define-metrics
Skill1 starsChanged 3 months agoMITView on GitHub
Define product metrics precisely — north star, input metrics and guardrails, exact formulas with denominators and segments, named funnels, and honest instrumentation gaps. Use before measuring a launch or OKR, or when two teams disagree about what a number means. Dispatches the product-analyst.
In this repository30 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- build-business-caseSkill
- build-roadmapSkill
- capture-feedbackSkill
- close-projectSkill
- coordinate-programSkill
- define-strategySkill
- design-dashboardSkill
- elicit-requirementsSkill
- groom-backlogSkill
- make-deliverableSkill
- manage-changeSkill
- manage-resourcesSkill
- plan-capacitySkill
- plan-launchSkill
- plan-portfolio-automationSkill
- plan-projectSkill
- plan-sprintSkill
- powerbiSkill
- prioritizeSkill
- review-okrsSkill
- review-portfolio-intakeSkill
- run-ceremonySkill
- run-discoverySkill
- run-gate-reviewSkill
- setup-pmoSkill
- track-financialsSkill
- track-portfolioSkill
- track-statusSkill
- write-prdSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 9
Before a launch or an OKR cycle needs measuring.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
/define-metrics — Make the numbers mean one thing
When to use
Before a launch or an OKR cycle needs measuring; when a metric is being quoted with two different values; when an experiment needs a primary metric and guardrails agreed before it starts.
Dispatches
product-analyst (lead) + product-strategist (which outcomes matter) + product-manager (what the feature
is supposed to change).
Inputs
knowledge/metrics.md (existing catalog), knowledge/product-context.md (goals/OKRs),
knowledge/prds/, launch plans in knowledge/launches/, knowledge/portfolio-measures.md (to avoid
colliding with a delivery measure).
Steps
- Tie every metric to a goal or OKR. A metric that serves nothing gets deleted, not defined — the catalog is only useful if it stays small.
- Propose the north star with its guardrails. A single optimizing metric without counter-metrics reliably produces the wrong behavior; name what must not degrade while it improves.
- Define precisely enough that a stranger could compute it — the event, the numerator and denominator, the filters, the period, and which segments the metric is and is not valid across. Ambiguity here is what produces two numbers for one name.
- Name the funnels. A "conversion rate" means nothing until the step sequence and window are fixed.
- Say measured or estimated, and name the limitations — what it excludes, where it misleads, what instrumentation it depends on.
- Check the boundary. Product metrics live here; delivery measures (SPI, CPI, milestone hit rate) live in
knowledge/portfolio-measures.md. Never define the same thing in both — if a dashboard needs a product metric, it cites this catalog rather than restating it. - Record instrumentation gaps honestly. Metrics you want but cannot currently measure belong in the gaps table, with the proxy in use if any. Being explicit here is what stops a proxy quietly becoming the truth.
Methods
knowledge/methods/metrics-and-experimentation.md (north star, metrics tree, hygiene, experiment design),
knowledge/methods/product-strategy.md (outcomes and OKRs),
knowledge/methods/lean-product-process.md (build-measure-learn).
Output
Updated knowledge/metrics.md — north star and input metrics, full definitions, funnels, experiment index,
instrumentation gaps. Reflect success measures into knowledge/product-context.md. Changing a definition
already in use is a breaking change: version it and log it in knowledge/decision-log.md. Follow
standards/document-standards.md.
Created by Colin Beck<br>
LinkedIn: https://www.linkedin.com/in/beckcolin/<br>
GitHub: https://github.com/link7373
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

