Define an Ops Metric with Formula, Owners, and Caveats

Full prompt

Document this operational metric for {{COMPANY_NAME}} so every team uses the same definition.

**Metric Name:** {{METRIC_NAME}}
**Owner:** {{OWNER_NAME}} ({{OWNER_SLACK}})
**Category:** {{CATEGORY: Revenue | Retention | Efficiency | Quality | Growth}}

**Definition & Formula:**
Write the exact calculation. Include numerator, denominator, and any filters (e.g., "exclude trials," "USD only").

**Data Source:**
- Primary source: {{SOURCE: Stripe | Segment | Mixpanel | Custom SQL}}
- Query/report name: {{QUERY_NAME}}
- Refresh cadence: {{CADENCE: Real-time | Daily | Weekly | Monthly}}
- Owner of data pipeline: {{DATA_OWNER}}

**Known Caveats & Failure Modes:**
List 3–5 scenarios where this metric is misleading or breaks:
1. {{CAVEAT_1}}
2. {{CAVEAT_2}}
3. {{CAVEAT_3}}

**Targets & Benchmarks:**
- Current value: {{CURRENT_VALUE}}
- Target: {{TARGET}}
- Industry benchmark (if known): {{BENCHMARK}}

**Related Metrics:**
List 2–3 metrics that move together or offset this one.

**Last Updated:** {{DATE}}
**Next Review:** {{REVIEW_DATE}}

How to use

1. Pick one metric your team argues about most (churn, CAC, NPS—whatever). 2. Spend 30 minutes filling this out with the metric owner and one data analyst. 3. Post the result in a shared wiki or Slack channel. 4. When someone uses a different definition, link them to this card. 5. Repeat for your next 5 most-debated metrics. 6. Review quarterly—metrics drift as your business changes.

Customization tips

• Add a 'Calculation Example' section with real numbers (e.g., "If we had 500 customers on Jan 1 and lost 16 by Feb 1, churn = 3.2%"). Kills ambiguity faster than prose.
• Include a 'Who Should NOT Use This Metric' subsection. E.g., "Don't use this to evaluate individual CSM performance—use cohort churn instead." Prevents misuse.
• Link to the actual query or dashboard. If it lives in Looker, Amplitude, or Tableau, embed the URL. Make it one click to verify the number.
• Add a 'When This Metric Changed' row: date, reason, and old vs. new definition. Prevents confusion when someone pulls historical data.
• Create a Slack bot command that posts the metric card on demand (e.g., `/metric churn`). Beats hunting through wikis.