Portion cost and over-portioning: the guide that separates theoretical food cost from the one you actually pay

Start by weighing the plate as it goes out today, not as the recipe claims

The first deliverable of this guide is a table of REAL portion weights: twenty services of each of your ten best-selling dishes, weighed at the pass, written on a sheet, with no warning to the cook on day one. If your spec card says 180 grams of tenderloin and the average of those twenty weigh-ins comes to 197, you have just found a 9.4% cost overrun on that protein, and across 3,400 plates a month that becomes a number no supplier negotiation will ever give back. You will know this step is done when you hold three figures per dish: spec weight, measured average and maximum deviation. With sector net margins running between 3% and 9% according to Statista, two points of food cost swallow a third of your profit. A portion cost is worth something only if it starts from the price on your latest invoice and subtracts what the cut loses during trimming.

Cost the portion with the REAL purchase price and with trim loss included

Take the kilo of tenderloin at your purchase cost, weigh the trimmed product, and work out the yield: if 1,000 raw grams give you 780 usable ones, your true cost per usable gram climbs 28%, and that percentage is what belongs on the spec card, not the invoice figure. Then multiply by service weight and add garnish, sauce and cooking fat. The deliverable is a five-line card per dish: ingredient, raw cost, yield, portion weight, portion cost. With food-away-from-home inflation at +4.1% during 2024 according to USDA ERS 2025 (via Apicbase), a card costed with last year's prices lies by construction. Averages hide precisely what you need to see. An overall food cost of 30% can be built from starters at 18% and proteins at 44%, and if your servers sell the protein well you are growing sales while sinking in margin. Load every card with its individual percentage, sort them from highest down, and flag in red anything above 32%, the ceiling under the Masterestaurant criterion: a maximum, not a recommendation.

Set the target food cost dish by dish, never as a menu average

Payroll, rent and utilities do not load onto the plate; those belong to break-even, and mixing them is why so many menus end up with inflated prices that scare the guest off while fixing nothing. The deliverable here is a ranked list: which dish to correct first, how much it weighs in sales and how many points it wins back. A 3,000-gram scale accurate to one gram costs less than two dinner covers and it is the tool that rewires the kitchen conversation. Before the spec card, talking about portions meant accusing: «you are overserving». With the card and the scale you say «the card reads 180, this week averaged 197, let us review the cut», and that sentence can be worked on. As Diego F. Parra, consultant and founder of Masterestaurant, puts it, the cook's judgment decides the flavor and the scale decides the grams, and those two things never competed.

Put the scale on the pass and turn portion weight into data, not into a complaint

Add calibrated ladles and scoops for sauces and sides, mark your portioning molds, and keep the scale in plain sight. The step closes when the pass weighs the proteins of your five highest-rotation dishes without anyone having to ask. Without variance, a spec card is decoration. Every Monday work out theoretical consumption —dishes sold times spec weight— and compare it against the real consumption your inventory reports for that protein. If you sold 620 tenderloins that theoretically required 111.6 kilos and the count says 124 went out, you have 12.4 unexplained kilos: an 11% deviation, and there sits your leak, split among overserving, badly executed cuts and kitchen mistakes. The operating rule I use: above 5% you investigate, above 8% you intervene in the process that same week. The deliverable is a variance sheet by ingredient family carrying three weeks of history, because one isolated figure cannot tell a bad cut apart from an installed pattern.

Four mistakes that wreck a well-intentioned portion control program

The costliest mistake is not skipping the scale, it is weighing for one week and then dropping it; intermittent discipline produces numbers nobody believes and a team that learns this too shall pass. Second: costing with the supplier's list price instead of the invoiced price with its discounts and its increases, which on volatile ingredients shifts several times a quarter. Third, and most common among owners who came up through the kitchen, is failing to update the card when the supplier or the caliber of the cut changes, because a 78% yield quietly becomes 71% with no announcement. Fourth: using portion control to shrink the plate. Cut grams chasing margin and the guest notices by the third visit, and you traded two points of food cost for a drop in frequency worth far more. Suppose you leave things as they are because the restaurant «is doing fine».

What happens if you do nothing: the full scenario, with numbers?

Three points of overserving on 120,000 dollars of annual food sales is costing you 3,600 dollars you will never see again, and since that leak dissolves inside the «cost of sales» line of your P&L, you never go looking for it.

Then the ingredient increase arrives —recall the +3.5% year-over-year in food away from home for May 2025 reported by the National Restaurant Association— and your natural reaction is to raise prices. The guest pays for an internal mess. A year later the margin has not returned, because the leak still lives underneath the new price. In 2025 at least eight restaurant brands filed Chapter 11 in the United States according to Restaurant Business, and none of them failed over a single dish: they failed on decisions built over numbers that measured nothing. You have finished this guide when you can answer five questions with paper in hand and without hesitating.

Closing checklist: how to know everything landed

One: does a costed spec card, with yield and portion weight, exist for 100% of the dishes making up 80% of your sales? Two: is your last ingredient price update less than thirty days old? Three: is weekly variance on your three main proteins sitting below 5%? Four: does the pass weigh product when you are not there, which is the only proof the habit took hold? Five: is your individual per-dish food cost under 32% across the whole menu, and do you know exactly how much each exception you chose to keep weighs in sales? Four yeses out of five is not a pass; it tells you which one is missing. Print that list, tape it in the office, review it the first Monday of every month. The difference is not the scale, it is the CONSEQUENCE of the scale: once grams are measured, the conversation with your cook stops being «you are serving too much» and becomes «the card says 180 and the week averaged 197, let us review the trim».

What actually changes once grams stop being an opinion?

You can fix that; an accusation you cannot. Before spec cards existed, raising prices was the only visible lever, and it is the worst one because it punishes the guest for an internal mess.

With cards in place, the levers become gram weight, trim loss and yield per cut, three variables the owner controls without touching the menu. Portion control does not lower perceived quality. A guest cannot tell 180 g of protein from 205 g, but will absolutely notice that today's plate differs from last week's; consistency is hospitality, and it happens to be margin too. Over-portioning concentrates in few ingredients: protein, cheeses, expensive sauces and fatty sides. Costing all 42 dishes in month one is a sequencing mistake; start with the eight that drive 70% of sales and the four ingredients eating half your purchasing spend. I got this wrong for years: I believed the food cost problem lived in purchasing, in negotiating harder with suppliers.

What actually changes once grams stop being an opinion — in practice

Negotiation buys one or two points and you refight it every quarter. Portioning buys three to five points and it holds by itself once the habit is built.