Predictive Customer Analytics and Churn Prevention Model

Act as a predictive analytics consultant. Design a customer churn prediction and prevention system for [COMPANY NAME].

Context:
- Business model: [SUBSCRIPTION / SAAS / E-COMMERCE / SERVICE]
- Active customers: [NUMBER]
- Monthly churn rate: [PERCENTAGE]
- Available customer data: [LIST DATA POINTS — usage, billing, support, demographics]
- Analytics maturity: [BASIC REPORTING / ADVANCED ANALYTICS / PREDICTIVE]
- Available tools: [PYTHON / R / EXCEL / DEDICATED PLATFORM / NONE]
- Retention budget: R$ [AMOUNT]/month

Deliver the following:

1) **Churn Definition and Measurement**:
   - Operational definition of churn for this business
   - Churn rate calculation methodology
   - Voluntary vs. involuntary churn separation
   - Cohort-based churn analysis approach

2) **Predictive Feature Engineering**:
   - Identify 20+ predictive signals across categories:
     - Usage patterns (frequency, recency, duration, feature adoption)
     - Engagement signals (login frequency, content consumption, email opens)
     - Billing indicators (payment failures, plan changes, upgrade/downgrade)
     - Support interactions (ticket volume, complaint types, resolution time)
     - Behavioral changes (sudden drops, pattern deviations)
   - For each feature: data source, calculation method, predictive power estimate

3) **Model Design**:
   - Recommended modeling approach (logistic regression, random forest, gradient boosting) with justification
   - Training data requirements (historical period, sample size)
   - Model evaluation metrics (AUC-ROC, precision, recall, F1 score)
   - Threshold selection for risk classification
   - Model retraining frequency

4) **Risk Segmentation**:
   - 4-tier risk classification (low, medium, high, critical)
   - Characteristics of each risk segment
   - Size and revenue at risk per segment
   - Recommended intervention per risk level

5) **Automated Intervention System**:
   - Trigger-based retention actions:
     - Low risk: [ACTION]
     - Medium risk: [ACTION]
     - High risk: [ACTION]
     - Critical risk: [ACTION]
   - For each intervention: channel, message, offer, timing, success criteria
   - Escalation workflow for high-value customers

6) **Implementation Plan**:
   - No-code/low-code approach for teams without data scientists
   - Advanced approach for teams with data science capability
   - Integration with existing marketing automation tools
   - Testing and validation process

7) **Performance Monitoring**:
   - Model accuracy tracking over time
   - Intervention effectiveness measurement
   - ROI calculation (retained revenue vs. intervention cost)
   - Monthly review process and model improvement cycle