Context

At SparkFit, a subscription fitness app, the growth team uses a binary classifier to predict which free users are likely to convert to paid within 30 days so they can target discount offers. The model reports very high overall accuracy, but campaign lift has been disappointing and the paid marketing budget is being wasted on the wrong users.

Current Performance

The Problem

Leadership sees 96.2% accuracy and assumes the model is strong. However, only a small fraction of actual future converters are being identified, which limits incremental revenue in a growth use case where missing high-intent users is expensive.

Task

1. Explain why 96.2% accuracy can still indicate a poor model for this use case.
2. Use the metrics and class imbalance to diagnose what is happening.
3. Describe which metrics matter more than accuracy for targeted growth campaigns and why.
4. Recommend how you would adjust evaluation, thresholding, and model selection.
5. Discuss the tradeoff between sending more offers and preserving marketing efficiency.

Constraints

• Discount offers cost $4 per targeted user.
• A converted paid subscriber generates $48 in 90-day gross profit.
• The CRM team can send at most 25,000 offers per month.
• Over-targeting low-intent users reduces brand trust and future campaign response.