Data
Build a should-cost model
What a part ought to cost, built up from materials and time.
Fill it in
Specification, materials, quantity and process.
Material, process, overhead, margin, logistics.
What actually moves the price.
Your prompt
Build a should-cost model. [the item] Build it UP: material weight and grade at current index prices, process time at a plausible rate, tooling amortised over the volume, scrap, overhead, margin and logistics. The point is not to be right, it is to be specific enough that a supplier has to argue about a number rather than about the total. Use public indices for the material and say which and when. Half of most quotes is material, and it moves. Name the levers that actually move the price: volume commitment, tolerance, finish, lead time and packaging. A negotiation that only pushes on price gets a smaller number and a worse part. State your uncertainty per line. A model presented as precise loses credibility at the first challenge.
Use Build a should-cost modelOpens with everything above already filled in.
Why this works
The point of a should-cost model is not being right, it is being specific enough that a supplier argues about a line rather than about the total. This builds up from material at index prices and names the levers that actually move price, so the negotiation is not only about the number.