Metric Sensitivity & Variance Report

Analyze metric sensitivity for {{metric}}:

1. Variance components:
   - Day-to-day standard deviation (coefficient of variation)
   - Week-to-week variance (seasonality contribution)
   - Cohort sensitivity (how much does metric change by user segment, geography, device?)

2. Power analysis:
   - Minimum detectable effect (MDE) at 80% power, α=0.05
   - Required sample size for {{use_case}} experiment
   - Recommended observation window (how many days to get stable estimate?)

3. Volatility drivers:
   - Identify external factors (weekends, holidays, campaigns, releases)
   - Estimate unexplained variance

4. Recommendation:
   - Should this metric be stratified by user cohort?
   - Should we use {{metric}} as primary or guardrail in experiments?

Deliver as 1-page summary table.