The window is short and it closes quietly

DFM at concept costs an hour. DFM after tooling costs a re-cut.

It is the same change. Same drawing, same engineer, same twenty minutes of thinking. What differs is only when it happens — and that difference is worth more than almost any other decision in the project.

The short version
  • The same change costs an hour at concept and a steel re-cut after tooling.
  • A review that does not name the process it reviews against is not a DFM review.
  • Factories send checklists because a real review is unpaid work before a PO. It is an incentive, not ignorance.
1 hrthe same change, at concept
re-cutthe same change, after T1
66engineers and programmers doing the reviewing

The same change, priced at four moments

cost of the change → Concept an hour Design frozen a CAD revision Tooling released re-quote + delay After T1 steel re-cut In market recall ← changes are nearly free · changes get expensive →
Nothing about the change gets harder. Only the number of things already built on top of it. This is why the highest-leverage hour in a hardware project is spent before anyone cuts steel.

What a real review looks at

A DFM review that does not name the process it is reviewing against is not a DFM review — it is generic advice. The same housing gets completely different feedback depending on whether it is going to be moulded, cast or machined.

Geometry against the process

  • Wall sections and uniformity — why thickness is the price
  • Draft, and whether it survives the texture — 0° / 1° / 3° compared
  • Parting line and where it lands on a visible face
  • Gate and ejector positions versus cosmetic surfaces
  • Undercuts, and whether they justify side actions

The commercial consequences

  • Tolerance chains and which ones drive the price — stack-up →
  • Cavity count against realistic annual volume
  • Whether the finish can exist on that alloy — anodizing on castings →
  • Cast-and-machine versus machine-from-solid at your volume
  • Which features can be loosened to pay for the two that cannot

Why most factories send a checklist instead

Founders often read a thin DFM report as incompetence. It usually is not. It is an incentive.

A genuine review is unpaid work performed before there is a purchase order. A careful review of a die cast housing is several hours of a senior engineer's time. Quoting fast wins the job. Quoting carefully loses it to whoever asked no questions.

Watch whether the factory asks you anything at all.

A supplier that comes back with questions about your application, your volume and your critical surfaces is showing you how they will behave for the next three years. One that returns a price in an hour is showing you that too.

The uncomfortable version, from the factory side: I am describing my own industry's incentive, and I have no clean solution to it. What I can tell you is that the fastest quote is the one that did not ask any questions — and in casting, that is very often the most expensive outcome.

Read: the factory that refused a $38,000 tool order →

Who does the reviewing

30 tool engineers, 15 project engineers, 15 CNC programmers and 6 product designers, working in UG, Pro/E, CATIA, SolidWorks, AutoCAD and Moldflow — on the same site as the machines that will make the part.

Engineer at a CAD workstation
Design review where the tool will actually be built
Mould open in the machine
Where the design decisions become permanent
CMM verifying a part
And where they get checked — six CMMs to ≤0.0005 mm

Send the drawing before you send the RFQ

You get back the two or three features driving your cost, the ones that will drive your rejects, and whether the volume says mould it, cast it or machine it. Quoted per part — and cheaper than one avoided re-cut by an order of magnitude.

DFM review & project rates

Producing at volume already? The managed department · First product? Free templates & checklists