Here’s a question we answered in the Interrupt Slack channel after the From Concept to Launch: What It Takes to Build and Ship a New Device panel discussion. We’re sharing the answer on our blog to make it available to a wider audience.
Starting from “I have a working MVP”, how many iterations do you recommend for scaling from 1-10 units up to thousands in production?
The number of iterations product and company dependent, but there’s a general formula that is commonly followed: the “NPI Process.” I’ll paint in broad brush strokes here – much more information on this process is available on our website and around the web.
The NPI process defines multiple build events/stages, and you can view these as the iterations that let you scale up from single, bespoke units to mass production.
- Proto (your working MVP)
- EVT
- Building small batches of works-like and looks-kind-of-like prototypes for internal development and validation
- Generally, you don’t care about things like cosmetics, you will use use 3D printed parts or milled parts instead of dye cast parts, etc.
- Testing is a secondary concern; you’re bringing up test stations and collecting data, but you might pass every unit through the line regardless of pass/fail status
- Small quantities: 5-50 units, but varies by company scale
- DVT
- You’re pretty solid on the design, you’re working on improving DFM and making final design tweaks
- Cosmetics might not be there, but you should be using production-intent parts if possible and things like water proofing etc. should work
- You’re running test stations with realistic limits to understand (and improve) process yield
- These units will typically be used for certification efforts, reliability testing, and environmental testing – the design needs to be good enough to make it through those efforts.
- 2-5x more units built than EVT – e.g., 200 units in 20 unit build batches
- PVT
- You’ve got the design locked and ready to produce, now you’re focused on tuning manufacturing process, improving yield + throughput, etc.
- All finished parts are used (e.g., you’re using custom tooled parts and not milling/printing anymore)
- These units are typically “revenue-able” (you can sell them)
- No significant changes to design at this stage (if there are big changes, you’re having another DVT event and not PVT)
- This will often be a lot more units than previous builds – 10% of your first mass production run, perhaps
- Mass Production
- We’re building units for shipping to customers
- There is still an effort to reduce cycle time, improve yields, and add test coverage to address “escapes”.
At each stage, you’re refining your product design and your manufacturing processes, and building larger and larger quantities of units. In an ideal world, you would move smoothly through the process. It’s more likely that you’re going to have two EVT events. And you often hope for one DVT event, but you may end up with two if there are major design changes (especially likely if there are certification challenges or reliability problems).
Three events at a given stage is not unheard of… but these development builds are expensive. You’re certainly better off spending more time improving the design between two build events than you are rushing through the process and needing a third one.