define-kpis
Skill4 starsChanged 7 days agoMITView on GitHub
Define or revise the company's metrics and KPIs using industry best practices — build the metric tree (north star, drivers, inputs), write rigorous catalog definitions, set targets and scorecard thresholds, and pair gameable metrics with guardrails.
In this repository21 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- analyzeSkill
- build-dashboardSkill
- build-modelSkill
- build-pipelineSkill
- causal-analysisSkill
- connect-dataSkill
- evalSkill
- experimentSkill
- forecastSkill
- health-checkSkill
- investigate-metricSkill
- make-deliverableSkill
- powerbiSkill
- privacy-reviewSkill
- research-domainSkill
- scorecardSkill
- setup-teamSkill
- tableauSkill
- triageSkill
- upgradeSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Ground in the business: Read knowledge/business-context.md (model, priorities, north star if set).
≈ “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 KPIs
Owner: metrics-steward. Args may scope it, e.g. /define-kpis marketing funnel or bare for the full company set.
Procedure
- Ground in the business. Read
knowledge/business-context.md(model, priorities, north star if set). If industry benchmarks would help set targets, run/research-domainfor current {{INDUSTRY}} benchmarks first — don't quote stale benchmark numbers from memory without checking. - Select the framework for the business model. SaaS → ARR/NRR/churn/CAC/LTV/activation; e-commerce → conversion/AOV/repeat rate/margin; marketplace → GMV/take rate/liquidity/two-sided retention; media → engagement/retention curves/ARPU; adapt for anything else. Within the framework, choose the vital few: 1 north star, 3–6 driver KPIs, supporting input metrics. Resist metric sprawl — every KPI must have a plausible owner and action.
- Stress-test each candidate KPI:
- Actionable? (someone can move it)
- Gameable? (then pair it with a counter-metric, mandatory)
- Computable honestly from available data? (check
knowledge/data-sources.md; if not, define the honest proxy and note the gap) - Leading or lagging? (the set needs both)
- Ratio metrics: are numerator and denominator individually visible somewhere? (ratios hide volume collapses)
- Write catalog entries in
knowledge/metrics-catalog.md, full template per entry: name, plain-English definition, exact formula, source table/model, grain, segments where valid, owner, target, scorecard thresholds (🟢/🟡/🔴), counter-metric, limitations, version + date. Definitions must be computable by a stranger from the entry alone. Give every entry aclassification(public/internal/confidential/restricted, as defined inknowledge/governance-policy.md) — who may see a number is part of its definition. Where the metric is reported for a standard population ("paying customers", "excluding internal accounts"), define that as a named segment in the entry (segments:plussegment_<name>:, with a/*SEGMENT*/marker ingeneric_sql) sopython scripts/run_metric.py "<metric>" --segment <name>applies the identical filter everywhere. Confirm withpython scripts/check_metrics.pyandpython scripts/check_governance.py. - Build/refresh the metric tree section of the catalog: north star → drivers → inputs, with one line on the causal logic of each link.
- Set targets and thresholds. Base on: history (trailing distribution), trajectory needed for stated priorities, and industry benchmarks. Where the user must make a judgement call (ambition level), present 2–3 options with implications and ask.
- Reconcile with existing reporting. Diff new definitions against what current dashboards/scorecards compute. Any change to an in-use definition is a breaking change: version it, quantify the restatement on the last 3 reported periods, list affected artifacts, and get user sign-off before switching.
- Record & propagate. Log decisions in
knowledge/decision-log.md. Update the scorecard KPI set. Hand affected dashboards todashboard-developerand affected marts toanalytics-engineerfor alignment.
Created by Colin Beck — https://www.linkedin.com/in/beckcolin/
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.

