How scoring works

GEO Readiness asks whether your page is prepared for AI engines — can they reach it, parse it, and identify who you are. It is computed from the audited page on every run, so it is always available.

GEO asks whether the engines actually know and cite you. It comes from the GEO tools you run, so it only exists once you have run them.

A site can be perfectly readable and still invisible, or well known and badly structured. Blending the two into one number hides which of those you are, which is why they are reported separately.

25 checkpoints across four groups. Each is pass, warning or fail, weighted by impact (high, medium, low), with warnings earning partial credit that is stricter on high-impact checks.

A group where most checks do not apply to your page type is marked not measured and left out of the total rather than scored as zero. The remaining groups are re-weighted so the score still reads out of 100.

If less than half the model could be measured, no score is shown at all. Reporting a confident number from a quarter of the checks would overstate what we know.

13 checkpoints, scored the same way, across three groups.

Weighted across the tools that have produced a result for your site. Visibility Monitor and Prompt Coverage carry the most weight because they are the two that put real prompts to real engines; the others are a step removed from whether an engine actually cites you.

Tools you have not run are excluded, not zeroed. Two tools scoring 90 give you 90, not 26 — you are not penalised for tools you chose not to run.

A score appears once at least 40% of the weight has been measured. Below that you will see how many tools have run instead of a number: one probe is not a visibility measurement.

Results older than 90 days are treated as stale and excluded. These tools read live engine behaviour, so an old run describes a web that has moved on.

Bands: under 35 poor · 35–59 average · 60–79 good · 80+ excellent.

Some tools are deliberately left out, and it is worth saying why rather than leaving gaps unexplained.

• Competitor Citations and AI List Inclusion produce findings rather than a single score. Inventing one to make them fit would be a number we made up.
• GA4 Traffic Attribution measures visits that arrived, which is an outcome of visibility rather than a measure of it.
• Fact Density does not currently store a result that can be read back after the fact.

30% GEO Readiness, 25% AEO Readiness, 25% On-Page SEO, 20% Technical SEO. Any section that could not be measured is excluded and the rest re-weighted, so the combined score is never propped up by a section that never ran.