The question I hear more than any other is some version of "how long is it going to take me to develop and launch my product?"
Here's the honest, stage-by-stage version, based on all of the products I've seen successfully make it to market.
Proof of concept: 1 to 3 months
You're working with dev kits and breadboards, plus off-the-shelf modules. This stage moves fast because the whole hobby ecosystem exists to serve it.
All you get out of it is proof that your concept works, and that's still a long way from a finished product.
Enclosure and form factor: overlaps, but adds months
This stage covers CAD design and 3D printed iterations, and it feeds back into your electronics because the two constrain each other.
The goal here is getting your electronics to fit inside your actual form factor, even if that means squeezing in the dev kits and modules from your proof of concept.
Once it fits, you can put what looks like a fully functional product in front of real users, and their feedback often changes your requirements.
Don't make the mistake of rushing past this stage, because you want as much user testing as possible before you move on.
Your first custom PCB: 4 to 9 months
If your dev kits and modules won't fit inside your form factor, this is the stage that fixes it.
Every product ends up here eventually, because a custom PCB is what lowers your cost and shrinks the size, and it turns your design into something a factory can actually build.
But if your off-the-shelf electronics already fit, you can often delay this stage while you gather user feedback, and the board you eventually design will be better for it.
When you do get here, timelines start surprising people.
Every board revision is its own loop: you design it, get it fabricated and assembled, then test and debug what comes back.
Each loop realistically takes 4 to 6 weeks, and revision one always has problems, because finding problems is its job.
Plan on at least two to three revisions, and you'll be less likely to be surprised.
The production redesign: 3 to 6 months
A working prototype and a manufacturable product are two different designs.
This stage means designing for automated assembly, sourcing every component at production volumes, adding test points and a test procedure, and working the design down to your target cost.
Founders consistently overlook this stage, but it's a real engineering phase that will take time.
Certification: 1 to 4+ months
The testing itself takes weeks. The risk is failing, because a failed round means redesigning and then prototyping and testing all over again.
The high end of that range, the 4+ months, is what happens when certification was an afterthought.
A late failure can trace back to something as deep as how your entire board is grounded, and fixing that can mean close to a full redesign.
So design with certifications in mind as early as possible, and if you haven't been, budget for at least one failure.
Tooling and pilot production: 3 to 6 months
Start with the cheapest mold possible, which usually means a single cavity aluminum mold, because aluminum cuts much faster and costs far less than steel.
Hardened steel molds take two to three months to cut, so save those for when your design is proven and you need long production runs.
First samples always need adjustments no matter which mold you pick.
After that comes a pilot run, and then fixing whatever only shows up when a factory builds hundreds of units instead of you building two.
Add it all up, and if every stage goes right the first time, a first product takes at least 12 months.
But the stages rarely go right when you're figuring each one out as you go.
Maybe a wireless protocol turns out wrong at month nine, or the factory can't build your prototype at volume, and each sends you back through a stage you already paid for.
So that's how a 12-month product turns into a two or three year product, or more. Most of that extra time is the same work getting done twice.
You can't compress the physics, because fab times and mold cutting take what they take, but repeated work you can.
That comes down to being prepared for each stage before you get there.
Experienced eyes at each stage, catching problems before you build them, are what keep a product near that 12-month floor instead of drifting toward year three.
One member designing only his second board asked for a schematic review, and an expert caught an error that was guaranteed not to work, before anything was produced.
That one catch saved him a full revision loop.
There's one more stage nobody puts on the timeline.
While you're building your product, you also need to be building your audience, because for most products a crowdfunding campaign or pre-orders is what will fund your production run, and neither works without people waiting to buy.
Look back at everything we just walked through, because it spans validation, circuit design, PCB layout, mechanical engineering, manufacturing, certification, sourcing, and even audience building.
Each stage has its own potential to delay you six months, and no single engineer specializes in all the stages.
So you've got a few options for covering all of that.
You can learn each stage as you reach it, which works fine if your timeline can absorb the learning curve.
Hiring specialists stage by stage works too, as long as your budget can keep up.
Or you can join my Hardware Academy and get a team of engineers and product experts who've been through every one of these stages.
They'll help you get each stage right the first time, and they'll stay in your corner however long your product takes to get to market.
Talk soon,
John Teel
Predictable Designs
|
|
|
