metric-definition

Define trustworthy business metrics with pinned definitions, edge-case rulings, and guardrail pairs.

1|Updated Jul 3, 2026
One-click install
npx skills add https://github.com/Nandansai08/skillz --skill metric-definition-nandansai08
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: metric-definition
Source: https://github.com/Nandansai08/skillz/tree/main/skills/data-analytics/metric-definition
Command: npx skills add https://github.com/Nandansai08/skillz --skill metric-definition-nandansai08

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Teams routinely fight over mismatched dashboards and ambiguous metrics like "active user" or "conversion rate." This Skill turns a vague metric idea into a single canonical definition with every noun pinned, edge cases legislated, gaming resistance built in, and one owned implementation. ## Core Features & Use Cases - Pinned metric sentence: Forces every metric into an explicit template covering aggregation, event, population, window, and time anchor so no noun stays vague. - Denominator and gaming red-team: Interrogates who can shrink the denominator, then pairs the target metric with a guardrail metric to prevent Goodhart-style gaming. - Canonical implementation and validation: Ships the definition in a metrics/semantic layer (dbt metrics, LookML, or one blessed SQL view) with an owner, versioning, 12-month backfill checks, and reconciliation against independent sources. - Use Case: When two dashboards disagree on "activation," use this Skill to arbitrate the definition, legislate edge cases like refunds and timezones, and ship one versioned definition both teams consume. ## Quick Start Ask the AI to define an activation metric for your PLG funnel, including the exact numerator and denominator, edge-case rulings, and a guardrail metric.

Frequently Asked Questions about metric-definition

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I define a metric that two teams won't fight over?▼

Write the metric as a sentence with every noun pinned: aggregation, precisely defined event, population, time window, and anchor. Then legislate edge cases like refunds, timezones, and late-arriving data in writing, and ship one canonical implementation in a metrics layer that all dashboards read.

What is a guardrail metric and why pair one with a KPI?▼

A guardrail metric is a paired measure that gaming the target metric would damage, such as reopen rate alongside tickets-closed. Shipping a KPI without its guardrail creates an incentive bug where teams optimize the number without creating real value.

Why do two dashboards show different values for the same metric?▼

Dashboards diverge when each re-implements the metric locally instead of reading one canonical definition from a semantic layer like dbt metrics or LookML. Unpinned nouns, undocumented exclusions, and silent definition changes also cause discrepancies.

When should I not use metric definition and use A/B test analysis instead?▼

Use metric definition when creating or arbitrating the metric itself so it is trustworthy. Use ab-test-analysis when you need experiment statistics on an already-defined metric, such as significance testing on an existing conversion rate.

How do I validate a new metric before shipping it?▼

Backfill 12 months and check the trend against known events like launches and outages, then reconcile against an independent source such as finance's revenue figures. Any unexplained gap means the metric is not measuring what you think.