Skip to content

Metrics glossary

Every number in the product has an ⓘ that explains it. This page collects those explanations in one place. It is generated from the same definitions, so the two always agree.

AEO issues

Answer-engine problems found on crawled pages — missing schema, unreachable content, stale pages.

What it is not Flags from measured signals, not a ranked to-do list.

Source Page crawl · checks + microdata

AEO score

A composite 0–100 over four measured dimensions: citation rate, citability, schema and technical access.

What it is not Our weighting, not a third-party metric. A dimension we could not measure is excluded and the rest renormalise — it is never scored as zero.

Source Our policy · fixed weights over the four measured dimensions

AI visibility

The share of measured AI answers that cited your domain or named your brand.

What it is not A share of the answers we could measure, not of all AI traffic, and not a share against competitors — adding a rival cannot move it.

Source Measured live · citations and brand-name matches across the tested prompt set

Answered

Of the prompts asked, how many this engine returned something measurable for.

What it is not Zero here means the engine did not answer — an outage, a rate limit, or a surface that was not shown. It does NOT mean the engine never cites you.

Source Measured live · answers with prose or at least one source

Answers measured

Answers that came back with text or a citation — the denominator behind every rate here.

What it is not Not the number attempted. An engine that returned nothing is excluded rather than scored as a miss, so this is usually lower than responses.

Source Measured live · answers with prose or at least one source

Asked

Prompts put to this engine in the run.

What it is not What we sent, not what came back. Read it against Answered — the gap is the engine's coverage, not your visibility.

Source Measured live · one request per question

Average position

Where your domain typically appears among an answer's cited sources — 1 is first.

What it is not Position within an AI answer's sources, not a Google ranking.

Source Measured live · index of your domain in each answer's citation list

Brand mentions

AI answers that named your brand in the text, whether or not they linked you.

What it is not Name-drops, not links. Read it next to citations — the gap between them is the opportunity.

Source Measured live · brand name matched in each answer's text

Citability

How well a page is structured for an answer engine to quote it, from measured signals only.

What it is not Our ranking of measured facts — headings, schema, freshness, reading ease — not a third-party metric and not a prediction of citation.

Source Our policy, over page-crawl microdata, headings and last-modified dates

Citation rate

The share of the AI answers we measured that cited your domain as a source.

What it is not Out of the questions we tested, not out of all AI traffic. Change the questions and this changes.

Source Measured live · AI answers across the tested prompt set

Citations

AI answers that linked your domain as a source.

What it is not Links, not name-drops. An answer that praises you by name without linking counts under brand mentions, not here.

Source Measured live · citation list of each answer

Crawler access

How many of the four crawler-facing files and rules are in place: robots.txt, sitemap.xml, llms.txt, and no AI-crawler block.

What it is not Four yes/no checks we run ourselves, not a third-party score.

Source Our policy, over direct fetches of robots.txt, sitemap.xml and llms.txt

Engines

How many AI engines returned a usable answer for this run.

What it is not Engines that answered — an engine that failed is excluded, not counted as zero visibility.

Source Measured live · engines with at least one response

Mentions

How many indexed AI answers reference this brand or domain.

What it is not A count of answers in an index — not how many people saw them.

Source AI answer index · mentions

Pages with schema

Crawled pages carrying JSON-LD structured data.

What it is not Presence, not correctness — a page with invalid schema still counts.

Source Page crawl · microdata

Questions

How many distinct questions this run graded the site against.

What it is not The tested set, sourced from measured AI demand. Change the set and every percentage here changes with it — which is why the set is listed on the page.

Source Measured AI prompt demand for this domain and market

Readiness score

Average citability across the pages we scored.

What it is not An average of our own citability ranking, not a measured outcome.

Source Our policy, averaged over scored pages

Responses

How many AI answers were measured in this run — the denominator behind every rate on this page.

What it is not The sample size. Every percentage here is out of this number, not out of all AI traffic.

Source Measured live · one per (question × engine)

Schema coverage

The share of crawled pages carrying JSON-LD structured data.

What it is not Measures presence, not correctness. A page with wrong schema counts as covered.

Source Page crawl · microdata

Seed queries

The questions we put to the AI engines for this project.

What it is not The tested set, sourced from measured demand. Change it and every rate on this page changes with it.

Source Measured AI prompt demand + People Also Ask phrasings

Share of voice

Your mentions as a share of the total across you and the competitors being compared.

What it is not Relative to the competitors you chose. Adding or removing one changes it without anything else changing.

Source AI answer index · share across compared targets

Standing

Answers where this domain was cited or named, as a share of the same measured answers used for every other domain here.

What it is not One denominator for everyone, so the rows are directly comparable. Not a share of a shared total — two domains can both be high.

