What does each check measure?

By Matt Kerr·Published 11 November 2026

Find what each Check Engine test reads, what it reports and what to review before changing your site.

Choose an area below, then open a check. Each reference explains the test rules, sample limits and severity. When an existing guide already owns the same explanation, the link leads to that guide.

Findings describe what the audit could check. A missing measurement is different from a measured problem. Read the scope and the saved example before deciding which fix applies.

AI search visibility

• Citable, specific claims: It looks for numbers backed by words that explain where the claim comes from.
• AI crawler access: It separates crawler restrictions you chose from restrictions inherited from a general robots rule.
• Dead AI crawler directives: It finds retired or unmatchable crawler names in the robots file.
• Server access for AI crawlers: It compares saved server responses with the crawler access your robots file allows.
• Entity definition and sameAs: It looks for business identity markup and recognised profile links in sameAs.
• Answer extractability: It counts substantial pages with question headings that are specific to the page.
• Dates AI engines can read: It checks whether substantial pages carry a date in machine-readable markup.
• llms.txt: It reports whether the saved site probe found an llms.txt file it could read.
• Content reachable without JavaScript: It finds indexable pages marked as dependent on JavaScript for their content.
• The AI control nothing can detect: It explains an AI visibility control that cannot be checked from outside the site.

Brand consistency

• Logo presence and markup: It checks how often a readable business identity appears and whether logo text names it.
• Brand name consistency: It compares declared business names while allowing common forms of the same name.
• Consistent positioning statement: It checks for shared subject words across page descriptions.
• Social profile consistency: It compares profile platforms linked on the site with those declared in sameAs.
• Typography and colour consistency: It counts typefaces and groups nearly identical colours from readable styles.
• Writing voice consistency: It compares average sentence lengths across substantial pages.

Content quality

• Generic AI-sounding copy: It flags pages containing multiple phrases from the engine's stock-copy list.
• Pages competing with each other: It finds groups of distinct pages whose title subjects overlap after several exclusions.
• Freshness signals: It looks for article pages missing a published or modified date in markup.
• Trust and authorship signals: It checks for trust pages, a route to contact you and article authorship signals.
• JavaScript-dependent content: It reports indexable pages whose initial HTML lacks their main content.
• Declared page language: It compares a page's declared language with conservative language clues in the text.
• Language coverage: It states when language-dependent copy checks were skipped.
• Placeholder content: It searches for known filler text while excluding many cases where a page discusses it.
• Readability: It finds substantial pages whose usable sentences are long on average.
• Thin content: It separates short content, sparse archives and intentionally indexable utility pages.
• Topics worth writing about: It compares saved demand queries with words already used in page titles and headings.

Conversion & landing pages

• Published email addresses that cannot receive mail: It reports published email domains that the saved mail test says cannot accept mail.
• Form friction: It finds long forms and checks whether key pages offer a contact route.
• Funnel coverage: It looks for page paths that suggest educational and comparison content.
• Title / headline consistency: It compares meaningful words in the title with a page-specific main heading.
• Calls to action: It checks for an action or useful enquiry form on the pages most likely to sell.
• Pricing transparency: It looks for a pricing route or recognisable cost language.
• Social proof: It looks for recognised proof language on key conversion pages.
• Phone links that will not dial: It checks phone-link destinations for forms that are placeholders or not defined by the link standard.
• Above-the-fold value proposition: It reviews the saved opening text for an absent or inward-looking statement.

Crawlability & indexing

• Broken pages (4xx/5xx): It separates server errors, missing pages, refusals and requests that could not finish.
• Googlebot's 2MB fetch limit: It compares decompressed HTML size and important markup offsets with a fixed byte budget.
• Canonical tags: It checks missing, conflicting, off-site and unusable canonical targets.
• Conflicting canonical tags: It compares a canonical in the response header with the canonical in the HTML.
• Click depth and orphans: It reports deep pages and sitemap pages with no observed internal links.
• Duplicate pages: It groups identical page bodies, titles and descriptions after checking canonical agreement.
• Client-side redirects: It separates JavaScript redirects from HTML meta refresh redirects.
• Pages blocked from indexing: It checks noindex directives while excluding several intentional and sitemap-only cases.
• Redirect problems: It finds redirect loops, longer chains and internal links pointing at a redirect.
• robots.txt blocks the whole site: It checks whole-site search restrictions and problems reading the robots policy.
• XML sitemap: It checks sitemap discovery, responses, URL quality and coverage against the crawl.
• Soft 404 handling: It reports the saved site probe for a missing page that answered as successful content.
• Soft 404 pages in the crawl: It finds short successful pages containing one of the stored missing-page messages.
• URL structure: It checks address spelling, length, query parameters and session identifiers.
• www / apex canonicalisation: It reports the saved result for host canonicalisation.
• hreflang language codes: It separates malformed, superseded, unsupported and unrecognised language-region codes.
• hreflang return tags: It checks reciprocal hreflang declarations between pages the crawl successfully read.
• hreflang self-reference: It looks for a self-reference in pages that already declare hreflang alternatives.
• hreflang target health: It checks measured hreflang targets for errors, redirects and noindex directives.

