How to Price a Menu When Food Costs Keep Moving

Your chicken supplier sends a new price list. Oil is up again. The produce invoice is eleven dollars higher than last week and nobody can tell you why. Meanwhile your menu has said $18.95 since the spring, and every time you think about changing it you picture the regular at table four doing the math out loud.

Most menu pricing advice hands you a multiplier — food cost times three, or a 30% target — and leaves you there. That is not pricing. That is a guess with a decimal point. The multiplier only works if you already know what the plate actually costs you, and most kitchens do not, because the number on the invoice is not the number on the plate.

Here is the mechanism. Learn it once and you can re-price anything, in any month, without waiting for someone to tell you what the new multiplier should be.

Cost the plate, not the invoice

Start with one dish. Write down every component that leaves the walk-in and ends up in front of the guest. Protein, starch, vegetable, sauce, garnish, the oil it was fried in, the butter finish, the bread on the side, the ramekin of dip nobody asked for but everybody expects.

Now price each one at your usable cost, not your purchase cost. This is where most sheets go wrong.

A case of romaine costs what it costs, but you throw away the outer leaves and the core. If you lose a quarter of the weight, your usable cost per kilo is a third higher than the invoice says. Same with a whole fish, a bone-in leg, a case of tomatoes where two are always soft. Trim loss, peel loss, spoilage, the portion that burns — all of it belongs in the cost of the dishes that survive.

The simplest way to handle this without building a spreadsheet you will never open again: take a case, prep it the way you actually prep it, weigh what you end up with, and divide. Do this once per major ingredient. You will not need to redo it unless your prep changes. The yield percentage stays the same even when the price moves — so when the invoice goes up, you multiply the new price by the same factor and you are done in ten seconds.

Then add prep labour. Not your line cook's whole shift — just the labour that exists because this dish exists. If somebody spends two hours a week braising short ribs, that is a real cost of the short rib dish and of nothing else. Wing sauce made in house, dough proofed overnight, stock simmered for six hours, pickles jarred on Sunday: all of it is labour attached to specific plates. House-made almost always beats bought-in on ingredient cost and loses some of that back in hours. You want to know which side of the line each item falls on.

What you end up with is a plate cost that is usually 15% to 30% higher than the naive add-up-the-invoice number. That gap is exactly why kitchens that "run 30% food cost" somehow have no money at the end of the month.

Decide what the plate has to earn

Now flip the direction. Instead of asking "what should I multiply by," ask "what does this plate need to contribute."

Every plate you sell has to cover its own cost and then hand you something toward rent, wages, gas, insurance, the POS fee, the accountant, and eventually you. That something is contribution margin — selling price minus plate cost, in dollars, not percent.

Dollars matter more than percent, and this is the single most useful shift in how you think about the menu. A $9 salad with a $2 plate cost runs a beautiful 22% food cost and hands you $7. A $34 steak with a $13 plate cost runs an ugly-sounding 38% and hands you $21. If your dining room turns a fixed number of covers a night, you would rather sell the steak. Percentage targets quietly push you toward selling cheap food to people who came in hungry.

So set a target: what does an average main have to contribute, at your average number of covers, to cover fixed costs and leave a profit? Work out your monthly fixed costs, divide by the number of entrées you sell in a month, and you have a floor. Anything below it is being subsidized by something else on the card. That is allowed — a loss-leader wing night that fills seats on a Tuesday is a strategy — but it should be a decision, not a surprise.

Price backward from the margin, then round

Plate cost plus required contribution gives you a number like $22.40. Now you make it a menu price.

Round up, not down, and round to something that reads cleanly. $23. Not $22.95 unless your whole card is .95 — mismatched endings make a menu look assembled rather than designed. Charm pricing signals value and cheapness at the same time; whole numbers signal confidence. Pick one convention and hold it across every line.

Then sanity-check against the room, not the street. Does $23 fit what someone sitting in your dining room, looking at your chairs and your lighting and your service, expects to pay? That is a real constraint. But it is a constraint about your room.

Do not price off the place down the street

The restaurant two doors down charges $19 for a similar dish. You know this because you looked. It is the most tempting data point on earth and it is close to useless.

You do not know their rent, which might be half yours on a lease signed in 2016. You do not know whether the owner's family works the line for free. You do not know if they buy frozen and you buy fresh, or if that dish is a loss-leader propping up a bar program you cannot see from the sidewalk. You do not know if they are quietly going under. Restaurants copy each other's prices into a slow collective decline, everybody anchoring to somebody who is losing money.

Use the street as a reality check on the category — if every comparable room is at $19 and your math says $31, something in your math or your model needs another look. Do not use it as a price.

Change the menu, not the price tag

The psychology matters as much as the math. Guests do not track absolute prices well. They track changes. A regular who orders the same thing every Thursday will notice that one dish went from $18.95 to $20.95 and will feel something about it.

So do not creep. Do not raise two dishes in March and three in June and four in September. Every one of those is a separate small betrayal, and together they teach people that your prices are unstable.

Move the whole card at once, on a schedule — twice a year is plenty for most rooms. And when you do it, reprint. New paper, adjusted descriptions, one or two dishes dropped, one or two added, maybe a different order on the page. When the menu is visibly new, the prices are new too, and nobody is comparing line by line against a memory. Same reason retailers change packaging.

Between reprints, absorb the movement — that is what your margin cushion is for. If an ingredient genuinely spikes beyond what the cushion can hold, change the dish instead of the price. Smaller portion of the expensive thing, more of the cheap thing, a different cut, a seasonal swap. Guests forgive a recipe that evolves far more readily than a number that climbs.

One more place the number moves without your permission: delivery apps. If your app menu mirrors your dine-in prices, the commission comes straight out of the contribution you just calculated. Price those channels separately, and read why delivery app orders cost more than you think before you decide what to list.

Check the card twice a year, not the spreadsheet every week

Build the plate costs once. Update ingredient prices when a big one moves. Sit down with the whole menu twice a year, recalculate contribution on every line, and reprint.

While you are in there, look at what sells. A dish with great margin that nobody orders is not helping you, and a dish that sells constantly at a thin margin is quietly eating your labour. Your POS or your online ordering system can tell you item-level counts — pull them before you re-price, so you know which lines are load-bearing and which are just taking up space on the page. Move the good-margin sellers to the top of their section. Cut the bottom three.

That is the whole discipline. Know the real cost. Know what each plate has to earn. Move everything at once, on your schedule, with a menu that looks new. Nothing about it requires a consultant or a subscription — just an honest yield test and an afternoon with a calculator, twice a year, before the invoices decide your prices for you.