Plate costing: the before and after of measuring every recipe

Step 1: build the recipe card with yield measured, not estimated

Before you calculate a single cost, weigh the raw ingredient and weigh what survives the trimming, because that gap is where the unaccounted money lives. Salmon coming in at 22 USD per kilo with 32% loss across skin, bone and trim actually costs 32,35 USD per usable kilo, and that 10,35 USD delta separates a menu that looks profitable on paper from one that is profitable at the bank. The deliverable is a card per dish with three columns — gross quantity, yield factor, net quantity — signed by the head chef. Verify it this way: take your best seller, have the cook prepare it and weigh the discards; if the real percentage differs from the paper by more than three points, the card is not usable yet and the measurement has to be repeated across two more runs. Theoretical cost is the sum of net ingredients from the card, and its mandatory companion is contribution margin: selling price minus cost, in money, dish by dish.

Step 2: calculate theoretical plate cost and contribution margin in DOLLARS

Take a real menu I reviewed this year. Lomo saltado costs 4,10 USD, sells at 14,90 and leaves 10,80 with a 27,5% food cost; mushroom risotto costs 3,20, sells at 9,50 and leaves 6,30 at 33,7%. Percentage arithmetic condemns the risotto. Now multiply by turnover: risotto sells 62 times a week and contributes 390 USD of margin, lomo sells 19 and contributes 205. Your deliverable is a table with cost, price, percentage, weekly units and weekly margin in dollars. If that last column does not exist, you are not costing yet. Reconciliation is the only proof that your recipe card describes the kitchen you actually have rather than the one you wish you had. It works with a simple subtraction: opening inventory plus purchases minus closing inventory gives real consumption, and that number faces the theoretical consumption obtained by multiplying each dish sold by its card.

Step 3: reconcile theoretical cost against real inventory every period

The gap is called food cost variance and it should stay under two percentage points. Past three, money is leaving through theft, over-portioning, waste or badly received purchases — in that order of frequency. Context helps here: according to The Restaurant HQ, an average restaurant wastes between 4% and 10% of the inventory it buys. Your deliverable is a monthly variance sheet by ingredient family, with the three largest deviations flagged and an owner assigned to each. Pricing is the last decision in the sequence and it depends on three inputs you already resolved: reconciled net cost, target margin in dollars and the observed elasticity of that dish in your market. Diego F. Parra insists on this order because reversing it is the most expensive mistake in the operation — a price built on a badly calculated cost ends up defended with discounts, and discounting eats EBITDA without warning.

Step 4: set the price AFTER the cost, never before or alongside it

There is a ceiling, of course: the National Restaurant Association recorded a restaurant price inflation peak of 8,8% in March 2023, the highest in over two decades, and the customer still carries the memory of that hit. The deliverable is a new price list with its effective date and, next to each dish, the dollar margin you have just decided to defend. Re-costing once a year made sense when prices moved once a year, and today it simply does not. The USDA recorded farm-level egg prices climbing 43,1% during 2024, and 80-90% ground beef going from 4,56 to 5,63 USD per pound by mid-2026 according to the same source. Under that volatility, a menu costed in January is stale by April. The fix that survives real operations is not re-costing everything: identify the twelve ingredients that concentrate roughly 70% of purchasing spend and review them on a fourteen-day cycle, leaving the rest of the pantry on quarterly review.

Step 5: re-cost every 14 days the twelve ingredients holding 70% of your spend

Your deliverable is a short board listing those twelve items, the last purchase price and the accumulated variation; when a line crosses 8%, the dish using it goes back to the pricing table. Almost every broken costing that reaches my desk fails on the same five things, and none of them is about software. First: loading payroll, rent and utilities onto the plate, when those expenses belong to break-even and not to the recipe. Second: costing with list price instead of the price actually paid on the last invoice. Third: ignoring yield loss, which as we saw can push an ingredient a third higher. Fourth: forgetting the invisible inputs — frying oil, salt, complimentary garnish, house sauce — that add between 20 and 40 cents per plate and vanish from the paper. Fifth: killing dishes by percentage without checking turnover. I got this wrong for years, defending percentage food cost as if it were the truth; percentage ranks the menu, but rent gets paid by dollar margin multiplied by units sold.

Closing checklist: how to know the costing came out right

You are done when you can answer six questions without opening a drawer. Does every menu item have a recipe card with a yield factor measured in the kitchen and signed off? Does the master table show contribution margin in dollars and weekly units sold, not just percentage? Did last period's variance between theoretical and real consumption land under two points? Do the twelve critical ingredients carry a review date within the last fourteen days? Does the live price list have a date and an owner? Does the costing exclude payroll, rent and utilities? Six yeses and the system is alive. Any single no, and the number you are staring at is an opinion formatted as a table. Start tomorrow with your highest-turnover dish: weigh it raw, weigh it trimmed and write down the difference — that alone will shift your pricing conversation. YIELD. The number one difference is not software or discipline, it is a scale.

Four differences that move cash

Salmon landing at 22 USD a kilo with 32% loss to skin, bone and trim actually costs 32.35 USD per usable kilo, and that 10-dollar delta is exactly what separates a menu that looks profitable on paper from one that is profitable at the bank. FREQUENCY. Recosting once a year made sense when prices moved once a year. Today, with fresh ingredients swinging by the quarter, a 14-day cycle across the twelve items holding 70% of food spend captures nearly all the variation for a fraction of the work; the rest of the pantry tolerates a quarterly look. UNIT OF MEASURE. I got this wrong for years: I defended percentage food cost as revealed truth, and helped retire plates that were holding the cash together. Percentage is a traffic light capped at 32%; the decision belongs to contribution margin in dollars times turnover, the only figure that reaches the EBITDA line.

Four differences that move cash — in practice

CLOSING THE LOOP. Costing without reconciling is writing a budget and never opening the bank statement. Theoretical cost says what should have been spent; inventory consumption says what was spent. The gap between them — theft, over-portioning, waste, purchasing error — is where lost profit lives, and it only shows up when somebody subtracts the two figures every fifteen days.