Metric Variance Analyzer

Analyze variance in {{metric_name}} over {{analysis_period}}. Current daily range: {{min_value}} to {{max_value}}.

Decompose variance into:
1. Day-of-week effects (which days are consistently high/low?)
2. Cohort effects (which user segments drive variance?)
3. Random/noise (unexplained variation)
4. Trend (is the metric drifting?)
5. External events (known campaigns, launches, holidays)

For each component:
- Quantify contribution to total variance
- Provide forecasting guidance (adjust for day-of-week in targets?)
- Recommend {{analysis_goal}} baseline or smoothing approach