Waste control: before vs after with the Masterestaurant method

Step 1: close a physical count at both ends of the period

Waste control starts by closing a physical inventory on the same weekday, at the same hour, before the first purchase of the next period comes through the door: without those two snapshots you don't have food cost, you have a purchasing rhythm. The formula is opening inventory plus purchases minus closing inventory, divided by net sales, and missing either end is enough to make the number lie by three or four full points. The DELIVERABLE is a sheet with the 40 or 50 SKUs that carry 80% of your spend, each with quantity, unit and unit cost from the latest invoice, signed by whoever counted and whoever supervised. Verify it this way: two people count the protein walk-in separately and their figures must land within 2% of each other. If they don't, waste isn't your problem yet; you simply cannot count your own storeroom, and that problem comes first and costs less to fix.

Step 2: recost every recipe card on servable weight, not purchase weight

A 6-kilo beef primal yields between 4.2 and 5.1 kilos of usable portion depending on the cut and the hand working it, so costing on invoice weight bakes a 15 to 25% undercost into every plate from the day the recipe was written. I got this wrong for years: I treated yield factor as textbook refinement, until the numbers at a grill house showed its three signature dishes were selling below cost. The work is to weigh the whole piece, break it down, weigh the trimmed product and divide; that quotient is your FACTOR and it multiplies purchase cost. It's done when every menu item carries a card with gram weight, yield factor and a dated unit cost. Verify by picking three dishes at random and rebuilding the portion on a scale: if actual grams beat the card by more than 5%, the card governs nothing.

Step 3: calculate the gap between theoretical and actual cost, in money

Theoretical cost comes from multiplying units sold of each dish by its recipe-card cost, and that sum gets compared against the actual cost your step-one inventory produced; the difference between the two, in money and in percentage points of sales, IS your waste. An average restaurant loses between 4 and 8 points of food cost inside that gap. With the 92-seat operation in Bogotá the number came out naked: they bought protein worth 38 million pesos a month while the card said 29, a gap of 9 million monthly, 108 million a year. Nobody was stealing. The deliverable here is one monthly line with three figures, theoretical, actual and gap, and you verify it by checking whether the gap moves when you change something. Fix portions and see no reaction across two periods, and your measurement is built wrong. Every food cost gap distributes across four buckets, and they deserve to be measured in this order: portioning without a scale, spoilage from cold chain or rotation, process waste in butchery and prep, and purchase prices that drifted away from the card.

Step 4: split the gap into four buckets and attack the biggest one

In Bogotá those 9 million split among eyeballed portions, a refrigeration rack running at 9 °C instead of 4, and three recipes written in 2023 with a supplier who had raised prices twice since. The FAO estimates roughly 14% of food is lost between harvest and retail, which confirms how much of this is structural to the system rather than a moral defect in your kitchen. The deliverable is your step-three gap broken into percentages that add to 100. Verify by weighing a full day's waste and contrasting it against the bucket you assigned it to. Three controls are enough to keep the gap from reopening, and none costs more than what a single mis-served portion burns: a digital scale on every plating station with the card's gram weight taped in plain sight, a cooler temperature log filled twice per shift against a hard 2 to 4 °C range in refrigeration, and a waste sheet where the cook writes down what got tossed, how much it weighed and why.

Step 5: install the three daily controls that hold the number in place

Diego F. Parra defends an order almost nobody respects, and it's the same order that runs this guide: theoretical cost first, actual cost for the period second, and only then a conversation about waste. Backwards it fails. The deliverable is three sheets filled for 14 straight days, and you verify by crossing the waste sheet against the closing gap: if staff report 200,000 pesos tossed and the gap says 3 million, there are 2.8 million nobody is looking at. The costliest mistake is treating waste as a discipline issue before it's an arithmetic issue: you can ask a cook to be careful, but you cannot ask him to compensate for a process that measures nothing, which is why so many control programs go dark within three months. The second is counting inventory whenever there's time, sliding the closing date, which makes two consecutive periods incomparable.

The four mistakes that sink 80% of waste programs

Third comes freezing the recipe card while suppliers keep raising prices; a 2023 recipe costed against 2026 invoices drifts four or five points on its own. And the fourth, the quietest one, is measuring only protein because protein is visible, when oil, dairy and aged cheeses hold another large share of spoilage. Correct all four and waste stops being a suspicion about your people and becomes a line in your P&L that you move at will. Your waste system works when you can answer five questions with a number instead of an opinion, and the evidence fits on one page. One: what was your actual food cost last period, calculated with inventory at both ends? Two: what was the theoretical cost according to cards dated within the last 90 days? Three: how many points wide is the gap, and what does it translate to in money per month? Four: how does that gap distribute across portioning, cold chain, process and price, in percentages adding to 100?

How do you know it all landed? Closing checklist?

Five: did the gap shrink versus the prior period? The threshold I demand is a gap under 2 points of net sales and a per-dish food cost that never crosses 32%, which is the MAXIMUM tolerable and not a target.

Once all five answers exist, stop auditing and move your energy to the next bottleneck. The first difference is arithmetic: actual food cost equals opening inventory plus purchases minus closing inventory, divided by net sales for the period. Without inventory at both ends you are measuring your buying rhythm, which is a different animal. That single error explains several points of deviation in any operation that never closes inventory, and it is the number one reason an owner swears food cost is 30% while the bank shows nothing left. The second is yield. A 6-kilo beef primal does not deliver 6 kilos of portions: it delivers between 4.2 and 5.1 depending on the cut and the hand working it.

The four differences that move cash

If your recipe card costs on purchase weight rather than servable weight, every plate is born with 15-25% of undercosting baked in, and no amount of pressure on the team closes a gap that lives on paper. The third difference is cadence. Monthly inventory gives you twelve correction opportunities a year; weekly counts on the eighteen SKUs holding 80% of spend give you fifty-two, and they surface deviation while something can still be done. Signal speed beats decimal precision. The fourth is the uncomfortable one: some dishes on your menu lose money on every sale and volume hides them. Once theoretical cost is compared plate by plate against menu price, two or three popularity stars with negative contribution margin almost always surface. Pulling them bruises the ego and repairs the P&L.