The price hike that erases your profit: catch it in 72 hours, before it eats the month

Step 1: freeze a base basket of 20 inputs and date it

Before chasing any price increase you need a SNAPSHOT, and that snapshot is twenty inputs with their unit price and the exact date you bought them. Pick the twenty that show up most often in your weekly invoicing, not the twenty most expensive ones, because the damage comes from rotation and not from the list price: an input that arrives three times a week multiplies any movement across twelve purchases a month. The deliverable is a sheet with five columns —input, unit, price, supplier, date— and it gets verified one way only: you pull a random invoice from last week and the numbers match to the cent. If you have to call the supplier to fill a cell, that line is not ready. With an operating margin of 5% to 6%, this sheet is worth more than your month-end accounting report. Costing at close is an autopsy; costing against Thursday's invoice is medicine, and the gap is measured in money.

Step 2: cost against the invoice, not against the month-end close

A venue billing 45,000 dollars a month with 30% food cost processes roughly 13,500 dollars of purchasing monthly, so forty-five days of blindness amount to some 20,000 dollars bought at a price you never checked. The procedure is simple and boring, which is exactly why it works: when goods arrive, somebody captures the unit price in the step-one sheet before the box goes into the walk-in. Seventy-two hours of reaction, not six weeks. The measurable deliverable is the last-updated date on each line; if any of them passes seven days, control broke and you went back to doing autopsies. Without a number that triggers action, cost control turns into after-dinner opinion. We use 1.5 percentage points of variance over the theoretical plate cost as the line that forces intervention that same week, and below that threshold nothing gets touched, because chasing 0.4% movements burns your team's attention on noise.

Step 3: set the variance threshold at 1.5 points and automate the alarm

The reason for setting it that low comes from the industry's arithmetic: with net profit hovering around 3% to 5% according to the National Restaurant Association, a point and a half of food cost eats a third of everything you keep. The deliverable is a formula cell that paints red any plate outside the range. You verify it by manually entering a price 10% higher and confirming that the alarm fires. Global food cost lies by construction, because it averages the plate that leaves you a 68% margin with the one that leaves 12% and hands back a soothing 31% that describes nothing real. Sort your menu by contribution margin in absolute dollars per unit sold, not by kitchen category, and you will see that eight to twelve lines hold up the register while the rest travel as stowaways. A warning belongs here: percentage margin and dollar margin rarely agree, and when they clash always follow the absolute dollar, because you pay rent with dollars and not with percentages.

Step 4: move from global food cost down to contribution margin per plate

The deliverable is a table sorted from highest to lowest contribution. Verify it by adding up the top ten lines: if they don't explain more than half your gross profit, recalculate the recipes. An increase detected without a written decision within the next 72 hours costs you the whole month again. You have four exits, and they're worth ranking by damage to the guest: renegotiate volume with the supplier, change the input specification, adjust portion weight inside a range the customer won't notice, and only at the end raise the selling price. In the United States 90% of full-service operators raised prices during 2024 and 60% also pulled dishes from the menu, according to the National Restaurant Association, which tells you the menu is a legitimate lever but a crowded one. Diego F. Parra insists at Masterestaurant on exhausting the first three before touching the fourth.

Step 5: turn the increase into a decision before moving the selling price

The deliverable is one dated line per increase: input, percentage, decision made, owner. The costliest mistake isn't botching a calculation, it's handing price capture to whoever receives the goods without giving that person two minutes to do it. Second: watching only the kilo of protein, which swings hard and gets noticed, while oil, onion and packaging climb between 5% and 9% in silence and settle into a plate you sell a hundred times a week. Third: ignoring waste, when foodservice accounted for 17.9% of the United States food surplus in 2024 and full-service restaurants contributed more than 43% of that surplus, according to ReFED. Fourth: costing the plate by loading payroll and rent onto it, an accounting vice that inflates the recipe and hides the real problem. Payroll and utilities belong to the break-even point, never to the plate. You'll know everything landed well when you can answer five questions without opening an email or calling anyone.

Closing checklist: how to know the system is actually running

One: which input was the last to rise more than 5% and what day did you find out? Two: does any line of your basket sit more than seven days without an update? Three: how many plates are outside the 1.5-point threshold today? Four: what written decision came out of the last variance and who signed it? Five: are your ten highest-margin lines still the same ones as last month? If a single one fails, the system isn't built, it's drawn. The definitive test is a calendar test: next Tuesday at nine in the morning open the step-one sheet and check that somebody other than you touched it the week before. That's the point where costing stops depending on the owner. FREQUENCY. Costing at month-end is an autopsy; costing against the invoice is medicine. Between a 45-day and a 72-hour reaction sits, in a venue billing 45,000 USD a month at 30% food cost, roughly 20,000 dollars of purchasing processed at a cost you did not know had changed.

The four differences that actually move cash

GRANULARITY. Global food cost lies by construction, averaging the dish that returns 68% margin with the one returning 12%. Deloitte and the wider industry have pushed item-level contribution margin for a decade, and most menus are still organized by kitchen category rather than by what each line contributes to the register. THRESHOLD. Without a number that triggers action, restaurant expense control turns into opinion. We use 1.5 points of deviation against the dish's target food cost: below that you note and observe, above it you redesign the card, renegotiate the ingredient or move the price. One written criterion, no debate in the kitchen. ORDER OF LEVERS. Recipe card and real grammage first, supplier negotiation second, menu engineering third, and menu price last. Reversing that order is the most expensive mistake I keep running into: owners raise prices on a cost that was never measured properly, and the guest pays for an internal inefficiency that could have been fixed for free.