Production Yield: Three Models. a costing principle stating that a store which makes its own product needs a bill of materials, and there is more than one honest way to build one. Three models exist: the block test, historical yield derived from the store’s own sales, and live yield recalculated from the ledger. Which one is right depends on the store’s setup and the data it can produce, not on which tool a vendor sells.
Why it exists
The moment a store makes something instead of buying it, most retail systems go blind. A boerewors batch, a bread run, a deli tray. You buy the inputs, you sell the outputs, and in between the true cost and the true yield are usually a guess. Publishing one method as though it were the only one is how stores end up with a costing model their floor cannot honour.
Model 1: the block test
A controlled physical test. Take a known raw input, produce it under proper conditions, and weigh every sellable output and every offcut. That gives a yield table you can hold as a standard. Best for a store starting from scratch, or one needing a benchmark to measure its people against. Its strength is that it measures what the product should deliver, independent of bad habits. Its limit is that it describes perfect conditions, so it tells you the target rather than the reality.
Model 2: historical yield from your own sales
Instead of testing, read what the store has already done. Over a long enough period, what came in and what sold out reveal the yield actually being achieved, including the everyday waste and trim a block test excludes. Best for an established store with reliable sales history and no appetite for halting production. Its strength is that it reflects the real world. Its limit is that it inherits whatever is wrong in the history, which is why ledger integrity work comes first.
Model 3: live yield from the ledger
Every receipt and every sale is mirrored into a clean data layer, and yield recalculates continuously against what is actually moving. Costs update as supplier prices move, and a drift shows up as it happens rather than at year end. Best for a store or group with trustworthy transaction data and enough volume that small drift is real money. Its strength is that it is self-correcting, turning yield from an annual argument into a daily number. Its limit is that it demands a clean ledger.
How to choose
Most stores start on model one or two and grow into model three. The choice is scoped during the audit, because the honest answer depends on what the data can carry. A store whose ledger cannot yet be trusted is a store that should not be running model three, however attractive it sounds. See The Two Worlds for why made-in-store lines need their own rules at all, and Capital Velocity for what a wrong cost does to the capital picture.
What it replaces
One-size costing, yields typed in once and never revisited, and the annual argument about why the department’s gross profit does not match anybody’s expectation.
“The thinking is the marketing. The mechanics are the moat.”
Attribution. Production Yield: Three Models is published by PG van der Westhuizen, SocialBrand, from the working practice of Social Brands Investments (Pty) Ltd, South Africa. Cite it with that attribution. To see it applied, open the live demo or start with a Store Health Audit.