From bad idea 👎 → pivot hell 🔥 → finally Trace 👾 the road to YC S25

Last year @Clustr looked “healthy” with 10,000+ active users, 10-minute sessions and $20,000 in MRR. Yet users churned quickly – traction didn’t equal stickiness. After hundreds of growth experiments we finally pulled the plug, trimmed the team from 12 to 2, and plunged into pivot hell. 🔥

Dozens of dead ends later, I flew to San Francisco hunting for answers. On a quick walk in Dogpatch, Tom Blomfield (Monzo Bank/GoCardless) hit me with the line that finally cut through the fog:

“Move fast – caution kills startups.”

We sprinted:

- 54 expert interviews in two weeks

- Shipped our first code

- Closed three big pilots

After a long, scrappy road we’re finally building something that feels inevitable: @Trace – a workflow automation platform that breaks down existing projects, assigns repetitive tasks to AI agents and routes the rest to the right people. Teams jump from idea to execution in minutes, without changing how they work.

We set the YC application deadline as our launch clock, prepped like mad, and finally got the call:

“Welcome to the YC Summer ’25 Batch” 🚀

What the journey taught us:

1. Kill the tarpit early. Metrics can mask fatal flaws.

2. Momentum beats certainty. Speed produces real signals.

3. Stay small and decisive. Two aligned builders beat a stuck dozen.

4. Talk to users daily. Conversations outclass conjecture.

If repetitive busywork drains your team, let’s chat. @Trace is looking for pilot teams ready to let AI agents handle the grunt work – so people can finally focus on what they do best. 💡

PS: Thanks to @rajiv_ayyangar & @aaronoleary for covering our story ⬇️

https://www.producthunt.com/stories/unwinding-our-clustr-fck