Why designs that work perfectly as a prototype often fail — or become expensive — at production scale, and the core principles that keep a physical product buildable and affordable.
Design for manufacturing (DFM) is the practice of designing a product so it's easy, reliable, and affordable to actually build, not just possible to build once by a skilled machinist or molder who already knows all of its quirks. A design that looks finished on screen, or even as a single 3D printed prototype, can turn out to be unbuildable at the volume and price your business actually needs.
Take a founder named Priya, developing a plastic housing for a handheld tool. Her prototype, printed on a desktop 3D printer, has a snap fit lid with a sharp interior corner and a wall that changes thickness abruptly partway down. It looks great and fits perfectly, because a 3D printer builds material up layer by layer and does not care about sharp corners or sudden thickness changes. When she sends that same design out for an injection molding quote at 5,000 units, the molder flags both problems: the sharp corner will crack under molding stress, and the thickness change will cause the plastic to sink and warp as it cools. Fixing both means redesigning the part, which is cheap to do on screen and expensive to do after a mold has already been cut around the old design.
That gap, between what a prototyping process forgives and what a production process demands, is exactly what DFM exists to close before it costs real money.
The core reason DFM matters as early as it does: a change to a CAD file costs an afternoon of a designer's time. The same change after a prototype has already been built and tested costs that time plus the wasted prototype. The same change after production tooling exists (a steel injection mold can run $15,000 to $80,000 depending on complexity) means paying to modify or scrap that tooling on top of everything else. Catching a manufacturability problem before tooling is cheap. Catching it after usually is not, and the gap between those two costs is often measured in tens of thousands of dollars, not hundreds.
Five principles that hold across almost every process
Mistakes that show up as surprise costs later
0/4The right process depends heavily on your expected volume
| Good for | Watch out for | |
|---|---|---|
| Injection molding | High volume (typically 10,000+ units), complex geometry, very low cost per part once tooling is paid off | Tooling costs $15,000 to $80,000 or more, and it takes 6 to 12 weeks before the first production part exists |
| CNC machining | Low to medium volume, tight tolerances, metals and engineering plastics | Unit cost stays roughly flat regardless of volume, and cutting time makes it slower per part than molding |
| 3D printing | Prototyping, one-off or very low volume parts, geometry that would be impossible to mold or machine | Rarely cost-effective once volume climbs into the thousands, and layer lines affect strength and surface finish |
| Sheet metal fabrication | Enclosures, brackets, and structural parts made from flat or bent metal | Design constraints around minimum bend radius, hole placement near bends, and material thickness |
The comparison above lays out the tradeoffs, but most founders are really asking one question: given my volume and my part, what should I actually quote this in right now? The decision tree below walks through that, including an option that often gets skipped over: bridge tooling, a lower cost aluminum mold that produces a few thousand good parts before wearing out, used to validate a design and early demand before committing to full steel tooling.
Which manufacturing process fits your part
What's your expected order volume for this part over its first year?
Walk through before finalizing a design for quoting
0/5Check your understanding
Priya's team designs a bracket as two injection molded pieces joined by two screws, because that was easier to model in CAD at first. A manufacturing engineer suggests redesigning it as one molded piece instead. What is the strongest reason to make that change before tooling is cut?
A design calls for a machined slot toleranced to plus or minus 0.01 millimeters, but the part only needs plus or minus 0.1 millimeters to function correctly. What is the practical effect of leaving the tighter tolerance in place?
A founder expects to sell about 3,000 units of a molded plastic clip in year one, then scale past 20,000 in year two if it sells well. Full steel injection tooling costs $60,000; a comparable aluminum bridge tool costs $12,000 but wears out well before 20,000 units. What is the most sound approach for year one?
A connector housing is symmetric enough that it can be plugged in two different ways, only one of which is electrically correct. Which DFM principle addresses this directly?
Previous
Choosing and Working With a Contract Manufacturer
Next →
Supply Chain and Logistics Fundamentals for Physical Products
Ask a question about this lesson or share your take.
Loading…