Prime cost: the 6 mistakes draining your EBITDA and the method that stops the leak

Chapter 1 — What prime cost really is, and why it isn't food cost plus payroll

Prime cost is the sum of food and beverage cost plus total labor cost —including benefits and payroll taxes— measured as a percentage of net sales. It is not a figure you close at month-end: it is a weekly indicator of theoretical versus actual cost. The rest of the structure (rent, insurance, depreciation) is sunk cost you can no longer move this week; prime cost you can. At Masterestaurant we set the healthy threshold between 55% and 60% of sales. Each point above that in a restaurant with USD 1.5M in annual revenue is USD 15,000 leaving EBITDA with no obvious accounting trace. The error I see again and again: owners who add the two P&L lines at month-end and think they measured. That isn't measuring, it's performing the autopsy. Chasing the prime cost average hides the volatility that bleeds you dry. A 61% average sounds healthy, but if it swings between 56% in slow weeks and 68% at the peaks, the operator lives on borrowed treasury: good weeks subsidize bad ones and cash flow never breathes.

Chapter 2 — Mistake 1: chasing the average instead of the variance

The number worth chasing is variance —the gap between theoretical and actual cost as a percentage of sales—, not the mean. Across hundreds of managerial P&L audits I have seen the same trap: a prime cost 'at 60%' that actually runs at 66-68% during peak-volume shifts, exactly when you sell most and lose most. That 6-to-8-point gap swallows all projected EBITDA. Variance sustained above 2 points isn't bad luck: it's a broken measuring instrument that must be repaired before touching recipes. Measuring prime cost monthly instead of weekly turns a fixable leak into a compounding hemorrhage. If you find on day 45 that labor cost jumped three points, you already lost six weeks of margin that never comes back. Prime cost is the only number an operator can move in under seven days: you adjust shift scheduling, renegotiate with two suppliers, change three high-volume recipes.

Chapter 3 — Mistake 2: measuring late turns a leak into a compounding hemorrhage

Rent, insurance and depreciation don't respond at that speed. That is why at Masterestaurant inventory counts and theoretical payroll close every Monday, with the weekend shift still fresh. A restaurant that measures monthly loses, on average, between 4 and 9 EBITDA points a year versus one that measures weekly. That gap, in a USD 2M operation, is between USD 80,000 and USD 180,000 a year of pure decision. Failing to calculate theoretical cost leaves the operator blind against actual cost. Theoretical cost is what the week's sales SHOULD have cost to produce according to standardized recipes and scheduled labor; actual is what genuinely left cash and inventory. The subtraction between them —variance— is the diagnosis. Without theoretical you only have a final number with no cause: you know you lost, not why. In practice, a 3% food cost variance usually breaks down into unrecorded waste, unstandardized portions and pilferage; a 2% labor variance is almost always over-scheduling in slow hours.

Chapter 4 — Mistake 3: not separating theoretical from actual leaves the problem undiagnosed

I have audited restaurants convinced they had a purchase-price problem when they were actually giving away 4 points at plating. Without the theoretical, that owner would have renegotiated with suppliers and kept losing exactly the same. Putting rent, utilities or depreciation inside prime cost breaks the tool. Those costs are fixed and sunk for the month: they belong to the break-even point, not to the live indicator you correct week by week. When an owner 'dilutes' rent into per-plate food cost, they get an inflated number that says nothing actionable: they can't lower rent on Tuesday. Prime cost must contain EXCLUSIVELY what responds to seven-day management: inputs and labor. At Masterestaurant we rigorously separate prime cost (the lever) from occupancy cost (break-even). A common attribution error: loading the general manager's salary into variable labor when it is fixed. That dirties the variance and sends you chasing ghosts.

Chapter 5 — Mistake 4: loading sunk costs into prime cost distorts the lever

The hard rule: if the cost doesn't change when the week's sales change, it doesn't belong in prime cost. A consolidated prime cost hides which time slot destroys margin. A restaurant can close the week at 60% overall and still lose money: lunch runs at 52% and dinner at 71% from over-staffing and an unbalanced menu. The average lies by design. At Masterestaurant we break prime cost down by daypart and by channel —dine-in, delivery, takeout— because each has a different labor structure: delivery can carry food cost 8 points higher from packaging and a platform commission of 18-30%. I have seen operators raise prices across the whole menu when the problem lived in two dinner-shift dishes. Segmenting prime cost by slot turns an anesthetic number into a map of where to cut. Without that granularity, every decision is blind and the opportunity cost is brutal.

Chapter 6 — The Masterestaurant method: from accounting figure to decision dashboard

The Masterestaurant method turns prime cost into an anticipatory decision dashboard, not an accounting autopsy. The difference is operational: the traditional approach tells you why you lost last month; this one lets you correct this week. The dashboard runs three numbers every Monday: theoretical prime cost, actual and variance, broken down by daypart and channel. On top of that we build a 90-day roadmap with KPIs at 3, 6 and 12 months, designed to take to the board: cut food cost variance below 1.5 points, close over-scheduling in slow hours and stabilize consolidated prime cost between 55% and 58%. In operations we have advised, recovering 3 to 5 prime cost points in a quarter is realistic and translates, in a USD 1.5M restaurant, into USD 45,000 to 75,000 of EBITDA recovered per year. Start by measuring theoretical against actual this Monday. The traditional approach treats prime cost as historical accounting data; the Masterestaurant method treats it as an instrument of anticipatory management.

Chapter 7 — The differences that decide the margin

The difference is not semantic: one tells you why you lost last month, the other lets you fix this week. In a restaurant with USD 1.5M in annual sales, each mis-controlled point of prime cost is USD 15,000 a year leaving EBITDA with no obvious accounting trace. Chasing only the average hides volatility. A 61% average prime cost sounds healthy, but if it swings between 56% and 68% depending on the week, the operator lives on borrowed treasury: good weeks subsidize bad ones and cash flow never breathes. Variance —the deviation between theoretical and actual cost as a share of sales— is the real KPI. Chasing the average is like measuring body temperature while ignoring intermittent fever. The third axis is the decision unit. Global food cost averages profitable dishes with bleeding ones; per-dish contribution margin exposes which product funds the operation and which drains it. Menu engineering fixes that: reposition, redesign the portion, or retire it. Without that granularity, lowering global food cost cuts exactly the star dishes and cheapens perception, destroying average check while the dashboard number improves falsely.