Start with the ambiguous labels
Words such as traffic, leads, users, and visibility often hide incompatible definitions. List every dashboard card and ask what exactly increments it. Is traffic a server request, a session, or a search click? Does a lead mean a form attempt or an accepted enquiry? Who is excluded? The dictionary should resolve those questions before anyone interprets a trend. Google's user-metric documentation illustrates why even apparently simple audience labels deserve precision: different user measures serve different purposes and should not be substituted without explanation.
Give each metric a compact contract
For every metric, record its display name, source system, unit, filters, time basis, counting rule, owner, and known limitations. Add a plain-language question it can answer. For a rate, specify numerator and denominator separately and explain whether they share the same population. A conversion rate calculated from accepted accounts and all hosting requests is not a standard reader conversion measure. Keep formulas available next to the report so a reviewer can inspect the calculation without reverse-engineering the dashboard or relying on its creator's memory.
Record reporting scope and freshness
A metric should say whether it describes first acquisition, the current session, or a later event attribution perspective. Google's acquisition comparison provides useful context for this distinction. Include the last successful refresh and the most recent complete period. Distinguish missing data from zero. If a connector fails, keep a visible stale-data state rather than displaying yesterday's values as live results. These details are especially important when founders use a compact dashboard and may not open the underlying platform before making decisions.
Use a hypothetical definition dispute
Imagine sales reports twelve qualified opportunities while the content dashboard reports twenty leads. The discrepancy may disappear once the team sees that the dashboard counts accepted forms and sales counts opportunities passing a qualification review. Rename the cards, retain both, and show the progression. This example uses illustrative values only. The point is not to pick the larger or smaller number as correct; each measures a different stage. A dictionary turns the disagreement into a tractable workflow question about quality and follow-through.
Version the definitions
Assign a version or effective date when an event, filter, or qualification rule changes. Preserve the earlier definition in the report notes so historical comparisons remain honest. Review the dictionary when launching a new signup flow, changing analytics configuration, or adding a data source. Keep it short enough that editors and founders will use it. The goal is shared understanding, not a technical encyclopedia. A useful dashboard card should make its meaning and limitations clear before anyone asks whether the number improved.
Practical review checklist
- For each dashboard card, record the unit, source system, time basis, filters, counting rule, owner, and limitation.
- Specify both sides of every rate and confirm that its numerator and denominator describe compatible populations and periods.
- Show freshness and missing-data states so a failed connector cannot present stale values as a current business trend.
- Version definitions when event triggers or qualification rules change, preserving the meaning of historical reports for future readers.
Every dashboard metric needs a unit, scope, counting rule, freshness indicator, and owner.
Sources & further reading
Official documentation informs this guide. Examples are illustrative, and product interfaces or limits can change.
From search signals to a useful first draft.
SERP.so is in development. Join the waitlist for a workspace that connects research, content review, and publishing.
Join early access ↗