Should-cost

When a buyer says the target is $1.10 on a part you priced at $1.41, the number did not come from nowhere. It came from a should-cost model, a bottom-up estimate built from the geometry, a standard routing and a table of regional labor and machine rates. The model is not malicious and it is not stupid. It is generic, and generic is the argument, because the model prices a theoretical floor while you are pricing yours.

In the negotiation a should-cost behaves like an anchor rather than a fact. It assumes clean CAD, a textbook routing, handbook cycles and rates averaged across a region, and it has never heard of your changeover reality, your plater’s minimum lot charge or the cleanliness spec in the notes. The wrong response is to discount toward it, which concedes that the model describes your floor. The strong response is specific, line against line, your measured cycle against its assumed one, your real outside processing quotes against its allowance, your dated bar price against its material index, because a model argues in averages and loses to receipts.

It costs money when nobody on your side has the receipts ready. A should-cost meeting with a quote that cannot show where its numbers came from turns into a haggle, and haggles split the difference. A quote built from your own closed jobs, the way what a burdened rate actually includes argues rates should be built, meets the model with something it cannot generate, evidence. On program work the same model returns every year wearing a price-down schedule, so the receipts are not a one-time defence.