Documented metrics are metrics with a written definition: what they measure, how they are calculated, which source they come from, which filters they use and what they are for. When people and AI systems read reports and dashboards, ambiguous definitions (“conversions”, “customers”, “return”) generate wrong interpretations. Documenting is a cheap way to increase confidence in and usefulness of analysis.
The essentials in a definition
Documented metrics are indicators with a written definition, formula, source and limitations, so that people and AI systems interpret them the same way. They follow the principle of helpful content, can rely on structured data when published and fit into dashboards such as Looker Studio.
Why ambiguity is costly
In the same company, “conversion” can mean a lead, a booking or a purchase; “new customer” can depend on the period; “return” may or may not include costs. Someone reading a dashboard without a definition draws different conclusions, and an AI assistant summarising the dashboard risks amplifying the error. Documentation also serves for auditing and training new colleagues.
The same applies to public content: clear definitions help get cited correctly.
What to document
For each metric: name, plain-language definition, formula, data source, filters and exclusions, time window, update frequency, owner and known limitations. Include examples of calculation and interpretation. Keep a single glossary, accessible and with change history, in the style of helpful content: clear, factual and updated.
If you publish metrics on public pages, use the same definition and, where it makes sense, appropriate structured data.
A metrics glossary people and machines understand
- List the main metrics
- Write definition, formula and source
- Show definitions in dashboards
- Review and control changes
Integrate into dashboards
Show the definition next to each indicator (tooltip, note or help page) in a dashboard such as Looker Studio. Use consistent names and avoid metrics with the same name and different meanings. If old versions exist, flag them as such.
Make definition changes go through approval and be recorded.
A metric without a definition is an opinion with numbers.
Review and maintain
Review the glossary quarterly and whenever measurement changes. Test clarity by asking someone outside the team to explain a metric based only on the definition. If an AI tool summarises reports, check whether summaries respect definitions.
Record discrepancies found and fix the source.
Decision table
| Definition element | Example | Why it matters |
|---|---|---|
| Name | New customers | Avoids confusing synonyms |
| Formula | Customers with no purchase in the previous 12 months | Makes the calculation reproducible |
| Source | Order database | Defines the reference |
| Limitations | Does not include in-store purchases | Avoids over-interpretation |
Common mistakes to avoid
- Two meanings for the same name. It generates arguments and wrong decisions.
- Changing the definition without a record. It breaks historical comparisons.
- Definitions that are too technical. Nobody reads them.
How to measure the result
Track the number of documented metrics, questions about definitions and discrepancies found.
Limits and exceptions
Documenting requires discipline and time, and not all tools make presenting definitions easy. Definitions evolve with the business. The value is in keeping them current and used.
Illustrative example
Illustrative example: a company finds that marketing and sales use “lead” with different meanings. It creates a glossary with definitions and examples, shows them in dashboards and records changes. Meetings start to discuss decisions, not numbers.
In practice: checkpoints
Ask a new team member to read the glossary and point out what they do not understand. Correct the definitions based on those doubts and publish the updated version. Mark which metrics are official and which are experimental, keep a changelog and link each definition to the dashboard where it appears, so that readers and automated assistants can find the exact meaning of any number they are shown.
Frequently asked questions
How many metrics should I document?
Start with those supporting decisions and reports to management.
Who maintains the glossary?
A responsible person, with input from teams.
Does this help with AI?
It helps summaries and answers use the right definitions, without guarantee.
How do I keep it updated?
Tie each measurement change to a glossary update.
Where should I store it?
In a single, accessible place with history.
Document main metrics with definition, formula, source and limitations, show them in dashboards and control changes, so people and AI interpret them the same way.
Sources and recommended reading
- Google Search Central: helpful, reliable, people-first content
- Schema.org: Organization
- Looker Studio for developers
Last editorial review: October 2026. Informational content; always confirm current conditions in the official sources before deciding.