Restaurant unit economics: before and after you know what each dish leaves behind

What is restaurant unit economics?

Unit economics is the margin left by ONE unit sold —a dish, a check, an order— after subtracting only its variable costs, and you calculate it by taking the selling price and removing food cost, packaging, platform commission and the direct service input.

An owner in Bogotá showed me a P&L with 14% operating profit and a bank account that refused to grow, and the contradiction did not live in the income statement: his three best sellers, which carried 41% of all orders, returned $2.30 of contribution margin while the rest of the menu sat near $8. He sold more and earned less. Accounting tells you what ALREADY happened, aggregated and late; this metric tells you what happens every time the kitchen bell rings, which is why I put it ahead of any other financial conversation with an operator. Before you calculate anything, draw the line: a dish only carries what DISAPPEARS when the dish is not sold.

Step 1: separate variable costs from structural costs

Recipe inputs, packaging, channel commission, side, sauce and attributable waste belong there; payroll, rent, utilities and software do NOT, because you pay them even if the kitchen sells zero. Almost nobody respects that boundary, and blending it produces an inflated plate cost that pushes price hikes into a market that cannot absorb them. The structural benchmarks help you place yourself: full-service payroll runs at a median 36.5% of sales (CostLab.AI 2025) and sector labor cost moves between 25% and 35% of revenue (U.S. Bureau of Labor Statistics). That money comes back at the restaurant's break-even point, never inside a single dish's cell. Deliverable for this step: two closed columns, variable and structural, with not one line item repeated in both. Contribution margin in money beats percentage, and half of a menu's profitability gets decided right here. A dish at 22% food cost returning $3.10 loses against one at 30% returning $9.60, because Saturday's register fills up with dollars and not with percentage points.

Step 2: calculate contribution margin in dollars, not in percentage

Percentages are useful to police purchasing and supplier negotiation; dollars are what tell you which item the server pushes and which one owns the best real estate on the menu. Take the sector reference carefully: average full-service food cost closed 2025 at 32.4% of sales (VantaInsights 2026), and inside our own contract 32% per dish is the CEILING, never the target. Deliverable: a table with your 20 best-selling items ranked by dollar margin, highest to lowest, with price and variable cost sitting in plain view. Weigh the portion three times during service, on a Friday at nine at night, and use that number —not the recipe card— for your costing. The gap between theoretical grammage and what leaves the line at peak usually runs between 8% and 15% on proteins and sauces, and that gap eats the margin without ever showing up in a report.

Step 3: cost the REAL portion, not the one on the spec sheet

I got this wrong for years: I costed with the recipe and audited the purchase, while the hole sat in the middle, in the hand of a cook who serves generously because nobody told him what his generosity is worth. Add attributable trim waste, real product yield and the courtesy portion —the bread, the extra sauce, the refill— which almost never gets costed. Verifiable deliverable: a corrected spec sheet with measured grammage, signed by the chef, and a plate cost that does not shift more than 3% when you repeat the measurement a week later. The same dish carries three different unit economics depending on where it goes out, and treating them as one is the most common way to lose money while believing you are growing. In the dining room you charge full price; on your own delivery you add packaging, bag and driver; on a marketplace you also add a commission that across the region moves between 18% and 30% of the check.

Step 4: cost every channel separately, because they are not the same business

With 70% of U.S. diners having ordered delivery in the past month (Escoffier, 2025 Consumer Dining Trends), the channel is not optional, though selling it at a loss certainly is. Run the number: a $12 dish with $4 of variable cost leaves $8 in the dining room and $3.40 on a platform at 30% with $1 of packaging. Deliverable: a dish × channel matrix with the dollar margin in every cell, plus a written decision on which items do NOT go to platforms. A unit margin without mix is an orphan number: what defines your cash is how many times each dish sells, not how much your best one returns. Multiply each item's dollar margin by its units sold last month and add it up; that total is your gross contribution, and structure gets paid out of it. The Bogotá case resolved there: he reordered the menu, raised the three anchor dishes by $1.80, moved two $8 items to the center of the layout, and within eleven weeks monthly contribution climbed without a single extra order.

Step 5: multiply by the real mix and your business appears

Compare the result against a healthy prime cost, which sits around 55%-65% of sales with a target near 60% (Restaurant365); if gross contribution does not cover structure comfortably, the problem is mix or price, and rarely volume. Deliverable: monthly gross contribution calculated and set against the location's fixed expenses. Mistake number one is prorating payroll and rent inside the dish, which inflates unit cost and triggers suicidal price hikes in price-sensitive markets. The second is costing from the supplier's price list instead of the last invoice actually paid, when protein moved 6% the previous month. The third is forgetting VAT or consumption tax inside the selling price, which makes the margin look 8-10 points fatter than it is. And the fourth, the most expensive of them all: scaling before measuring.

The four mistakes that wrecked this calculation most often

With the independent sector contracting 2.3% in 2025 and a net loss close to 9,500 locations (Technomic via Nation's Restaurant News), opening a second site —median cost around $275,000, roughly $3,046 per seat in a leased space, per RestaurantOwner.com— on negative unit economics multiplies the loss by the number of branches. Deliverable: the four mistakes reviewed one by one against your own file, dated. You are finished when you can answer, without opening any software, how much your number one dish returns. Verify these six points before you call the work closed: every item has a variable cost built on measured grammage rather than theory; no structural line appears inside a dish; the margin table is ranked in dollars and not in percentage; a dish × channel matrix exists with delivery and marketplace costed apart; monthly gross contribution is calculated against real fixed costs; and at least one decision has been executed —price, portion, recipe or menu removal— on the three worst-margin items.

Closing checklist: how to know the job is done right

Diego F. Parra built the Masterestaurant framework on that sequence, and the sequence matters: first the unit, then the model, and only then growth. Schedule the review every 90 days, or whenever a key input moves more than 5%. FIRST: unit economics separates variable costs from structural costs, and that border is the one almost nobody respects. The plate carries only what disappears when the plate is not sold: ingredients, packaging, channel commission, garnish and attributable waste. Payroll, rent and utilities do NOT belong on the plate, because you pay them even at zero sales. Mixing them produces an inflated per-dish cost that pushes prices up in a market that will not carry them, and I have watched kitchens close over that arithmetic. SECOND: contribution margin in dollars outranks the percentage. A dish at 22% food cost leaving $3.10 loses against one at 30% leaving $9.60, and Saturday's till fills with dollars rather than percentage points.

Five differences that move cash rather than talk

Percentages police purchasing; dollars design the menu. THIRD: with no measured unit, scaling amplifies the error. If your star dish loses $0.80 per unit, the second location doubles that bleed with all the solemnity of an expansion plan. Correct order: measure, fix, then open. FOURTH: unit economics translates your value proposition into numbers. You may claim you sell product-driven cooking and warm hospitality; the unit margin says whether the market will pay for it or whether you are subsidizing that promise out of your own pocket. FIFTH: it rewrites the capital conversation. A restaurant investor discards revenue-projection pitches in the first meeting, because billing is easy; what gets evaluated is unit margin, speed of recovery and how sensitive that margin is to an 8% protein increase.