Churn Prediction Model Scoping

Scope and design a churn prediction model from scratch, covering feature selection, modelling approach, evaluation metrics, and how outputs should inform business action.

Data AnalystClaudeCo-PilotChatGPTGeminiHighUpdated Mar-26

480·

Prompt

I want to build a churn prediction model. Please help me scope and design it end to end. Product/service context: Definition of churn: Prediction horizon: Available data: Current churn rate: Intended use of the model: Technical environment: Please produce a comprehensive model scoping document covering: 1. Problem framing — binary classification vs. survival analysis, and which is more appropriate here 2. Feature engineering recommendations — behavioural, transactional, demographic, and engagement signals to include 3. Data preparation requirements — label creation, train/test split strategy, handling class imbalance 4. Recommended modelling approaches ranked by complexity and suitability (e.g. logistic regression baseline, XGBoost, survival models) 5. Evaluation framework — which metrics matter most (precision, recall, AUC, lift) and why, given the business context 6. How to translate model scores into actionable segments (e.g. high / medium / low risk buckets) 7. Monitoring plan — how to detect model drift and when to retrain 8. A list of assumptions and risks that could undermine model performance Before starting, identify any gaps in the information I've provided and ask me to fill them.