Audit intermediate Works with: Any AI

KPI Definition Audit

Table output Has example

• 📊 Excel
• 📈 Google Sheets

Check that each metric is defined, owned and comparable.

The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.

[ADD SOURCE] is an editorial flag, not a blank to fill. Replace or resolve each flag before you use the result.

Audit our KPI list. Each KPI with its formula, source, owner, target and frequency: [KPI_LIST]. Decisions the KPIs are meant to inform: [DECISIONS]. Disagreements about numbers: [DISAGREEMENTS].

Check whether definitions are unambiguous, sources are single and traceable, targets have reasons, owners can influence the metric, and KPIs conflict with each other. Find vanity metrics and metrics with no decision attached.

Return a table with KPI, Problem, Severity (red, amber, green) and Fix, a score from 1 to 10 for clarity and usefulness, and a proposed data dictionary format.

Also identify which KPIs are leading and which lagging, which can be gamed, and which should be retired. Suggest a review date for each KPI and the person who can change the definition, so that history stays comparable.

Also write one sample data dictionary entry in full for the most important KPI, including formula, source, owner, refresh time, known limitations and the decision it informs. Suggest how to retire a KPI politely when it no longer serves, and how to keep historical comparisons honest when definitions change.

Then design the measurement framework that results from the audit. Define a small set of leading and lagging indicators linked to goals, with a one-sentence purpose, formula, data source, owner, frequency, target logic and the decision each informs. Produce a data dictionary template and fill one entry as an example using my information. Describe a governance process for changing definitions, with a version history and notice to readers. Propose a monthly review format of forty minutes in which each owner explains movements and agrees actions, with a rule that metrics which never prompt action are reviewed for removal. Add a section on gaming risks, such as targets that encourage cutting corners, and countermeasures like paired metrics. Suggest how to present KPIs to different audiences without changing the numbers. Avoid inventing benchmarks; mark any target without a basis [ADD SOURCE].

Fill in before use

Optional: type your details and the prompt above updates. Nothing is saved or sent.

Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.

KPI table (invented)
"Active users": formula unclear; marketing counts logins in 30 days, product counts actions in 7 days. Issue red. Fix: one definition.
"Churn": calculated monthly from cancellations over start-of-month customers; excludes downgrades. Amber.
"NPS": response rate 4%. Amber (small sample).
"Pages published": owner can influence; no decision attached. Vanity. Retire.
"Time to first value": new, no baseline.
Data dictionary entry (sample): Churn rate = customers who cancelled in the month / customers at month start; source billing system; owner finance; monthly; limitation excludes pauses.
Score for clarity 4; usefulness 5.

• audit