See who's leaving
before they're gone.
Drop in a CSV of your customers. In seconds, ChurnLens shows you who's about to churn, the reason behind each one, and exactly what to do — trained right in your browser.
Nothing is uploaded. The model runs entirely in your browser.
- 97%Acme CorpSupport calls (6) — 3× the retained average
- 91%Globex LtdOn international plan — churns at 4.1× the base rate
- 84%InitechHeavy day usage, low engagement
Do this first — 23 at-risk customers contact support often; this group churns at 3.6× the base rate. Reach out before they cancel.
Three questions, answered
Most tools stop at a score.
A red dashboard tells you something is wrong. It doesn't tell you what, or what to do. ChurnLens answers all three.
Who
Every active customer scored 0–100% on churn probability, ranked so you act on the riskiest first.
Why
Each at-risk customer comes with the specific, human-readable factors driving their risk.
What to do
We cluster the at-risk into cohorts and tell you the one retention play that moves each one.
The difference
Enterprise insight,
without the enterprise.
The incumbents cost $1,000–$180,000 a year, need a dedicated success team, and take weeks to set up. ChurnLens gives a small team the same core insight in under a minute — and never sees your data.
Private by design. The model runs in your browser. Raw customer data never leaves it.
Explainable. Every prediction carries the human-readable reasons behind it.
Actionable. Cohort-level plays, not just risk scores.
Free to try. No signup to run a full analysis.
The model, in plain terms
- 1
Balance the data
Churners are rare, so we synthesise realistic examples (SMOTE) so the model learns them properly.
- 2
Boost the trees
Gradient boosting — the family that won our benchmark at 94% — finds the patterns.
- 3
Score & explain
Honest accuracy on held-out data, then a reason for every customer.
Find your at-risk revenue
in sixty seconds.
One point of monthly churn reduction is worth ~$100k a year on a $10M ARR business. Start with the customers most likely to leave.