Metrics defined
Every RankX AI metric is a ratio with a stated denominator. Mention rate, share of voice, Google position, AI Overviews, readiness and audit scores defined.
Every RankX AI metric is a ratio with a stated denominator, and the denominator decides what the number means. A metric with nothing to divide by is reported as unknown, never as zero.
Mention rate and share of voice divide by checks with a known verdict. Shopping visibility divides by answers that showed products at all.
This page defines each metric once, so the rest of the documentation can refer to it.
What are known checks and total checks?
Known checks are the checks that produced a definite verdict, and they are the denominator RankX AI uses for every visibility rate. They appear beside almost every figure.
- Total checks is every check that completed, read or not.
- Known checks is every check that produced a determined verdict.
A check that ran but could not be read counts in neither the numerator nor the denominator, and it is counted separately so the gap stays visible. If known checks is zero, every rate built on it is unknown, which is why a fresh Website shows blanks rather than zeros.
Mention rate
Mention rate is the share of known checks in which an AI assistant named your brand, reported per assistant over a 7, 30 or 90 day window.
How it is calculated. Mentioned checks divided by known checks, per assistant.
What it does not mean. It is not a share against competitors, and it is not one number across assistants. The assistants often disagree, so a 40% mention rate on one and 10% on another are two findings rather than an average of 25%.
When it is blank. No check in the window produced a verdict yet. Reporting that as 0% would tell you that you are invisible when nothing has been measured.
Share of voice
Share of voice is how often your brand is named in AI answers compared with the rivals named in those same answers. It is RankX AI's headline visibility figure, and it is comparative on purpose.
How it is calculated. Your mentions divided by known checks, shown beside the same figure for each rival brand found in the answers. Rivals are grouped by brand, so different spellings of one company count once.
Why comparative. "This rival appears in 21 of 25 answers and you appear in none" is actionable in a way a bare 0% is not, because it names who is being recommended instead.
Two values that mean opposite things. A blank share of voice means no verdict has been produced yet. A share of voice of 0% means RankX AI measured and your brand was absent from every answer it read.
The first is a gap in the sample; the second is the finding worth acting on, and RankX AI never shows them the same way.
Shown beside it. Your rank among the brands compared, how many brands were compared, how many assistants named each rival, and the rival named most often.
Position in an answer
Position in an answer is where your brand sat in an AI answer that contained a ranked list.
It exists only when the answer actually ranked things. Many AI answers are prose, and prose has no position, so a blank position means the answer had no ranking, not that you ranked last.
Google position
Google position is where your site ranks in ordinary Google results for a tracked keyword, checked to the top 20. Every keyword is in one of three states, which are never merged:
| State | Meaning |
|---|---|
| Ranked | Found, with a position from 1 to 20 |
| Not in top 20 | Checked, and not in the first 20 results |
| Not checked | No check has run for this keyword yet |
Merging the second and third into "no position" would turn an unmeasured keyword into a failing one. When a scheduled check skips a keyword, RankX AI shows why. See rank tracking.
AI Overview detection rate and verdicts
The AI Overview detection rate is the share of RankX AI's rank checks on which Google showed an AI Overview. It describes the query rather than your brand.
On each check where an overview appeared, RankX AI records two separate verdicts: whether your brand was named in the overview's text, and whether your site was cited as one of its sources. Each verdict can also be unknown, which means the overview was detected but not captured. Unknown is never counted as "not mentioned".
The data comes from the rank tracker, on tracked keywords, alongside the ordinary position check. It is not part of the AI visibility run. See AI Overviews.
Shopping visibility
Shopping visibility is how often your products appear when a shopper asks an AI assistant, per surface, over the last 28 days.
How it is calculated. Answers containing your products divided by answers that showed products at all. An answer with no product cards says nothing about your catalogue, so counting it against you would invent a loss.
When it is blank. No answer showed product cards in the window, so there is nothing to divide by. It is never shown as zero.
Shown beside it. Checks run, checks that produced an answer, checks that showed products, checks that could not run, and how many answers included your products, per surface. See AI Shopping.
AI Readiness score
The AI Readiness score is a 0 to 100 score for whether AI assistants can reach, read and understand your site. Two rules make it unusual:
The overall score is the weakest category, not the average. An average would let easy passes hide one real defect. The weakest category names the thing to fix, and the score always shows beside that category's name.
Below a coverage floor there is no score at all. Not 0, which reads as "your site is broken", and not 100, which reads as "nothing to do". The score always shows with its coverage beside it, and a check that could not run moves it in neither direction.
The checks RankX AI runs, with each check's category and weight, are in the AI Readiness check reference, which is generated from the product.
Website Audit score
The Website Audit score is a 0 to 100 score for the technical health of the crawled site.
How the penalty is weighted. Each issue carries a weight, and its penalty is a ratio of the right denominator for its class: pages crawled for a page issue, the sitemap for a sitemap issue, and the site itself for a site-wide one. A missing sitemap is one site-wide fact, so it is not diluted by the number of pages crawled.
Not assessed is not passed. A check that could not run is reported as not assessed. See what "a check did not run" means.
Content quality and GEO scores
Content quality and GEO scores are two scores on an article generated in Content Studio: one for the quality of the writing, and one for whether the piece is structured to be quoted and cited by AI assistants.
A blank score on a draft means it has not been computed yet, not that it scored zero.
Sentiment
Sentiment is how an AI answer speaks about your brand, as one of four outcomes: positive, negative, mixed or undetermined. RankX AI shows it on the AI Visibility page and on each prompt's page.
How it is decided. Only from words in the answer that are about your brand itself, quoted word for word. A sentence that praises a competitor, or that only lists your brand, does not count, and a verdict has to clear a confidence floor.
When it is undetermined. The answer did not say enough about your brand to support a verdict. Undetermined is a real result, and RankX AI never fills it with a guess or a "neutral" bucket.
Which figures does RankX AI leave out?
RankX AI leaves out two figures other tools publish, because neither can be measured without guessing:
- A single blended "AI visibility score" across the assistants. The per-assistant figures are the real ones, and a blend would hide the disagreement between assistants, which is often the most useful finding.
- Estimated traffic from AI answers. Where RankX AI reports traffic, it comes from your own Search Console and Google Analytics data, and is labelled as such.
Where to go next
- Null is not zero, for the rule every metric on this page follows.
- How RankX AI measures visibility, for where the checks come from.
- Glossary, for one-sentence definitions of every RankX AI term.
Questions
Why does a new Website show blanks instead of zeros?
Because nothing has been measured yet. RankX AI divides every rate by checks with a known verdict, and with no known checks the rate is unknown, which shows as a blank rather than 0%. See null is not zero.
Can I average mention rates across the assistants?
It is better not to. The assistants you switch on often disagree with each other, so a 40% rate on one and 10% on another are two findings, and an average of 25% hides both.
Does RankX AI estimate traffic from AI answers?
No. Traffic figures in RankX AI come from your own Google Search Console and Google Analytics data, and are labelled as such. See Traffic.
Last updated