The Scenario
A subscription company has 50,000 customers. 8% churned last quarter. The CEO wants a model that predicts who will churn next quarter so the retention team can intervene.
The Brief
Describe (or code) how you would build a churn prediction model. Focus on feature selection, model choice, evaluation, and the business implications of false positives vs false negatives.
Deliverables
- Feature selection: 5-7 features you would use and why each predicts churn
- Model choice: the algorithm you would use and why (logistic regression, random forest, XGBoost, etc.)
- Evaluation: explain precision, recall, and the confusion matrix in the context of churn prediction
- Business impact: should the model optimize for precision or recall, and what is the cost of getting it wrong?
Submission Guidance
A model with 92% accuracy sounds great until you realise 92% of customers do not churn anyway. Show you understand class imbalance.
Submit Your Work
Your submission is graded against the rubric on the right. If you pass, you get a public Badge URL you can share on LinkedIn. There is no draft save, so work offline first and paste your finished response here.