5 BOM Errors That Cost Stamping Tool Shops Money
No stamping tool shop has a line item in its accounts called BOM error.
That is precisely why the problem survives. The loss shows up somewhere else — as air freight on a casting, as a weekend of overtime before tryout, as a bar of tool steel cut down for a smaller job, as a rack of parts nobody will ever book out. Each gets explained on its own. None gets traced back to the spreadsheet where it started.
A press tool or stamping die is a one-off product with 100–250 parts: long-lead castings, catalogue components with fixed lead times, and three or four design revisions before it ever sees a press. Whether the design lives in Siemens NX or Solid Edge, the Bill of Materials is the single document that converts all of that into purchase orders. When it is assembled by hand — exported from the CAD assembly, re-typed into a spreadsheet, revised under delivery pressure — it fails in five predictable ways.
1. The part that exists in the model but never made it onto the BOM
What it looks like: Assembly is underway. Someone reaches for a pusher pin, a wear plate, a spacer — and it is not in stores. It was never ordered. It was never on the BOM.
Why it happens: The part list is transcribed from the CAD assembly by a person. Multi-body solids are the usual culprit — a single NX or Solid Edge part file that physically contains four separate components. Unless somebody notices and splits it, three of those four pieces have no row, no stock size, and no purchase line. A 120-body casting solid is not unusual in NX stamping die assemblies. Neither is missing one.
What it costs: Never the price of the part. It costs the tryout slot. You pay expedite freight, overtime to recover, and if the part is long-lead you do not recover at all — the delivery date moves. A ₹2,000 pin can move a ₹25 lakh tool by two weeks.
How to stop it: Every physical part that must be bought or made needs its own row on first release. Multi-body parts have to be identified and split systematically — and the decision about which to split should be a short, explicit list someone signs off, not something an engineer scrolls past at 7 pm.
2. The stock size that is a few millimetres wrong
What it looks like: The bar arrives. The block is 8 mm short on one face after allowances. It cannot be used for the part it was ordered for.
Why it happens: Somebody opened each component in NX or Solid Edge, read the bounding box dimensions, added an allowance, and typed three numbers — 150 times in an afternoon. Transposition errors and stale figures after a design change are not lapses of care; they are the arithmetic of doing the same manual task 150 times under time pressure.
What it costs: Two ways, and the second is the one nobody counts.
- Too small: re-order long-lead material, re-plan the machining slot, and slip everything downstream.
- Too large: you never notice. You pay for material you did not need and machining hours you did not quote. It shows up as a quietly poor margin on a job that "went fine."
Undersized stock costs you delivery. Oversized stock costs you margin on every single tool, silently, forever.
How to stop it: Stock size should be derived from the model geometry — not measured by a human — and re-derived automatically every time the model changes. Assemblies and catalogue parts should be skipped automatically; manufactured components should be calculated. It is arithmetic, not judgement.
3. The photo sitting next to the wrong row
What it looks like: A vendor makes a part that matches the picture in the BOM, not the part number beside it. Or the tool room grinds to a profile from two revisions ago.
Why it happens: Photos are captured one at a time — orient the model in NX or Solid Edge, screenshot, crop, paste into a spreadsheet row. Insert a row, sort a column, and the images no longer line up with the parts. When a part is revised, the old image stays. Nothing marks it as stale.
What it costs: This error is dangerous because it looks verified. Nobody double-checks a row that has a picture — the picture is the verification. A wrong part gets made, inspected against the wrong reference, and discovered at assembly.
How to stop it: The image has to be bound to the part record — generated from the model and regenerated whenever the part changes. If a picture can drift away from its row, eventually it will.
4. Buying against a revision that has already moved
What it looks like: Purchasing does everything right, on time, against the BOM they were sent. The casting arrives at the height it was in Revision 1. Design moved to Revision 3 eleven days ago.
Why it happens: The design changed. Someone emailed an updated sheet — or meant to. Nobody told purchasing precisely what changed, because producing that list means comparing two 200-row spreadsheets by hand. That is not a task anyone does reliably twice.
So purchasing gets a new file and no diff. They either re-check everything (no time) or trust the changes were minor (they are not always).
What it costs: This is the single biggest generator of junk inventory in a stamping tool shop. The superseded part is already on order and cannot be cancelled. The replacement was never ordered because nobody knew it changed. You pay for both, and one of them goes on a shelf permanently.
