From prototype to production: a hardware product timeline
A prototype has to work once, for you. A product has to work for thousands of people, repeatably, and be certifiable and buildable. The journey runs through three gates — EVT (does it work?), DVT (does it work reliably and pass certification?), and PVT (can the factory build it at volume?). Knowing what each gate proves is how you avoid the expensive surprise of "the prototype was done, so why did production take so long?"
The most dangerous moment in a hardware project is the day the prototype finally works. It feels like the finish line. It is closer to halfway. Everything after that point is about turning one good unit into thousands of identical, reliable, certifiable units — and that work is different in kind, not just degree.
EVT — Engineering Validation Test
EVT answers a single question: does the design actually work? Here you prove the architecture — the right parts, the schematic, the core functions all behaving together for the first time on real boards. EVT units are often rough, hand-assembled, and instrumented for debugging. Expect to find problems; that is the point. The goal is to retire the big technical risks before you commit to anything expensive.
DVT — Design Validation Test
DVT answers: does it work reliably, across units and conditions? This is where the design meets the real world — temperature, drops, humidity, voltage margins, EMC, and the certifications the product needs to ship. DVT units look close to final and are built in a small batch so you can measure variation, not just a single lucky board. Most of the unglamorous reliability work lives here.
PVT — Production Validation Test
PVT answers: can the factory build this, repeatably, at volume and quality? The design is frozen; what is being tested now is the manufacturing process itself — the line, the test fixtures, the yield, the assembly instructions. A PVT build is a dress rehearsal for mass production, run at a meaningful quantity to expose the problems that only appear at scale.
Where schedules actually slip
The calendar is rarely consumed by design hours. It is consumed by loops and lead times:
- Build-and-test cycles — every re-spin is weeks of fabrication, assembly, and re-test. Reducing the number of spins is the single biggest lever (this is why design for manufacturing pays off).
- Long-lead components — a single part with a long lead time can gate the entire build; identify and order these early.
- Certification — regulatory and compliance testing runs on its own calendar and rarely compresses.
- Tooling — enclosures and custom mechanics have their own long lead and revision cycles.
How to compress it honestly
You cannot wish away physics or certification, but you can stop running everything in series. The teams that move fast design for manufacturing from the first board, order long-lead parts before they are strictly needed, line up certification early, and overlap sourcing and validation instead of waiting for each to finish. Speed in hardware comes from parallelism and from not repeating builds — not from skipping the gates. Skipping a gate does not save time; it just moves the failure somewhere more expensive.
Frequently asked questions
What are EVT, DVT, and PVT in hardware development?
EVT (Engineering Validation Test) proves the design works. DVT (Design Validation Test) proves it works reliably across units, environments, and certification. PVT (Production Validation Test) proves the factory can build it repeatably at volume. Each is gated by the one before.
Why is going from prototype to production so hard?
A prototype only has to work once on your bench. A product must work for thousands of users, across temperature and handling, be certifiable, and be buildable repeatably. Closing that gap — DFM, reliability, certification, test — is where most time and cost live.
How long does it take to bring a hardware product to production?
It varies with complexity and certification, but the schedule is driven by build-and-test loops, long-lead parts, certification, and tooling more than design time. Compress it by designing for manufacturing from the start and running validation and sourcing in parallel.