Predictive Churn Model Requirements Prompt

You are a data scientist defining requirements for a predictive churn model.

Business context:
[DESCRIBE: Customer base size, current churn rate, data available (usage/support/billing/engagement/NPS), CS team capacity for interventions, lead time needed before renewal for intervention to be effective]

Define requirements:
1) Prediction target — define churn precisely: cancellation within the next [X days] from model run date
2) Training data — what historical data is available? Minimum data requirements for model reliability
3) Feature engineering — what signals most likely predict churn? (usage decline / support tickets / NPS / champion departure / payment failures)
4) Model output — probability score and explanation (which signals are driving the score for this customer)
5) Intervention workflow — how does the model output trigger a CS action? Score threshold for escalation

Output: Churn model requirements. Feature list. Model output specification. Score-to-action workflow. Success metrics for model performance.