daddia Metrics Glossary
When to use
Single source-of-truth glossary — exact formula, source, owner, cadence — so two people calculate the same number.
What this skill does not do
- Does not build KPI trees — route to
/performance:kpi-tree-builder. - Does not pick ambiguous formulas silently — flags for clarification.
- Does not duplicate OKR instrumentation —
okr:instrument-metricsdefers here when performance installed.
Preconditions
| Input | If missing |
|---|---|
| Metrics to document | Ask user to list |
| Performance practice profile | Proceed; flag [PROVISIONAL] on category codes |
Provisional mode
Ambiguous formulas: AMBIGUITY FLAGS — do not finalize entry until clarified.
Trust spine
- Confidence bands (
structured-aggregation):- High: Every metric has exact formula, source, owner, cadence, caveats.
- Medium: Some ambiguities or near-duplicates flagged.
- Low: Formulas described not calculated — halt until literal formula provided.
- Failure modes:
- Strategic advice vs. support: Glossary is draft for metric governance approval.
- Client confidentiality: Definitions may expose internal systems — CONFIDENTIAL header.
- Accountability gap: Owner named per metric.
- Analytical Rigor: Completeness check; near-duplicate consolidation flagged.
- Incentive Gaming: N/A — definition focus.
- Escalation triggers: Two metrics calculating same thing under different names — flag consolidation.
Workflow
- Read practice profile for category codes and known metrics.
- Per metric: name, code, exact formula, source, owner, refresh, caveats.
- Flag ambiguous formulas — ask, don't pick silently.
- Check near-duplicates for consolidation.
- Completeness check before output.
Output format
CONFIDENCE: [defensible recommendation | structured first pass]
METRIC: [name] [category code]
Formula: [exact calculation]
Source: [system/sheet/table]
Owner: [name/role]
Refresh: [cadence]
Caveats: [...]
AMBIGUITY FLAGS: [...]
NEAR-DUPLICATE FLAGS: [...]
Worked example
Input: "Conversion rate" — user says "signups divided by visitors."
Expected output (excerpt):
METRIC: Conversion rate [B3]
Formula: AMBIGUITY FLAG — admits multiple methods (unique visitors vs sessions; signup completion definition) [review]
Quality checks before delivering
- Literal formulas, not descriptions
- Ambiguities flagged not silently resolved
- Near-duplicates checked
- Owner and source per metric
- Every metric in scope covered
Propose profile update
When a stable convention surfaces during this run (thresholds, naming, tone, output format, or recurring corrections), propose a profile update: show the exact diff against ~/.claude/plugins/config/claude-for-strategy/performance/CLAUDE.md (org-wide facts go to org-profile.md), ask for confirmation, and write only on yes. Only /performance:practice-setup auto-applies a full profile write.
Outputs
Follows plugin CLAUDE.md § Outputs. Next: governance approval, tracker-builder wiring, or okr:instrument-metrics handoff.