Customer Churn Early Warning System

Act as a customer retention analyst for small-business subscription models. Build a churn prevention system from my data.

Business: type [YOUR_BUSINESS_TYPE] · average customer lifetime [MONTHS] · monthly churn [PERCENTAGE or "not sure"] · pricing [monthly / annual / usage-based]
What I can actually offer to keep a customer [DISCOUNT, PAUSE, DOWNGRADE, CALL, EXTRA HELP, or NOTHING]
Data (customers who cancelled and, for comparison, customers who stayed):
[PASTE CUSTOMER/CANCELLATION DATA]

Deliver:
1. Churn pattern analysis — what cancelled customers share: timing, usage, plan, acquisition channel, support history
2. Risk scoring framework — a 1-5 score for every active customer; each factor and its weight
3. Early warning checklist — the 5 behavioral signals that predict churn 30-60 days out
4. Save-campaign playbook for scores 3, 4, 5 — when to send, channel, offer, word-for-word message templates

A trait counts as a churn signal only when it is clearly more common among customers who cancelled than among those who stayed; show both counts. A trait every customer shares predicts nothing, so list fewer than five signals in step 3 if fewer qualify. If my data holds no customers who stayed, say so at the top, label every pattern "unconfirmed: no comparison group", and tell me which rows to add. Save offers in step 4 come only from what I said I can offer; with nothing listed, the templates ask and listen and make no offer.

Success means every pattern in step 1 is supported by rows in my data (cite them), and the scoring weights reflect the strength of those patterns rather than textbook defaults.

Apply every rule, limit, and format above to every item in the deliverable, not just the first one. Deliver the complete result in one response — every section filled in, nothing deferred to a follow-up.

If a required input is missing, ask for it once before starting; otherwise proceed and label any assumption you made.

Notes: R40 refresh; tuned to vendor prompting guidance