Source Measured live · citations and domain-label matches across the tested prompt set

Topic clusters

Groups of related questions this project is tracked against.

Source Our clustering of the tracked question set

12-month trend

Monthly search volume across the last year, so seasonality and decline are visible.

What it is not Not your traffic — this is how often the phrase is searched by anyone. A flat line at low volume is not the same as no data; a gap is a month that was not reported.

Source Keyword index · monthly search volume

Change

Positions gained since the previous check. Positive means moved up the page.

What it is not Blank when there is nothing to compare — a first check, or a keyword unranked on either end. Never rendered as 0.

Source Our comparison of the two most recent checks

Checked

When this position was last measured.

What it is not Positions drift between checks; an old reading is not today's rank.

Source Time of the most recent check

Competition

How contested this keyword is among ADVERTISERS, from 0 to 1.

What it is not Not organic difficulty. A keyword can be 1.0 here and trivial to rank for — use keyword difficulty for that.

Source Google Ads · advertiser competition

Competitors

Domains tracked as competitors for this project.

What it is not The list being compared against — confirmed or auto-discovered, not every rival that exists.

Source Your project's competitor list

Content gaps

Keywords a tracked competitor ranks for and you do not.

What it is not Relative to the competitors on your list. It is not every keyword you are missing.

Source Search index · rankings compared across domains

CPC

Average price advertisers pay per click on this keyword.

What it is not A paid-search price. It says nothing about the organic value of the keyword.

Source Google Ads · cost per click

Critical issues

Issues we classify as blocking — a page that cannot be indexed, titled or read.

What it is not Our severity banding over measured checks, not the crawler's.

Source Our policy, over page-crawl checks

Declined

Keywords that moved DOWN since the previous index, plus ones that dropped out entirely.

What it is not Not necessarily your doing — a competitor publishing, or the SERP changing shape, moves these too.

Source Search index · keywords that moved down or dropped out

Domain rank

A 0–1000 score for a domain's overall backlink strength.

What it is not Not Google's PageRank, and not comparable to Ahrefs DR or Moz DA.

Source Backlink index · domain strength score

Improved

Keywords that moved UP since the previous index, plus ones newly ranking.

What it is not Not a traffic change. A keyword can climb from 80 to 60 and bring nobody — movement says the direction, not the value.

Source Search index · keywords that moved up or newly rank

Intent

What a searcher is trying to do: informational, navigational, commercial or transactional.

What it is not A model's classification of the query, not a measured behaviour.

Source Search index · classified search intent

Keyword difficulty

A 0–100 estimate of how hard it is to break into the top 10 for this keyword.

What it is not An estimate from the current top 10's backlink profiles — not a prediction that you will or will not rank.

Source Search index · keyword difficulty

Organic keywords

How many distinct keywords this domain ranks for in the top 100.

What it is not Counts rankings, not pages — one page can hold hundreds of these.

Source Search index · count of top-100 rankings

Organic traffic

Estimated monthly visits from the keywords this domain ranks for in Google's top 100.

What it is not An estimate from ranking positions — not measured traffic. Search Console clicks are the measured figure.

Source Search index · traffic modelled from ranking positions

Position

Where this URL ranked in Google's organic results when the SERP was last checked.

What it is not A snapshot for one location and language — not a personalised or live rank.

Source Search index · absolute position on the last checked SERP

Search demand

Monthly searches for this question where an engine answers with AI. For Google AI Overviews it is the Google search volume of the query; for ChatGPT it is estimated from People Also Ask.

What it is not A count of prompts typed into AI assistants — no engine publishes that. This is the closest measured proxy, and it is derived from Google search data.

Source Search demand index · derived from Google search data

Search volume

Average monthly searches for this keyword over the last 12 months.

What it is not Total searches by everyone, not visits you would get — and rounded into buckets by Google Ads.

Source Google Ads · 12-month average search volume

SEO issues

Technical problems found on crawled pages, from the page crawler's own checks.

What it is not Flags raised by measured checks — not a ranked to-do list, and not weighted by impact.

Source Page crawl · checks

SEO visibility

A 0–100 index of where your ranked keywords sit. 100 means every one of them is in the top 3.

What it is not Our weighting of measured position counts — not a third-party metric, and not traffic. It moves when rankings move, not when visits do.

Source Our policy (CTR-like weights) over the search index's position distribution

SERP features

Which non-organic blocks appeared on the page — AI Overview, snippet, PAA, and so on.

What it is not What was on the SERP, not what you own. An AI Overview above you is still a feature.

Source Live SERP · item types

Top-3 positions

How many of your ranked keywords sit in positions 1–3.

What it is not A count of keywords, not of pages or of visits.

Source Search index · rankings in positions 1–3

Traffic value

What your organic traffic would cost to buy as ads, at each keyword's CPC.

