Most recipe costs are wrong in the same direction. The spreadsheet prices a kilogram of salmon at the invoice price, the kitchen loses a fifth of that kilogram to trim and skin, and every portion sold quietly costs more than the menu believes. Multiply that by batch preps costed on guesswork and portions that drift with whoever is on the pass, and a menu that promises 68 per cent gross profit delivers 63 and nobody can say why.
This guide covers the layer of recipe costing where the errors actually live: yields, batches, sub-recipes and portion standards. It assumes you know why costing matters; if you want the system it sits inside, start with our food cost management system pillar, and for keeping costs current as prices move, our guide to recipe costing when supplier prices change weekly is the companion piece.
Definition: A standard recipe cost is the agreed, documented cost of producing one portion of a dish, built from current ingredient prices adjusted for yield, the true cost of any batch preps it contains, and a fixed portion specification. It is the number every site and every shift is held to.
The word standard carries two meanings and both matter. It is standard as in documented, one version of the truth rather than the head chef’s memory. And it is standard as in enforced, the spec the plate is checked against. A cost that is accurate but not enforced is trivia; a cost that is enforced but not accurate is a target nobody can hit.
It is tempting to blame the market, but the market is currently the quiet part. The Office for National Statistics put food and non-alcoholic beverage inflation at 1.7 per cent in the 12 months to June 2026, the lowest annual rate since August 2024. The persistent errors are structural: costing ingredients as purchased rather than as usable, treating preps as free, and letting portions float. Each error compounds through the menu, and none of them shows up until the variance report or the P&L.
Accurate dish costs climb four rungs. Skip a rung and every number above it is flattering fiction.
|
Rung |
What it prices |
Where it goes wrong |
|
1. Purchase price |
The invoice cost per unit, kept current from supplier invoices |
Prices typed once and never updated; pack size changes missed |
|
2. Usable cost after yield |
Cost per usable unit after trim, bones, skin, peel and cooking loss |
Costing as-purchased weight; guessing yields instead of testing them |
|
3. Batch and sub-recipe cost |
Preps, stocks, sauces and doughs costed as recipes, then used as ingredients at true portion cost |
Preps treated as free; batch yields never measured |
|
4. Plated portion cost |
The dish at fixed portion spec, including garnish and accompaniments |
Portion drift; garnish and sides left out of the card |
Yield is the percentage of what you buy that ends up sellable. Divide the purchase price by the yield to get the usable cost, and cost every recipe at usable cost, never invoice cost. Yields are tested, not guessed: weigh a representative item before and after prep, twice a year per key line and whenever you change supplier or spec, because a different trim on the same product is a different yield.
Worked example (illustrative only). A 5 kilogram whole salmon costs 60 pounds, 12 pounds per kilogram on the invoice. After filleting, skinning and trim, 3.25 kilograms is servable: a 65 per cent yield. The usable cost is 60 divided by 3.25, which is 18.46 pounds per kilogram. A 150 gram portion therefore costs 2.77 pounds, not the 1.80 the invoice price implies. Costing at invoice price understates this one ingredient by more than a third. Figures are illustrative to show the method, not benchmarks.
Kitchens run on preps: stocks, sauces, marinades, doughs, spice mixes. Cost each one as a recipe in its own right, ingredients at usable cost, then record what the batch actually produces so it can be used inside other recipes at a true per-portion cost. The batch yield is measured the same way ingredient yield is: what went in, what came out, in usable portions.
Worked example (illustrative only). A stock batch uses 14 pounds of ingredients at usable cost and produces 10 litres after reduction. A risotto uses 250 millilitres per portion: 40 portions per batch, so 35 pence of stock per plate. Treat the stock as free and every risotto sold understates its cost by 35 pence; across a section of prep-heavy dishes, that is how a menu misses its GP target while every individual card looks fine. Figures are illustrative.
Write the spec in units the pass can check: weights and measures, not adjectives. Include the garnish, the sides and the sauce, because pennies per plate at volume are the difference between the costed margin and the banked one. Then give the spec a version: when a recipe changes, the old card retires everywhere at once, which is the difference between a standard and a suggestion, especially across multiple sites.
Two loops keep the ladder honest. Price currency: ingredient prices update from supplier invoices, and every affected recipe recosts when they do; done manually this is the job that stops happening first, which is the honest case for automation over templates, as our comparison of recipe templates vs manual costing sets out. And reality checking: theoretical vs actual variance is the audit of your standards, because a recipe costed perfectly and portioned loosely shows up there within a week. Divergence on a dish is the signal to re-test its yields and re-check its spec.
Everything above is arithmetic, and none of it is hard once; the difficulty is keeping hundreds of recipes true while prices, yields and menus move. Recipe management software holds yields per ingredient, costs batches as sub-recipes usable inside other recipes, reprices every affected dish when a scanned invoice changes a price, and pushes one master spec to every site. StockTake Online does all of this cloud-based from a phone or tablet, with the same recipe data feeding variance so the audit loop closes itself.
Test your own ladder first: take one prep-heavy dish and run it through the free Food and Beverage Cost Calculators at invoice price, then again at usable yield cost, and compare. If the gap surprises you, Book a Demo and bring that dish; we will build it with yields, batches and versioned specs live.
How do you cost a recipe accurately? Price every ingredient at its usable cost after yield, cost batch preps as sub-recipes with measured outputs, fix the portion spec including garnish and sides, and recost the dish whenever a supplier price or yield changes.
What is a good yield percentage? There is no universal figure; yield depends on the product, cut and prep. The working rule is that every key ingredient has a tested yield on file, refreshed periodically and whenever the supplier or spec changes, rather than a guessed one.
What is the difference between a recipe and a sub-recipe? A recipe produces a sellable dish; a sub-recipe produces a prep, such as a stock, sauce or dough, that other recipes use as an ingredient. Sub recipes carry their own ingredient costs and measured batch yields so their true per-portion cost flows into every dish that contains them.
How often should standard recipe costs be reviewed? Continuously for prices, ideally automated from supplier invoices; twice a year for yields, plus on any supplier or spec change; and immediately for any dish flagged by theoretical versus actual variance.
Should garnish and sides be included in recipe costs? Yes. Small per-plate items compound at volume, and leaving them off the card is one of the most common reasons a menu’s planned gross profit never reconciles with the banked one.