Expertise · Measurement
Measure enough to choose what comes next.
Useful measurement does not search for one master number. It connects every indicator to its source, scope and the decision it can actually inform.
What is measured
Separate data universes.
Google performance, a brand mention in an AI answer, a cited link, a visit and a commercial enquiry do not combine automatically. We show totals, subsets and unclassified data without renaming them.
Measurement helps decide which page to deepen, which problem to fix or which hypothesis to check. It does not turn correlation into attribution.
/guides/example1846 2402,9 %/services/example963 8802,5 %/contact345406,3 %Scope · June 2026
Measurement material
Reporting that keeps its limits.
Every view names its source, period and coverage.
Google reading
Clicks, impressions, queries and pages with explicit filters and scope.
AI observations
Dated question sample with mentions, citations and cited pages.
Website journey
Available visits and actions, subject to consent and real configuration.
Proposed decisions
Pages to inspect, hypotheses to test and evidence still required.
Coverage, stated clearly
Missing and unclassified data remain visible in the analysis.
Priority pages updated
At six checkpoints, the curve shows the share of 100 fictional priority pages marked as updated. This measure is separate from the AI answer sample.
100 priority pages on example.com · 1–30 June 2026
Illustrative example · fictional data
A report that separates the evidence.
Three lines, three different questions.
A report places Google results, a sample of AI observations and identified visits beside each other. It does not add them together: each line supports a different decision.
- ✓ A query-filtered subset is not total traffic.
- ✓ An AI citation is neither a click nor revenue.
- ✓ A referring visit alone does not prove the cause of a conversion.
- ✓ Missing and unclassified data remain visible in the analysis.
Three sources, three decisions
The rows use different denominators. They are not added together.
From signal to decision.
Measurement should answer a decision.
- 01
Define
Question, source, period, segment and decision owner.
- 02
Check
Quality, gaps, duplicates and method changes.
- 03
Read
Totals, segments and observations without improper blending.
- 04
Decide
Next action and the criterion for reviewing it.
Questions before interpreting
Keep the scope of an observation in view.
Will you access all our data?+
Only the necessary, authorised sources. Access scope, responsibilities and missing evidence are established before analysis.
Can revenue be attributed to an AI answer?+
Only when real journeys and events support a defensible attribution. A citation or isolated visit is not enough.
Do you offer a GEO score?+
We favour interpretable observations. A score helps only when its definition, sample and limits are transparent.
What does reporting contain?+
Sources, periods, definitions, available results, limits and agreed next decisions. Its form follows the real scope.
Your evidence
Start with the decision you need to inform.
Tell us what you are trying to understand and which sources exist. We will frame a proportionate reading.
Request an audit ↗