What it is not A price tag on the traffic, not revenue and not what the traffic is worth to you.

Source Search index · estimated traffic × cost per click

URL

The page of yours that ranks for this keyword.

What it is not Google's choice, not yours — a surprise here is usually the finding.

Source Live SERP · the matched organic result

Avg. position

The mean position of your best result, weighted by impressions.

What it is not An average across every impression, not where you rank today.

Source Google Search Console · searchAnalytics · position

Clicks

How many times someone clicked through to your site from Google.

What it is not A count Google made, not an estimate — this is the measured figure.

Source Google Search Console · searchAnalytics · clicks

CTR

Clicks divided by impressions, as Google reports it.

What it is not Ranking higher does not by itself raise this; the snippet does.

Source Google Search Console · searchAnalytics · ctr

Impressions

How many times a result for your site appeared in Google.

What it is not Appearing is not being seen; a result on page 4 still counts.

Source Google Search Console · searchAnalytics · impressions

Crawl depth

How many clicks from the homepage it takes to reach this page.

What it is not Distance in links, not in URL folders.

Source Page crawl · click depth

On-page score

The crawler's own 0–100 score for the page's technical health.

What it is not The crawler's number, not ours — we surface it rather than inventing a competing score.

Source Page crawl · on-page score

Reading ease

Flesch–Kincaid reading ease, 0–100. Higher is easier to read; 60+ is plain English.

What it is not Measures sentence and word length only. It cannot tell whether the writing is any good.

Source Page crawl · Flesch–Kincaid reading ease

Word count

Words in the page's main content, excluding navigation and boilerplate.

Source Page crawl · main-content word count

Accessibility

Lighthouse's 0–100 score for automated accessibility checks.

What it is not Automated checks only. A perfect score does not mean the site is usable with a screen reader.

Source Lighthouse · categories.accessibility.score

Best practices

Lighthouse's 0–100 score for security, browser and console-error checks.

Source Lighthouse · categories.best-practices.score

CLS

Cumulative Layout Shift — how much the page jumps around while loading. Good is under 0.1.

What it is not Unitless. It is not a time.

Source Lighthouse · audits.cumulative-layout-shift

FCP

First Contentful Paint — when the first text or image appeared. Good is under 1.8s.

What it is not The first thing on screen, not the main content — that is LCP.

Source Lighthouse · audits.first-contentful-paint

INP

Interaction to Next Paint — how long the page took to respond visually to an input. Good is under 200ms.

What it is not Lighthouse only reports it when its run included an interaction; a dash means the run had none, not that the page is instant.

Source Lighthouse · audits.interaction-to-next-paint

LCP

Largest Contentful Paint — when the biggest thing above the fold finished rendering. Good is under 2.5s.

What it is not A lab measurement, not field data from real users.

Source Lighthouse · audits.largest-contentful-paint

Lighthouse score

Google Lighthouse's own 0–100 score from a single simulated page load.

What it is not Lab data from one run on emulated hardware — not what your real visitors experienced, and it varies between runs.

Source Lighthouse · categories[].score

Performance

Lighthouse's 0–100 performance score, weighted mostly by loading speed metrics.

What it is not One lab run on emulated hardware — not what real visitors experienced.

Source Lighthouse · categories.performance.score

SEO (Lighthouse)

Lighthouse's 0–100 score for basic crawlability and markup checks.

What it is not A dozen mechanical checks — it says nothing about rankings, content or competition. Not our SEO analysis.

Source Lighthouse · categories.seo.score

TTFB

Time to First Byte — how long the server took to start responding.

What it is not Server and network only. It says nothing about how fast the page then renders.

Source Lighthouse · audits.server-response-time

Engines citing you

How many of the AI engines that returned an answer cited your domain in it.

What it is not Out of engines that ANSWERED, not out of every engine we asked. An engine that answered from memory with no sources cites nobody, so it is not counted against you.

Source Measured live · AI answers, per engine

Gaps found

Fixable problems this audit measured and could name evidence for.

What it is not What we test for, not a full site audit. Zero means nothing we check was wrong — not that nothing is wrong.

Source Our thresholds · applied to measured crawlability and citation signals

Indexed citations

Citations of this domain in AI answers already in an index, with the date each was last served.

What it is not A count of citations, not of distinct answers or questions: one answer can cite a domain more than once. History from an index, not today's live scan, and not exhaustive — none found means none in this index, not never cited.

Source AI answer index · mentions

Search demand, last 12 months

The change in monthly search demand across the questions this domain is indexed for, where an engine answers with AI.

What it is not Demand for the questions, not your visibility in them — this can rise while your citations fall. And not a count of prompts typed into assistants: it is derived from Google search data.

Source Search demand index · derived from Google search data