Target Food Cost by Restaurant Type: −6.1 Points of Theoretical-vs-Actual Gap with the Standard Recipe Generator

The diagnosis: healthy sales, tight cash

The problem wasn't the food cost, but the absence of the right target for this model. The case file: a family trattoria with 14 tables (56 covers), 9 employees, a $19 USD average check, 11 years operating in a mid-sized city in central Mexico, with 78% of sales in the dining room. The owner, a cook by trade, filled the house on weekends, yet real EBITDA hovered around just 3% (per the case diagnosis), below the 2.8% pretax profit that full-service median reports show according to the National Restaurant Association (2025). In a sector Mexico values at 300 billion pesos according to CANIRAC (2024), holding 12.2% of the country's businesses according to CANIRAC/INEGI (2024), strong revenue isn't enough: without a defined residual, the cash drains on its own. There is no universal food cost target: it depends on the restaurant type. Food cost isn't a moral number, it's the residual left after paying labor, rent and commissions while leaving healthy EBITDA; each model leaves a different residual.

Is there a universal food cost target?

A white-tablecloth trattoria with wine sales tolerates 30–34% because its contribution margin per cover is high; a fast-food dark kitchen needs 26–29% to absorb delivery commissions that reach 25%.

That gap by model shows in reported profitability: 4.0% pretax profit in limited service versus 2.8% in full service according to the National Restaurant Association (2025). For this $19-check trattoria, Diego F. Parra of Masterestaurant set the target at 31%, not the textbook 28% that would have strangled the culinary proposition. The lost capital lived in the gap between what the recipe said it cost and what the cash register actually paid. Recosting with the week's supplier prices, the menu's theoretical cost came out at 30.5%; the real cost measured against purchases and inventory reached 37.2% (per the case audit), nearly 7 points of leakage. In an operation with $58,000 USD in monthly sales, those 6.7 points equaled $3,886 USD vanishing every month with no visible invoice.

The invisible leak: theoretical vs real cost

Waste explains part: foodservice surplus food represented 14% of sales according to ReFED (2024). The rest was unrecorded shrinkage, portions without weights and uncontrolled purchasing. Nobody was stealing; the system simply wasn't measuring. That's the leak traditional accounting never shows, because it averages the month into a single number. The intervention attacked measurement first and margin second, in that order. Using the Masterestaurant menu-engineering tool, all 34 menu items were recosted plate by plate, setting weights and technical sheets with a per-plate food cost target of ≤32%. A weekly inventory count was installed, along with purchasing against projected menu rather than habit. The six plates with real food cost above 40% were redesigned: portions were adjusted, two suppliers renegotiated, and three menu stars raised in price by 8% without resistance, because their contribution margin justified it. The cost context was tight: the services producer price index rose 3.2% according to the U.S.

The action with the Masterestaurant method

BLS (2025). The goal wasn't to cut quality, it was to close the gap between theory and cash. Real food cost dropped from 37.2% to 31.4% in 90 days and EBITDA rose from 3% to 9.6% (per the case follow-up). In money, the operation recovered roughly $3,400 USD a month from the leak that used to evaporate, enough capital to reinvest without a loan, no small detail when regional variation in SBA loan default rates for restaurants reaches 8.7 percentage points according to Crestmont Capital (2026). The average check held at $19 USD because the increases were surgical, not across the board. The cash lesson: the right target wasn't the lowest possible food cost, but the one that left a healthy residual for this model. Chasing the textbook 28% would have degraded the food and driven off the tablecloth diner sustaining 78% of sales.

Transferable lessons by operation size

The recommendation changes with size, but the principle is identical: set your target by your residual, not your neighbor's. Small independent (1 location, under 20 tables): this week, recost your five best-selling plates by hand with today's supplier prices and compare against your real cost from the last purchase; that's where you'll see your leak. Mid-size (2–4 locations): implement a weekly inventory count and a technical sheet per plate with fixed weights before touching prices. Multi-unit group: standardize the food cost target by format —not a single one for the whole brand— because a delivery location with 25% commission needs 3–4 points lower than the dining-room one. In all three cases, the first step isn't cutting: it's measuring the theoretical vs real gap, which is where the money lives. This result isn't replicable in any context, and confusing that would be survivorship bias.

Limits of this case

First, it worked because the trattoria had a high contribution margin from wine sales and a $19 check; a set-lunch business with a $6 check and no margin beverages lacks that cushion, and raising prices there does scare off the clientele. Second, the operation was mostly dining room (78%); a pure dark kitchen with dominant delivery faces 25% commissions that completely change the arithmetic and demand a 26–29% target, not 31%. Third, there were 11 years of history and stable recipes; in an opening with no data or technical sheets, you stabilize the operation first and only then optimize food cost. The right target always depends on the model, the channel and the check, never on a textbook number. Target food cost isn't a moral number: it's the residue left after paying labor, rent and commissions to leave a healthy EBITDA. Each business model has a different residue.

Why target food cost changes with restaurant type?

A trattoria with a $19 check and wine sales tolerates 30–34% food cost because its contribution margin per cover is high; a dark kitchen with 25% delivery commission must drop to 26–29% or its break-even becomes unreachable.

The mistake of comparing your food cost to your neighbor's ignores that labor, CapEx and channel mix are different. The useful benchmark is your own theoretical vs actual cost, not the restaurant next door. In limited service, the National Restaurant Association (2025) reports 4.0% pre-tax income versus 2.8% in full service: the lower target food cost of the fast format is what sustains that profitability gap.