Metric Catalog & Taxonomy — Starter Kit
A practical step-by-step guide, naming conventions, calculation-spec template, governance checklist, and an 8-metric example catalog to help teams name, define, and link metrics to outcomes and owners.
Welcome — Why a Metric Catalog Matters
If multiple teams measure the same idea differently, people argue about numbers instead of improving outcomes. A lightweight, shared metric catalog prevents that by making definitions, owners, calculations, and expected signals explicit. This starter kit helps you stop metric proliferation, align measures to outcomes, and create clear ownership so your organization can act with confidence.
What this Starter Kit Gives You
- Simple naming conventions and taxonomy to reduce duplicate metrics.
- Clear metric types (north-star, leading, lagging) and when to use each.
- An owner map pattern so every metric has a responsible steward.
- A calculation-spec template you can copy into your catalog.
- An example catalog with 8 practical metrics across product, ops, sales, and support.
- Governance checklist and rollout tips to keep the catalog healthy.
Start with a Hunger — Link Each Metric to an Outcome
Every metric in the catalog should answer: what decision or outcome does this support? Capture that connection in one sentence next to each metric. When a number changes, the team should be able to explain which decision or customer outcome it informs.
Naming Conventions & Taxonomy (Make duplication hard)
Use a predictable, human-friendly key for each metric. Consistency makes searching and governance easier.
- Structure: domain.component.metric_name_unit
Example: product.activation_rate_pct or ops.mttr_minutes - Lowercase and underscores: product_feature_xxx not ProductFeatureX.
- Include unit only when helpful: append _pct, _minutes, _count, _usd.
- Avoid synonyms in names: pick canonical words (use 'conversion' not 'signup_rate' and 'activation' both).
Keep a short tag set for the domain: product, ops, sales, support, finance, marketing. Use those as the first token to group metrics.
Metric Types — Choose the Right Signal
- North-star: the primary outcome-oriented measure you want to maximize over time (strategic).
- Leading: early indicators that predict future north-star movement (actionable).
- Lagging: outcome or result measures that confirm impact after the fact (diagnostic).
Prefer a small set of north-stars (1–3) and several leading indicators teams can act on quickly.
Metric Owner Map (Who looks after the metric?)
Assign at least one Owner (primary responsible) and one Steward (data/technical contact). Optionally include Interested Parties for visibility.
Owner responsibilities:
- Ensure the metric definition remains correct.
- Lead reviews each cadence (weekly/biweekly/monthly depending on cadence).
- Propose changes and communicate impacts to stakeholders.
Calculation Spec Template (copy into your catalog)
Use this template for every metric so anyone can reproduce the value.
| Field | What to record |
|---|---|
| Metric Key | product.activation_rate_pct |
| Display Name | Activation rate (%) |
| Type | Leading / North-star / Lagging |
| Outcome Linked | Retention at 30 days; higher activation should improve retention |
| Definition | Share the plain-English definition of the event or state measured. |
| Calculation | Numerator: users who completed X within 7 days; Denominator: users who signed up in period; Calculation: (Numerator/Denominator)*100 |
| Frequency | Daily / Weekly / Monthly |
| Target / Thresholds | Target: 25% (green >=25%, yellow 15–25%, red <15%) |
| Owner | Product Manager — Jane Doe |
| Data Source | Events DB: activation_event_v2; ETL: analytics_pipeline.activation_rollup |
| Notes / Caveats | Exclude test accounts; timezone: UTC; new tracking added 2025-03-01. |
Example Catalog — 8 Practical Metrics
Below are example entries you can adapt. Each row shows the key, type, brief definition, owner, and outcome link.
| Key | Type | Display | Owner | Outcome Link |
|---|---|---|---|---|
| product.daily_active_users_count | North-star | Daily Active Users (DAU) | Head of Product | Indicator of product engagement and monetization opportunity |
| product.feature_adoption_pct | Leading | Feature adoption (%) | Product Manager | Predicts retention if adoption increases |
| ops.mttr_minutes | Lagging | Mean Time to Repair (minutes) | Ops Manager | Reflects operational reliability and customer impact |
| ops.change_failure_rate_pct | Lagging | Change failure rate (%) | Release Engineer | Shows release quality and risk to uptime |
| sales.conversion_rate_pct | Leading | Sales conversion (%) | Sales Manager | Early signal for revenue pipeline health |
| sales.avg_deal_size_usd | Lagging | Average deal size (USD) | Sales Ops | Measures revenue quality and targeting effectiveness |
| support.first_response_minutes | Leading | First response time (minutes) | Support Manager | Better response predicts higher customer satisfaction |
| support.nps_score | Lagging | Net Promoter Score | Head of Support | Outcome measure of customer loyalty and referral likelihood |
Governance & Lifecycle — Keep the Catalog Healthy
- Adopt a review cadence: monthly for operational metrics, quarterly for strategic metrics.
- Change process: proposed edits submit a short rationale and expected impact; owner triages and schedules changes for the next review.
- Archive rather than delete: keep historical definitions with effective dates to avoid calculation surprises.
- Publish a canonical CSV or small web page and ensure dashboards reference the catalog keys (not ad-hoc SQL names).
Common Pitfalls & How to Avoid Them
- Too many metrics: teams track everything. Limit dashboards to the few signals that support current decisions.
- Overloading north-stars: a north-star should be outcome-focused, not a proxy for vanity metrics.
- Undefined calculations: undocumented SQL or ETL transformations cause disputes. Use the Calculation Spec template.
- No owner: metrics without owners drift or become stale. Assign a primary owner and a data steward.
How to Roll This Out (Practical Steps)
- Pick a pilot: select 2–3 teams and import their top metrics into a shared document using the Calculation Spec template.
- Run a short alignment workshop: confirm definitions, identify duplicates, assign owners.
- Publish the catalog and update one dashboard to use canonical keys. Validate values with owners for two cycles.
- Expand gradually and add governance once the pilot succeeds.
Next Improvements — Where interactivity helps
This guide is intentionally a compact reference. Consider converting the catalog into an interactive form where teams can submit metric records and the platform stores them centrally. That makes review, search, and lineage easier and enables automated dashboards to pull canonical metric keys.
Quick Checklist Before You Finish
- Does each metric have a single canonical key and clear owner?
- Is the calculation reproducible from the information provided?
- Is the metric linked to a decision or outcome it supports?
- Have you set a review cadence and change process?
Use this starter kit as a living resource: adapt the taxonomy, evolve your targets, and keep ownership explicit. When teams stop arguing about definitions and start using the same language and numbers, decision speed and quality improve.
Discussion
Comments and conversation will live here.