Cooking Yield: How to Measure Cooking Loss and Cost the Cooked Portion

A recipe that lists a cooked portion — five ounces of brisket, an ounce and a half of bacon, four ounces of pulled pork — needs more raw product than the portion weighs. Costing that portion at the raw price per pound understates it by exactly the weight lost in the pan, the oven or the braise. The fix is one measurement and one division, taken at the same stage the kitchen portions from.

The three formulas

USDA's Agricultural Research Service defines cooking yield the same way in its Table of Cooking Yields for Meat and Poultry: the hot cooked weight, recorded after a brief rest, divided by the raw weight. The formula uses the weight of the whole cooked sample; USDA records the edible portion separately for its nutrient calculations. For a bone-in or skin-on item measured this way, the published cooking yield still includes bone or skin you may not serve.

Worked example: a braised brisket portion

USDA measured that brisket 20 times, and the individual yields ran from 65% to 76%. At the same $7.00 raw price, that range alone moves the portion from $2.88 to $3.37 — a 49-cent spread before anyone changes the supplier, the price or the portion. A table value is a sensible first estimate. The yield you cost with should come from your cut, your braise and your holding time.

USDA-published cooking yields for common restaurant cuts

The full table has 173 rows; the dataset is published by USDA with sample counts, standard deviations and moisture and fat changes. Three limits apply before you borrow a number from it. The studies followed fixed protocols — the ground beef patties, for example, were cooked to 160°F with no added fat — so a different method, end temperature or rest gives a different yield. Several rows rest on one to four samples, and the two chicken rows above come from a 1975 USDA handbook with no sample count. And the yield is a hot weight after a brief rest: a roast that is chilled, sliced and reheated for service is portioned at a different stage, so its usable yield has to be measured at that stage. For fruits and vegetables, the produce yield chart lists USDA's as-purchased to ready-to-use yields.

How to run a cooking yield test

1. 1.Pick one item and write down its exact specification: cut, grade, trim level, fresh or frozen, and whether it is enhanced (injected with a solution). Record the supplier, invoice date and price per pound.
2. 2.Weigh it raw at the stage the recipe starts from. If you trim first, weigh after trimming and record the trim separately. If it arrives frozen, weigh after thawing and draining.
3. 3.Cook it the way service does: same equipment, temperature, time, end-point temperature and rest. A test cooked more carefully than a busy Friday measures the test, not the kitchen.
4. 4.Weigh at the stage you portion from — hot after resting, chilled before slicing, or after slicing or pulling. Remove bone, skin, fat cap and unservable end pieces first if they never reach the plate.
5. 5.Divide the servable weight by the raw weight. Record the stage, the date and who ran the test.
6. 6.Repeat on separate deliveries and use the average, keeping the lowest and highest results beside it. One test is one observation, not a standard.
7. 7.Test again when the supplier, specification, cooking method or holding time changes.

When trimming and cooking both remove weight

Many items lose weight twice: first to trimming — fat cap, silverskin, bone — and then to heat. Multiply the two yields; do not add or subtract the percentages. An 81.25% trim yield followed by a 70% cooking yield leaves 0.8125 × 0.70 = 56.9% of the purchased weight as servable product.

Without the credit, the same test costs $5.27 per cooked pound and $1.32 per portion. The credit is honest only when the trim goes into something you sell and would otherwise buy; the butcher's test shows how to value it, and trim loss separates recoverable trim from waste. A credit for a stock you never make lowers the cost on paper and nowhere else.

Keep the portion cost current when the invoice changes

A yield test stays useful through many price changes, because the yield belongs to the product and the method, not to the price. Express it as a cost factor and each new invoice price converts straight to a cooked cost.

The shortcut assumes every part of the calculation moves with the purchase price, and a fixed-dollar credit does not. In the pork example the factor is $4.7253 ÷ $3.00 = 1.5751. If the price rises to $3.30 while the trim is still credited at $2.50 per pound, the recomputed cost is $5.25 per cooked pound; the factor shortcut gives $5.20, about 1% low. Recompute the whole test when a credit matters, or update the credit's price at the same time.

In a costing tool that stores purchased quantities, record the raw quantity each portion needs — about 7.25 oz of raw brisket for a 5 oz cooked portion — against the ingredient's purchase price. Yield is already included in that raw quantity; do not apply the loss twice. In Marji, every extracted invoice line requires operator review before confirmation records a cost, regardless of OCR confidence. A confirmed price can then recost affected dishes; the recipe's quantities, units, yields and completeness remain your inputs to check.

Decisions a yield test changes

Which ground beef blend costs less per cooked pound

Leaner ground beef costs more per raw pound and loses less weight in cooking. USDA's “broiled or grilled” patties — oven-broiled to 160°F in its ground beef study — kept 63% of their raw weight above 22% fat, 69% at 12–22% and 73% below 12%. With these hypothetical prices, the cheapest raw blend is still the cheapest cooked blend, but by far less than the raw prices suggest.

The break-even premium is the ratio of the yields. At these yields, 90/10 costs the same per cooked pound as 80/20 when its raw price is 5.8% higher (0.73 ÷ 0.69 = 1.058), which is $4.87 against $4.60. Pay less than that premium and the leaner blend is cheaper to serve; pay more and it is not. Taste, juiciness and the patty the guest sees are separate decisions, and with four samples per fat level, run the test on your own grill before switching. To turn the patty cost into a menu price, see how to price a burger.

How much raw product to order

Yield also sets the order. For 120 brisket portions at 5 oz cooked and a 69% yield, you need 120 × 5 ÷ 0.69 = 869.6 raw ounces, or 54.3 lb of trimmed flat, before any allowance for holding loss or a busy weekend. The par level calculator turns that usage into an order-up-to quantity.

Portions specified by cooked weight

Bacon shows the effect most clearly. USDA's oven-cooked bacon (its “baked or roasted” row) kept 32% of its raw weight across three samples. A club sandwich spec of 1.5 oz cooked bacon needs 1.5 ÷ 0.32 = 4.69 oz raw; at a hypothetical $5.80 per raw pound that is $1.70 per sandwich, not the $0.54 you get by pricing 1.5 oz at the raw rate. If your spec counts raw slices instead, cost the raw weight and skip the yield.

Mistakes that distort the result

• —Applying yield twice. If the recipe already lists the raw quantity, cost it at the raw price. Dividing by the yield again overstates the dish.
• —Using a table value as your yield. Published yields come from set protocols and sometimes small samples. Use them to sanity-check your test, not to replace it.
• —Mixing stages. A hot-weight yield applied to a chilled, sliced product understates its cost. Record the stage with every yield.
• —Forgetting what is not served. Bone, skin and fat cap weigh something in the pan and nothing on the plate.
• —Comparing suppliers per raw pound. Two products at the same raw price can cost differently per cooked pound when their specification or added solution differs. Compare the usable cost instead.
• —Testing once, forever. Retest after a change in supplier, specification, method or holding time.

For a purchasing decision using the measured yield, compare whole and boneless chicken costs, including prep time and demand for the resulting cuts.