Title: I built a product nobody wanted. Now I'm starting over — this time validating first.
3 weeks ago I started building FlowForge — a visual AI workflow builder. Spent weeks on the UI, the execution engine, the node canvas.
Then I searched the market properly.
Zapier. Make. n8n. Gumloop. All doing exactly what I built. With years of head start and thousands of users.
Killed it immediately.
Here's what I actually learned from building something nobody needed:
I can ship a real product UI fast
I know my stack end to end now
Building without validating first is just expensive journaling
So I'm starting over. New idea: NotionAudit — an AI tool that scans your Notion workspace and tells you exactly what's cluttered, unused, and slowing you down.
This time I'm validating before writing a single line of code.
Landing page is live. No product built yet. Just trying to find out if real people feel this pain before I commit another 3 weeks.
https://notion-audit-ruddy.vercel.app/
Has anyone else killed a product after building it and started smarter the second time?
One thing I'd be careful with:
The expensive mistake may not be building before validating.
It may be assuming the lesson from FlowForge automatically applies to NotionAudit.
Those sound similar, but they can lead to very different decisions about what signal you're actually looking for this time.
I'd be careful not to make that call too quickly.
that's a fair pushback.
with FlowForge the market was saturated — clear cut. with NotionAudit the gap might exist but I haven't confirmed real willingness to pay yet, just that the pain exists on Reddit.
so you're right — the signal I need this time isn't "does the problem exist" but "will someone pay to solve it before I build anything."
still figuring out what that proof actually looks like in practice.
Possibly.
The reason I'd still be careful is that I don't think the interesting question is what proof looks like.
I think there's a more important decision sitting underneath that question.
I'd be careful making that call too quickly.
I wouldn't try to unpack it properly in a thread.
If you're curious, drop your email and I'll put together the tighter version.