Tuned: A Roadmap That Marks Where to Quit First
A solo founder's roadmap is closer to a map of kill-gates than a schedule. Cheap, decisive checkpoints before costly stages, and permission to quit early.
On this page
Only after I realized I had built backwards (last post) did I do the thing I should have done first. I wrote the roadmap. When I finished, though, it wasn’t a build schedule. It was a map of the places where I’m allowed to quit.
I now think that is what matters most in a solo founder’s roadmap. It is less a plan to build everything than a map of kill-gates. Cheap, decisive checkpoints sit before the expensive stages, and there you decide whether to keep going or stop. The worth of that map was never in a finish schedule. It was in permission to quit early.
A solo founder can’t finish the normal roadmap
The usual path for a hardware product is long. Plan, validate, build, fund, manufacture, operate. A normal company puts people and money on each stage and passes through them one after another. One person can’t walk that whole path as a straight line. Manufacturing and certification are not walls you clear by adding an AI to one person.
And the cost of each stage is very uneven. Planning costs almost nothing. Building takes time. Manufacturing means molds and certification, money you can’t get back once it’s spent. Operating is a cost that never ends. When cost leans this hard to one side, what matters in a roadmap stops being the stages themselves. It’s the space between them.
So people tend to fall into one of two places. It looks too big, so they never start. Or they jump into building without seeing the end. I was the second kind. Looking back, both mistakes share a root: reading the roadmap as a plan to finish.
This isn’t new: Stage-Gate
The idea already has a name. Stage-Gate, the standard new-product framework Robert Cooper laid out in 1988. You put a gate between stages and decide, on criteria set in advance, to Go, Kill, Hold, or Recycle. The point is to filter weak ideas out before they get expensive. What I called keep, pivot, or kill above is exactly Cooper’s Go, Recycle, Kill.
Stage-Gate was built for big companies, though. It is a portfolio tool: line up dozens of projects and filter which ones get budget. A solo founder has no portfolio to filter and no budget committee. Even so, I think it matters more for one person.
The reason is that building got cheap. Build cost used to be the brake. Making things was expensive, so before building you naturally stopped and asked “are we sure?” The AI harness took that brake off (last post). Once cost stops acting as a natural gate, you have to set one up yourself. In a big company Stage-Gate could look like bureaucracy. For a solo founder plus an AI, it stands in for a brake that’s gone.
Map the gates, not the stages
So I redrew the roadmap, from a list of stages into a list of gates.
There turned out to be more than one gate. There is one at every boundary: the end of validation, the end of funding, right before manufacturing. Each gate asks the same question. Given what I’ve seen so far, do I keep going, change direction, or stop?
Cost climbs steeply as you move right, so the later the gate where you stop, the more it costs. If you can fold at the earliest gate, you fold for the least. Seen that way, the map of gates was a map of where you can stop cheaply.
A gate’s job is simple. It lets a bar you set in advance make the call, instead of sunk cost. You write down “below this number, stop” before you enter the stage. Writing it after you’re in is too late. You let it through because you’ve already poured so much in.
The cheapest gate before the most expensive stage
The most decisive gate goes right before the most expensive stage. For hardware that’s just ahead of manufacturing, the point where money for molds and certification goes out and can’t come back. I set two gates at that door.
One is demand. The bar here can’t be a firmware question. It has to be “does a frustrated player pay for the coaching?” The survey’s “I’d buy it” is not enough; there has to be a real charge. The other is technical. Does the signal survive in the real form factor, not just on a lab board? That one only gets an answer once you build a full-size development sample. Both have to pass before I put money into molds.
I pushed these gates behind the build. That is what I meant by backwards (last post). Once I redrew it as a roadmap, the shape alone made it plain that they had to come before the build. Put the cheapest, most decisive questions at the door of the expensive stage. I think that is how one person saves months and serious money.
Manufacturing and certification I can’t clear alone anyway. They need outside help and partners, and the time the physical world takes, one step after another. The roadmap shows in advance where that wall sits. The gate has to be alive before I reach it.
Where the map fails
A gate only works when I’m honest. People want to pass their own gate. What they’ve already spent feels like a loss, and their own idea looks pretty. So the bar that means “stop” has to be set early and calmly, before the wish to pass shows up.
The harness makes this harder. When building gets cheap, gates get easy to skip too (the same trap as last post). The cheaper the build, the weaker the gate.
And a roadmap doesn’t know the future. The stage past the gate you just cleared can stay blurry. There is no need to freeze trying to draw all of it. Seeing clearly to the next gate is enough.
Wrapping up
So now I draw the gates before the stages. I decide what “stop” means before I enter a stage, not after. The cheapest decisive gate goes before the most expensive stage.
Looking back, the purpose of the roadmap was never to reach the end. It was permission to quit early. I think that permission is what keeps a solo founder going.
Comments
Loading comments...