How to rewrite a recipe as a structured bill of materials, scale it to any batch, and measure the yield and waste a real run gives you — the numbers your production, inventory, and costing actually run on.
Real help from real food people
Chef Diego runs a real food plant. If this page didn't get you there, tell us — a person reads every message.
After this lesson you can take one product's recipe and turn it into a bill of materials a system can run from — scale it up or down without the numbers drifting, tell the difference between the units a batch should make and the units it actually made, and know why a change to that one document has to reach purchasing, inventory, costing, and your label. It helps to have sketched a product description and a process flow first — the early steps of a HACCP plan — but you can build a BOM without them.
A recipe is written for a person. It carries technique, order, judgment: bring it to a simmer, fold until just combined, cook until it pulls from the side of the kettle. A person reads it and does the work.
A is written for everything else in your operation. It strips the recipe down to a list: this material, this quantity, this unit, to make this much output. It says nothing about how to stir. It says exactly what goes in and how much.
You need both, and they are not the same document. The recipe runs the hands on the floor. The BOM runs the numbers: it tells purchasing what to buy, tells the batch what to consume, tells inventory what to draw down, and gives costing the quantities it rolls a price from. When people say a recipe "drives production," what actually drives it is the BOM — the structured version underneath.
The reason to separate them is that a quantity you can only read in prose is a quantity no system can act on. "A good handful of the spice blend" is a fine instruction and a useless BOM line. The BOM forces every material into a number and a unit, and that discipline is what lets the rest of this lesson work.
Because a BOM is quantities, you can scale it with arithmetic instead of guesswork. Pick the batch size you want, work out the factor against the BOM's base batch, and multiply every line by it.
The figures here are made up to show the shape. Say a hot-sauce BOM is written for a base batch of 120 twelve-ounce jars:
You need a run of 300 jars. The factor is 300 ÷ 120 = 2.5, and it applies to every material line:
The factor is exact, so the material math is exact — as long as two things hold. First, you scale off the structured quantities, not a memory of "about a case of peppers." Second, you round to the unit you can actually measure: nobody weighs 1.875 lb by eye, so the floor needs that as grams on a scale, or the BOM needs a unit it can hit. Handling those unit conversions cleanly — pounds to grams, eaches to cases — is its own subject, and a later lesson in this course takes it on.
Not every line scales one-for-one
Materials that go into the product scale with the batch. Some things around it don't: a line-startup flush, a fixed sanitation step, or a fermentation starter can be close to constant whether you run 120 jars or 300. Scale the ingredients by the factor, but check any fixed or step-change input by hand rather than assuming everything moves 2.5×.
The BOM says a batch makes 120 jars. That is its . It is the honest planning number, and it is almost never what you end up with.
What you end up with is the . Say that 120-jar run gives you 111 good jars. The gap — 9 jars, or 9 ÷ 120 = 7.5% — is not a rounding error to wave off. It is where a chunk of your margin lives, and it is invisible unless you write both numbers down on every run.
Two numbers, captured the same way each time: what the BOM said the batch should make, and what the batch actually made. Do that for a few runs of the same product and the average gap becomes a number you can plan around — you order materials for the yield you really get, and you cost against the units you can really sell, not the paper count.
The 9 lost jars did not vanish evenly. They left the batch at specific points, and the reason to record where is that the pattern tells you what to fix. Break the loss down and it might read:
That total gap — the difference between what a batch should have produced and what it actually did — is the batch's . On a cooked product, some of it is not even a mistake: heat drives off moisture, so a kettle that takes in a certain weight of batter hands back less finished weight. That evaporation is real yield loss and belongs in the count, not written off as "how cooking works."
Recording loss by where it happened turns a vague "we always lose a few" into something you can act on. Loss clustered at startup points at the filler setup. Loss clustered at QA points at a spec that keeps catching product. You cannot improve a number you only ever see as a single lump at the end of the day.
Here is the part operators feel most and plan for least. Change a product's — swap a supplier's blend, cut the sugar, add a stabilizer — and you have not just changed the recipe. You have changed the BOM. And everything that derives from the BOM is now, quietly, wrong until it catches up.
Purchasing is still buying the old material. Inventory is still depleting the old one. The batch cost still reflects the old quantities — re-costing a recipe when its make-up changes is exactly the work the true-COGS lesson on costing from your recipes walks through. And the one that bites hardest: your label may now be wrong.
A label carries the formulation whether you meant it to or not. Federal rules require the ingredient list to run in descending order of predominance by weight (21 CFR 101.4), so adding more of one ingredient can change the order the list has to print in. And the food allergen law requires every one of the nine major food allergens to be declared — sesame became the ninth on January 1, 2023 under the FASTER Act. Reformulate to add a spice blend that contains sesame and never update the label, and the label is now missing an allergen: misbranded on paper, and a real hazard to a sesame-allergic customer.
Confirm your own label against the current rule
The label is its own subject — a later track covers it in full, and the counts and dates above move, so verify them on the source rather than trusting any page, including this one. The point here is narrower: the BOM sits upstream of the label. When the formulation changes, the label is one of the downstream documents that has to change with it, and it is the one with legal teeth.
All of that downstream damage has a single cause: the recipe living in more than one place. A binder on the floor, a spreadsheet in the office, a label file with the printer, a cost estimate in an email — change one and the others go stale, and nobody notices until an audit, a recall, or a margin that quietly slipped.
The fix is not a better binder. It is making the BOM the one place the formulation lives, and having production, inventory, costing, and the label all derive from it — so a change happens once and flows everywhere. Software built for food does this by holding the recipe as a single bill of materials that a batch runs from, inventory depletes against, and cost rolls up from; Bettr Manager, the operations platform this site is part of, works that way. The tool is not the point. The discipline is: one BOM, one source of truth, everything downstream derived from it — and that discipline is enforceable on paper too, it just takes far more hand-work to keep honest.
The BOM tells you what a batch should consume. The next lesson takes the batch you actually ran and ties the specific ingredient lots it consumed to the finished lot it produced — the moment the BOM stops being a plan and becomes a traceable record. First, build the BOM and watch one real run against it.
Do this in your operation