How to stop it: Every revision needs a formal approval record and a difference list — added, removed, changed, with the exact field, the old value and the new value. Not "here is Rev 3" but:
Changed — UPR SHOE CAST: height 490 → 565; stock size 1380 × 890 × 490 → 1380 × 890 × 565
Changed — ADAPTER LIFTER: qty 4 → 3
Removed — PUSHER PIN D322
That is a purchase action list: order this, cancel that, amend this PO before the casting is poured.
5. The standard part nobody recognised as standard
What it looks like: A catalogue gas spring gets drawn up and quoted as a special. Or the same nitrogen cylinder is bought twice in six months under two different part names.
Why it happens: Knowing which parts are purchased and which are manufactured lives in the heads of your most experienced engineers. A new engineer re-learns it one mistake at a time, and the parts library grows faster than anyone can memorise.
There is a version that is worse: splitting a standard purchased assembly into its bodies in NX or Solid Edge creates phantom components that purchasing will dutifully try to have manufactured.
What it costs:
- A catalogue part made as a special costs several times the catalogue price plus lead time you did not need.
- Duplicates under different names inflate inventory and guarantee that stock already on your shelf is never found when needed.
- Phantom parts from a wrongly split assembly generate purchase orders for things that do not exist.
How to stop it: Standard part identification has to be a shared, growing company asset — a library that every BOM checks against automatically and improves every time an engineer confirms a part.
The pattern behind all five
Read them together and the common cause is clear. None of these are carelessness. Every one of them is what happens when a 200-row document that drives your entire purchasing spend is assembled by hand, from memory, in a spreadsheet, under delivery pressure, and revised three times before tryout.
The errors are the designed output of that process — which means they are fixable at the source, not by asking people to check more carefully.
| Error | The manual step that creates it | What removes it |
|---|---|---|
| Missed part | Transcribing the part list by eye from the NX or Solid Edge assembly | Automatic extraction + explicit split decisions |
| Wrong stock size | Measuring and typing 150 bounding-box sizes | Stock size derived from the model, every revision |
| Wrong or stale photo | Screenshot and paste, row by row | Photo bound to the part, regenerated on change |
| Superseded revision | Comparing two spreadsheets by hand | Revision compare: added / removed / changed |
| Unrecognised standard part | Personal memory of the catalogue | A shared standard parts library that learns |
What to do about it on Monday
You do not need a system to start. You need a number.
- On your next release, time it. How long from "assembly is ready in NX or Solid Edge" to "BOM is with purchasing"? Split by step: extracting the part list, splitting multi-body solids, identifying standard parts, calculating stock sizes, capturing photos, formatting for ERP and vendors. Write the hours down.
- Multiply by revisions. Three or four times per tool, times the tools you release in a year. That is the real figure.
- Audit one finished tool backwards. Take a job you have already delivered and list every part ordered and never fitted. Count — do not attribute blame. That number is your annual junk inventory rate on one tool.
- Ask purchasing one question: "When a design changes, how do you know exactly what changed?" The answer tells you how exposed you are to error 4.
Most stamping tool shops are surprised twice: once by how many engineer-hours go into producing the BOM, and once by how much of the material on the shelf traces back to a revision nobody diffed.
Where BOM automation software helps
The useful test for any BOM automation tool is not whether it produces a BOM — it is whether it removes the five manual steps above while keeping engineers in control of decisions that need judgement.
SmartAPQP — available through DDSPLM — is BOM automation software built for tool rooms working in Siemens NX, Solid Edge, and other major CAD platforms. Standard parts are matched against a shared library before anything is touched. Multi-body solids surface as a short decide-list, then split automatically with traceable names. Stock sizes are calculated from model geometry and recalculated on every revision. Photos are generated from the model and bound to their rows so they cannot drift. Every revision is published to one shared link with full history, approve/reject, and a precise comparison against any earlier revision — added, removed, changed with exact field values. ERP and vendor exports are defined once, then produced in two clicks.
CAD to ERP, in under five minutes.
Test it on a tool that actually hurt you. Send us one assembly from a stamping die or press tool you have already built, and we will return a published BOM — photos, stock sizes, standard parts identified, ERP-ready export — so you can compare it line by line against what your team released.
← Back to all posts