe card — that's the complete recipe, and in most kitchens it finds one dish that alone carries half the problem.
Once you know your real costs, feed them straight into your recipe costing and your menu engineering — a menu that prices what's actually on the plate is the only menu that predicts your margin honestly.