metrics-treelisted

# Metrics Tree ## Goal Define how success will be measured inside the PRD before or alongside the tracking plan. ## Workflow 1. Start from the product goal. 2. Assign metric IDs such as `M1`, `M2`, and trace each metric to a goal or requirement. 3. Choose one primary metric and state why it is the best proxy for success. 4. Add secondary metrics that explain user behavior. 5. Add guardrail metrics for possible harm. 6. Add diagnostic metrics to explain movement. 7. Define metric formulas, data source, aggregation grain, measurement window, owner, and refresh cadence. 8. State baseline, target, or directional expectation when available; otherwise mark it as unknown rather than inventing a number. 9. Record measurement limitations, especially when no analytics taxonomy exists or when privacy rules intentionally exclude user-level or sensitive properties. 10. For dashboards, workload summaries, activity summaries, recency indicators, or other aggregate views, define the aggregation grain, permission filter, small-sample privacy risk, excluded detail fields, and whether users without detail access can see the aggregate. 11. For activity or recency summaries, prefer coarse states such as active/inactive buckets over exact timestamps or specific object names unless an approved privacy rule allows more detail. 12. For attribution, contribution, performance decomposition, or funnel breakdowns, define the attribution window, denominator, residual/unattributed bucket, missing-data