A founder's guide to design-for-manufacture

Most founders hear “design-for-manufacture” for the first time from a contract manufacturer’s quote — usually right after the quote comes back three times higher than expected. By then the part has already been drawn, the investor deck already shows the render, and changing the design feels like admitting failure. It isn’t. It’s the normal sequence when DFM gets skipped at the start.

What DFM actually is

Design-for-manufacture is the discipline of designing a part so it can be made repeatably, at your target cost, on the process you intend to use. That’s it — no more mystery than that. The mystery comes from the fact that “can it be made” and “can it be made the way you drew it, a thousand times, for the price you need” are two very different questions, and most first-time hardware founders only ever answer the first one.

At Sony, every headphone earcup I drew went through the same filter before tooling: wall thickness consistent enough to avoid sink marks, draft angles on every vertical face so the part releases from the mold without damage, snap-fits sized for the actual tolerance of injection-molded ABS rather than the tolerance of the CAD model. None of that is visible in a render. All of it shows up in your first production run if it’s missing.

The three questions I ask before opening CAD

1. What’s the target unit cost, and at what volume? A part that’s fine at 500 units can be unworkable at 50,000 — different tooling, different material grade, sometimes a different process entirely (CNC vs. injection molding vs. die casting).
2. What’s the actual tolerance you need, not the tolerance you assumed? Most consumer enclosures don’t need ±0.01mm. Most snap-fits and living hinges fail because someone specified that anyway, out of caution, without knowing what it costs.
3. Who is going to manufacture this, and what have they actually built before? A factory that’s good at sheet metal is not automatically good at precision plastics. The DFM rules change with the process and the supplier.

Where founders lose the most money

The single most expensive DFM mistake I see isn’t a bad draft angle — it’s sequencing. Founders often finalize the industrial design (how it looks), then hand it to an engineer to “make it real,” then send that to a manufacturer to quote. Each handoff treats the previous stage as fixed. By the time DFM feedback arrives, the tooling budget is already being discussed, and feedback that should have shaped the design eighteen months earlier becomes a fight over change orders.

The fix is not complicated: bring manufacturing-aware engineering into the room while the industrial design is still a sketch, not after it’s a finished CAD file. It costs less to redraw a wall thickness on day three than to requote a mold on day ninety.

If you’re earlier than that — still at the napkin-sketch stage — that’s the right time to start this conversation, not the wrong one.