Vedansh

Building ChurnSense - churn prediction for B2B SaaS. Solo founder.

[Vedansh @vedansh_builds]

@12andDone You're exactly right. That's the core of it. The real dashboard shows exactly that: specific signals ("usage dropped 40%", "login gap", "payment failures"), not a bare risk number. The demo video just didn't show that part.
Want to see it?

[Vedansh @vedansh_builds]

I’m building ChurnSense: it tries to flag B2B SaaS customers drifting toward a cancel weeks early.

Still early, zero paying customers yet (demo data in the video).

What would make this actually useful to you?

[Vedansh @vedansh_builds]

If you want to look: https://t.co/oZUB5P8y3T. I’ll also run it on one cancelled customer of yours and tell you honestly whether the drift was visible.

[Vedansh @vedansh_builds]

If you run a B2B SaaS with 50–500 customers and churn keeps you up at night:

I'll set up ChurnSense for you free for 3 months.

You get: daily risk scores, Slack alerts, AI save emails.
I get: feedback and (if it works) a testimonial.

DM me or reply here.

[Vedansh @vedansh_builds]

Our product ChurnSense is now on @indieneedcom.

I'd love to hear your thoughts.

https://t.co/jelj7mPpce

[Vedansh @vedansh_builds]

@elonmusk https://t.co/oZUB5P80el

[Vedansh @vedansh_builds]

@verbove https://t.co/oZUB5P80el

[Vedansh @vedansh_builds]

@Joshua_WD Shipped it. New accounts now get dampened based on actual first-activity, not just when we ingested the row, so a fresh signup and a year of imported history don't get treated the same way anymore.

Appreciate you pushing on this, better product for it. Wanna try this out?

[Vedansh @vedansh_builds]

Building ChurnSense in public.

Generic churn models miss context: daily logins can mean “healthy” for a dev tool, but not for an HR platform.

So I built per-tenant models that learn each customer's own baseline.

Shipped it. Here’s what it looks like:

[Vedansh @vedansh_builds]

@Joshua_WD Honest answer: nothing protects them yet. There's tenure logic at training time, but it doesn't carry through to the live score you'd actually see on a new account's dashboard. Real gap, not solved, tightening it up now.

[Vedansh @vedansh_builds]

@Joshua_WD Good question. Two clocks:

Tenant: heuristic cold-start until ~20 confirmed churns, then real-outcome training.

Account: needs enough activity for a reliable baseline. Still refining the exact minimum.

Happy to dig into specifics.

[Vedansh @vedansh_builds]

most "health scores" are weights someone invented in a spreadsheet and never checked. churnsense trains a model on your own churn outcomes and names the 5 factors behind each at-risk account, so the slack alert is actually actionable: https://t.co/PsXhuguYBA

[Vedansh @vedansh_builds]

@DEVIN__SETH https://t.co/oZUB5P8y3T

[Vedansh @vedansh_builds]

@StevBuilds https://t.co/oZUB5P8y3T

[Vedansh @vedansh_builds]

@MalikAbdul94020 https://t.co/oZUB5P8y3T

[Vedansh @vedansh_builds]

@0xVibly https://t.co/oZUB5P8y3T

[Vedansh @vedansh_builds]

@edinsoncode https://t.co/oZUB5P8y3T

[Vedansh @vedansh_builds]

@itsalicesoul https://t.co/oZUB5P8y3T

[Vedansh @vedansh_builds]

@niqudrift https://t.co/oZUB5P8y3T

[Vedansh @vedansh_builds]

@nikhilx22 https://t.co/oZUB5P8y3T