Prime cost: the mistakes draining your EBITDA and the right method to protect it in 2026

Chapter 1 — What is prime cost and why does it decide your margin?

Prime cost is the sum of ingredient cost —food and beverage— plus direct operating labor in kitchen and floor, and it is the only lever an operator controls in real time.

At Diego F. Parra we call it the block of truth: it represents between 55% and 68% of sales in an average restaurant across Latin America and Spain in 2026, and it is where 3% to 15% of net margin is won or lost. Rent, utilities and depreciation are fixed costs that never load onto the plate: they belong at the break-even point. Confusing both worlds is the first methodological error. A healthy prime cost runs below 62%; once it crosses 68%, the double-digit EBITDA vanishes and the operation starts surviving month to month on strained cash flow, even when the monthly P&L still looks correct on paper. The gap between real prime cost and what the operator believes he has runs 6 to 10 percentage points, and that invisible difference separates a profitable business from one that barely breathes.

Chapter 2 — The invisible 6 to 10 point gap

I have audited dozens of restaurants where the owner swears he is at 58% while the real number lives between 62% and 68% of sales. Those 6-10 points never surface in the monthly P&L because the average hides daily variance: a weekend with 5% waste is offset by two slow Tuesdays and the result looks stable. On sales of 100,000 USD a month, an 8-point gap is 8,000 USD evaporating with no accounting explanation. Multiplied by twelve months and three locations, the leak exceeds 288,000 USD a year. That is the difference between reinvesting and closing the doors. The structural error is not having a high food cost: it is measuring it at month-end against the sector average instead of daily against your own theoretical cost. The sector average is someone else's number and it never exposes your leak: a 31% food cost looks excellent, but if your theoretical cost per recipe was 27%, those 4 points of variance are waste, theft or overportioning that no benchmark would have caught.

Chapter 3 — Measuring at month-end is the structural error

The monthly close arrives thirty days late; by the time you see the number, the money already walked out the back door. The Masterestaurant method flips the logic: theoretical cost as baseline, daily close, and variance as an early signal. An operator who checks variance every 24 hours corrects in three days what the reactive model discovers in thirty and can no longer recover. The comparison that matters is not your food cost against the sector, but your real cost against your theoretical cost recipe by recipe, calculated from the standardized costing sheet. Theoretical cost is what that dish SHOULD cost with correct portions and purchase prices; real cost is what it actually cost according to inventory. The distance between the two is pure variance, and that is where 100% of the controllable leak lives. A restaurant with a 28% theoretical and 33% real cost has 5 points of variance: on food sales of 70,000 USD a month that is 3,500 USD monthly in waste, overportioning or theft.

Chapter 4 — Theoretical cost versus real cost: the only comparison that matters

Without theoretical cost there is no baseline, and without a baseline any food cost is an orphan number impossible to judge. The costing sheet, updated every time an input rises, is the heart of predictive control. Fixed costs —rent, utilities, depreciation, insurance— never load onto the plate because they distort food cost and lead to wrong pricing decisions. The classic error I see again and again: raising the price of a profitable dish to 'cover' a rent that does not belong to it, killing its turnover and losing more than it saves. Those fixed costs belong at the break-even point, not in unit costing. If your rent is 8,000 USD and your utilities 3,500 USD a month, those 11,500 USD are covered by aggregate contribution margin, not by loading 2 USD extra onto every burger. Food cost must reflect only ingredient and waste; direct labor enters prime cost, but administrative labor does not.

Chapter 5 — Why fixed costs never load onto the plate

Isolating each line in its correct box is what lets you price by real profitability instead of intuition. Moving from reactive control to predictive mitigation means no longer discovering the problem in cash flow but anticipating it with theoretical cost as baseline and daily variance as the signal. This white paper is aimed at owners, expansion directors and CFOs of restaurant groups running several locations who cannot afford to find out late. The reactive model detects the leak at 30 days, once you have already lost between 3,000 and 8,000 USD per location; the predictive one detects it in 24-72 hours and caps the damage below 500 USD. The annual difference, in a group of five locations, is around 200,000 USD of recovered EBITDA. At Masterestaurant we build the variance dashboard on three signals: theoretical cost per recipe, daily close of key inventory and labor deviation per shift.

Chapter 6 — From reactive control to predictive risk mitigation

With those three, risk stops being a surprise. The variance dashboard is built on three signals that, together, shield the cash: food cost variance by product family, labor deviation against sales per shift, and turnover of the ten inputs that concentrate 70% of spend. An operator watching only those ten references —Pareto applied to the pantry— controls the bulk of the leak without auditing the four hundred SKUs in inventory. Operating labor must move within a band: between 22% and 30% of sales depending on format; if a shift crosses 32% two days running, it signals overstaffing or falling sales. I have seen groups cut their prime cost from 66% to 61% in one quarter just by installing this dashboard, without firing anyone or lowering quality: five points that on 1.2 million in annual sales are 60,000 USD straight to the bottom line. The core difference is not the food cost figure, but the FREQUENCY and the BASELINE of the measurement.

Chapter 7 — The differences that decide the margin

The traditional approach compares the monthly result against a foreign industry average; the Masterestaurant method compares the daily close against each recipe's own theoretical cost. A 31% food cost can be excellent or a silent leak depending on its theoretical cost: if the theoretical was 27%, those 4 variance points are waste, theft or spoilage that no industry average would have exposed. The second difference is the accounting treatment of fixed costs. Loading rent, utilities and depreciation onto the plate distorts food cost and leads to wrong pricing decisions: you raise the price of a profitable dish to 'cover' a fixed cost that does not belong to it. The method isolates prime cost as the variable block and sends fixed costs to the break-even point, where you truly decide how many covers are needed not to lose. The third difference is the nature of the signal: reactive versus predictive. When the leak appears in cash flow, the capital is already gone. Daily variance is an early signal that lets you mitigate the risk before it materializes in EBITDA.