Schema inference
Messy exports to a single readable report.
Churn diagnostic · For SaaS founders
Upload customer exports, inspect inferred schema coverage, and generate a ranked retention queue with calibrated probability — editorial clarity, not dashboard noise.
Capabilities
Messy exports to a single readable report.
Profile-specific models — selected automatically from what your file contains.
Explainable drivers per account so GTM and CS see why risk moved — not a black-box percentile.
Ranked retention queue, coverage warnings, and PDF summary.
Process
CSV from billing, product analytics, or a blended warehouse pull — no template ceremony.
Inspect inferred columns before compute. Gaps are surfaced explicitly, not silently imputed as signal.
Get churn probability, risk bands, and the top accounts to save — then download the report.
Signals
Canonical fields
Model profiles
Queue depth cap
Your CSV is processed in memory during the audit and deleted immediately after results are generated. We do not store, log, or retain any uploaded data.
You don't need to share emails or names. Replace them with anonymous IDs before uploading — the audit works just as well. We only need behavioural patterns, not personal information.
For enterprise audits, we sign an NDA before you send anything. Request one by emailing churniq.team@gmail.com and we'll send it within 24 hours.
Your data is never shared with third parties, used to train models, or accessed by anyone other than you. Each session is fully isolated.
Access
$0
1 audit per year · self-serve
$49/mo
2 audits per month · per seat
$500
one-off · human-reviewed
No slides — upload a file and walk away with a ranked queue.ChurnIQ keeps the surface area narrow on purpose.
Start auditLive audit
The API must be running locally (localhost:8000). Mapping preview hits /preview-mapping; audit runs /run-audit.
De-identified data is enough. You do not need to share emails, names, or any personally identifiable information.
Minimum required (3 columns)
Better predictions (add any of these)
Replace real emails with user_001@anon.com and real names with IDs before uploading if you prefer. The audit works just as well.