Bill of materials management: one structure, several audiences, permanent disagreement

Why the design structure and the build structure part company

Engineering describes the product as it is designed. Manufacturing describes it as it is assembled, which introduces sub-assemblies that exist because they are stocked or subcontracted, consumables that are used but not designed, packaging, and a sequence the design has no view on. Attempts to force both audiences onto one structure end with a bill that serves neither: engineers add stocking levels they do not believe in, or planners work around a structure that cannot generate a coherent order. Two linked structures with a controlled transfer between them is more work to set up and far less work to live with.

Structure decisions you will live with for years

Depth is the first choice. A deep structure with many levels gives planning visibility of intermediate stock and generates an order at every level; a flat one is simpler and hides where material actually sits. Phantom levels let you keep an engineering grouping without creating stock, which is elegant until costing or traceability needs the intermediate to exist. The practical rule is that a level should exist where material is genuinely held, counted, subcontracted or traced, and nowhere else. Reversing these decisions later means reworking planning parameters, standard costs and every historic transaction that referenced the old structure.

Alternates, substitutes and approved sources

Bills need to express that a component may be replaced by another, and to constrain when. An alternate approved by engineering for all uses is different from a substitute approved by production for a single shortage, which is different again from a second approved manufacturer of the same specification. Systems conflate these regularly. The consequences are practical: an undocumented substitution corrupts cost, breaks traceability and can invalidate a customer approval. Record who may authorise each type, and make the substitution an entered transaction rather than a decision taken at the rack and mentioned afterwards.

Quantities, yield factors and units of measure

Quantity per assembly looks trivial and generates a surprising share of stock error. Bulk material consumed by mass or length needs a conversion that matches how it is bought and how it is issued. Yield or scrap factors built into the bill inflate demand permanently and are rarely revisited once the process improves, so they slowly become a hidden buffer. Items consumed but not counted still belong on the bill for costing even where they are not planned. Every one of these choices flows straight into purchase quantities, so a lazy entry becomes stock somebody has to explain.

Where-used is the query that prevents damage

Reading a bill downwards tells you what to make something from. Reading it upwards tells you what is at risk. A component obsoleted by its manufacturer, a raw material failing a quality check, a supplier price increase, a substance restricted by regulation — each becomes a bounded problem if you can list every parent, and an open-ended one if you cannot. That query only works on a maintained structure, which is the unglamorous argument for keeping bills current on products you no longer actively sell but still support with spares.

Frequently asked questions

Do we really need separate engineering and manufacturing bills?

If your products are simple and assembled in one step from bought parts, one structure is enough. The case for two appears as soon as you stock intermediates, subcontract an operation, package in variants, or consume material the design never mentions. The decisive question is whether engineers are being asked to model things they have no view on, or planners are working around a structure they cannot generate orders from. Either symptom means one structure is being made to do two jobs.

Who should own the bill of materials?

Ownership splits by layer rather than by document. Engineering owns what the product consists of and the revision at which it changes. Manufacturing owns how that content is decomposed for production, which levels are stocked, and the consumables and packaging attached. Finance owns none of it but depends entirely on both. Name individuals rather than departments for each layer, because unowned bills accumulate quiet edits that nobody can explain during a costing review.

How should consumables that we do not track appear on a bill?

Include them where they carry real cost, and set them to consume without generating a planned order or a stock transaction, so costing sees them and planning does not chase them. Leaving them off entirely understates unit cost and hides genuine spend on adhesives, gases, abrasives and cleaning agents. Reviewing that category occasionally is worthwhile, because the total is often larger than anybody assumed and no single person is watching it.