I killed my own SaaS after 17 days.
Not because the code was bad. Mostly because I realized I was building for a market that barely exists.
The original idea was a “developer CRM” — kind of a lighter HubSpot for solo devs.
But solo devs don’t really wake up thinking:
“man I need a CRM today.”
Most people just use Notion until they either stop caring or move straight to something bigger like Attio. There wasn’t really a middle space to win.
What was interesting though:
during those 17 days, I kept building little scripts for myself.
Checking GitHub traffic.
Exporting waitlist CSVs.
Searching for my domain on X.
Looking through comments to see if the same people kept showing up.
I wasn’t managing a pipeline.
I was just trying to figure out:
“who is actually paying attention?”
That ended up feeling way more real than the CRM itself.
So I scrapped v1 and rebuilt it into something else called Trace.
The core model is pretty simple:
Person → Event → Signal → Timeline
Instead of dashboards everywhere, the idea is:
show me 5 people that matter today and why they showed up.
That’s it.
Still early. Still rough around the edges.
But this direction feels a lot more honest than trying to force myself into “sales software for developers.”
A few things I learned from killing v1:
Market category matters more than clever features.
I should’ve spent more time studying failed products in the space before building.Words shape products.
Once I banned words like “pipeline”, “lead”, “deal”, and “stage”, the product naturally stopped drifting toward fake-sales-team software.Autonomous systems can easily become self-referential nonsense if you’re not careful.
A surprising amount of rebuild time went into fixing loops where the system kept reinforcing its own assumptions.
Anyway — Trace is pre-launch right now.
Waitlist only for now.
Would genuinely love feedback from other builders who track users in weird manual ways like this.