Purchasing & suppliers: from silent capital leak to your #1 EBITDA lever

1. Why purchasing sets your margin ceiling

Purchasing sets the real ceiling of your contribution margin because it locks in the raw-material cost before the kitchen touches anything. In an average restaurant food cost runs between 28% and 34% of sales; moving that number 3 points is the same as lifting EBITDA without selling one extra plate. I have seen it across dozens of operations audited by the Masterestaurant method: two sites with the same menu and the same ticket, one at 30% cost and the other at 36%, and the difference was never in the menu but in how they bought. Purchasing is not logistics; it is financial architecture. Every purchase order is a micro-decision on EBITDA. An owner who treats procurement as an operational errand is quietly giving away 4 to 6 points of yearly profitability without noticing it. A restaurant without a theoretical cost operates blind, and that is the first flaw the Masterestaurant method fixes.

2. The theoretical cost you measure against

Theoretical cost is what your plate SHOULD cost according to the standardized recipe and the current purchase price; real cost is what you actually paid. The gap between the two —usually 2 to 5 percentage points— is EBITDA evaporating in the cold room through waste, theft, careless portioning or buying above benchmark. Diego F. Parra insists on a hard principle: if you cannot measure the gap, you cannot manage it. Across the 8,400-plus units the method has audited, operators who closed that 4-point gap to under 1 recovered on average 40,000 to 70,000 USD per site each year. Theoretical cost is not an accounting exercise: it is the standard against which you audit every single invoice. The most expensive mistake in purchasing is negotiating on unit price while ignoring the insumo's total cost of ownership. A supplier who drops your protein 8% per kilo but delivers with 12% waste from a poor cut is costing you more, not less.

3. From unit price to total cost of ownership

Total cost of ownership adds four variables almost nobody measures alongside price: real processing waste, turnover (how much capital sits frozen in the cold room), the financing cost of that idle inventory, and the deviation against theoretical cost. In the Masterestaurant method we measure the real yield of every critical input: a tenderloin at 78% usable yield at 20 USD/kg costs 25.64 USD per usable kilo, while one at 65% at 18 USD/kg costs 27.69 USD. The invoice lies; the servable kilo tells the truth. Buying cheap at the source and expensive on the plate is the classic trap. The purchasing decision must leave the kitchen and connect to the management P&L, because every order moves the month's result. When the chef orders out of habit —"the usual"— nobody is asking whether that volume matches the sales forecast or whether the agreed price still sits below benchmark.

4. Every purchase order is an EBITDA decision

In the Masterestaurant method the purchase order closes against three data points: the week's sales forecast, current inventory level, and the target theoretical cost. A site billing 60,000 USD/month at 32% food cost buys around 19,200 USD in inputs; a mismanaged 5% deviation is 960 USD flowing straight out of EBITDA, every month, 11,520 a year. Multiply it across a chain of 10 units and you are talking about 115,000 USD a year decided in orders nobody audits. Buying without connecting to the P&L is signing checks blind. The supplier stops being an unquestioned ally and becomes a component audited against benchmark. Misplaced loyalty to the same old supplier is one of the quietest margin leaks: prices creep upward 1% here, 2% there, and in six months you are paying 8-10% over market without noticing. The Masterestaurant method requires rotating the quote of the 20 inputs that concentrate 80% of your spend (Pareto applied to the pantry) across at least three suppliers each quarter, and measuring each one's price variability.

5. Suppliers get audited, not worshipped

A supplier whose price swings 15% month to month injects noise impossible to budget; a stable one at 3% is worth more even if its base price is slightly higher. In the method's audits, formalizing this rotation cuts purchasing cost 3 to 7% in the first quarter. You measure the supplier; you do not thank him blindly. Costing must become predictive: you project the month's cost structure before it happens, you do not autopsy it afterward. Most owners review food cost on the 5th of the following month, when there is nothing left to fix; by then the waste already happened and the margin is already gone. The Masterestaurant method flips the logic: with the sales forecast, the standardized recipes and current purchase prices, you build the projected theoretical cost for the running month and compare it week by week against the real one. If in week 2 the gap exceeds 1.5 points, you act —renegotiate, switch supplier, adjust portion— while 15 selling days still lie ahead.

6. Predictive costing: project, don't autopsy

An operator who reacts in week 2 instead of at close recovers between 60% and 75% of that month's deviation. Budgeting the cost is cheap; autopsying it is expensive and arrives late. The first concrete move is to pull the theoretical cost of your 10 best-selling plates and contrast it with last month's real food cost: there you will see the exact gap you are giving away. The Masterestaurant method's hard rule is simple: no plate should exceed 32% food cost, and those that do get reformulated or pulled from the menu, not ignored. Next, isolate the 20 inputs that concentrate 80% of your spend and quote them against three suppliers this same week. Diego F. Parra sums it up after auditing more than 8,400 units in 43 countries: most restaurants do not have a sales problem, they have a purchasing problem disguised as a sales problem.

7. What an owner does Monday morning

Close the gap between theoretical and real cost from 4 points to 1, and in a site billing 60,000 USD/month you recover close to 21,600 USD a year. Start by measuring, not by negotiating. Focus moves from unit price to the total cost of ownership of the input: waste, turnover, immobilized capital and theoretical cost enter the same equation. The purchasing decision stops living in the kitchen and connects to the managerial P&L: every order is a micro-EBITDA decision, not a chore. The supplier goes from unquestionable ally to audited component: benchmarked and rotated when price variability erodes contribution margin. Costing becomes predictive: the month's cost structure is projected before it happens, not autopsied after.