Context

StreamWave uses a binary classification model to predict whether a subscriber will churn in the next 30 days so the retention team can send targeted offers. The current logistic regression model looks strong on overall accuracy, but the retention team says it is still missing too many customers who actually cancel.

Current Performance

Confusion Matrix

This evaluation is based on 10,000 subscribers from the most recent validation month.

The Problem

The VP of Growth wants to know whether accuracy is an appropriate headline metric for this use case and which metrics should guide model decisions instead. You need to interpret the current results, explain the tradeoffs among common classification metrics, and recommend how StreamWave should improve evaluation and model usage.

Requirements

1. Explain what each reported metric says about model performance.
2. Identify why the model can have high accuracy but still be weak for churn intervention.
3. Use the confusion matrix to quantify the business impact of false positives vs. false negatives.
4. Recommend which 2-3 metrics StreamWave should prioritize going forward and why.
5. Suggest concrete next steps to improve model evaluation and operational decision-making.

Constraints

• Retention budget allows offers to at most 900 users per month.
• Missing a true churner is estimated to cost $120 in lost contribution margin.
• Sending an unnecessary retention offer costs $8 per user.