Links

• Broken outbound links: It separates broken external destinations from refusals, sharing links and malformed addresses.
• Insecure internal links: It finds internal links written with an insecure scheme on a site served over HTTPS.
• Internal links into and out of each page: It checks sparse outgoing links and widespread pages with a single observed incoming link.
• Internal links to the pages that sell: It finds non-homepage selling pages with exactly a single observed internal inlink.
• Navigation reachable without JavaScript: It checks the saved navigation-in-HTML result across successful HTML pages.
• Outbound link handling: It checks whether external links are absent, dominated by a host or almost all nofollow.
• Pagination handling: It looks for numbered listing pages whose canonical points to a non-paginated address.

Local business SEO

• Local business detection: It checks whether your public business details call for local business tests.
• Contact and conversion signals: It checks tappable contact routes, maps and enquiry forms where local checks apply.
• Location targeting in content: It checks place names in titles and postal codes for a local market.
• NAP consistency: It reviews phone numbers and addresses while allowing genuine multiple locations.
• Review signals: It looks for customer reviews and for item reviews that could need markup.
• LocalBusiness schema: It checks local identity markup and important fields while respecting business type and ownership.

Tracking & campaign readiness

• Advertising pixels: It distinguishes detected pixels, staged pixels and incomplete page coverage.
• Analytics installed: It separates missing analytics, consent-gated tags, opaque containers and measured coverage gaps.
• Consent and privacy compliance: It looks for a consent banner and privacy policy alongside detected tracking.
• Email list building: It looks for recognised email tools, signup forms or invitation wording.
• Marketing stack inventory: It inventories detected tools and raises severity when a page carries many of them.
• UTM tagging hygiene: It reports indexable pages linking within the site with UTM parameters.

On-page SEO

• Internal link anchor text: It finds vague link text and a larger set of links with no text.
• Meta descriptions: It checks whether descriptions exist and fit the engine's display-width range.
• Heading structure: It checks missing main headings, distinct repeated main headings and gaps in the outline.
• Image optimisation: It checks missing alternative text, dimensions and lazy loading on image-heavy pages.
• Language and viewport declarations: It checks for a language value and a mobile viewport declaration.
• Title tags: It checks missing titles, display width and excessive repeated keywords.

Speed & Core Web Vitals

• Layout shift risk: It identifies image patterns associated with shifting layout, rather than measuring a browser score.
• Compression and caching headers: It checks response headers for missing compression and caching settings.
• Image formats: It counts pages with several JPEG or PNG images lacking a modern source alternative.
• Page weight: It compares decompressed HTML bytes with the engine's document-size cutoffs.
• Redirect latency: It checks how much of the crawled URL set travelled through a redirect.
• Render-blocking resources: It counts resources marked as blocking the initial render.
• Server response time: It compares first-response samples, repeat requests and slower-page outliers.

Structured data

• Incomplete structured data: It checks required properties by type, including empty fields and unresolved references.
• Invalid JSON-LD: It reports parsing errors, CDATA wrappers and unfilled template values.
• Structured data vs visible content: It checks whether declared ratings have visible reviews or rating text.
• Missing structured data: It checks for schema coverage, business identity and breadcrumb markup.
• Rich result opportunities: It finds pages with several qualifying question headings and no FAQPage declaration.
• Schema price vs visible price: It compares nonzero declared price amounts with visible and saved rendered text.

Security & trust

• Security headers: It checks selected response headers and recognises some equivalent protections.
• HTTPS: It checks HTTPS, the plain-HTTP route and certificate validity.
• Mixed content: It separates insecure active resources from pages with insecure images only.

Social & sharing

• Favicon: It reports a confirmed missing favicon from the saved site probe.
• Open Graph tags: It checks Open Graph identity, images, descriptions and measured image failures.
• Twitter/X card: It checks widespread missing card-type declarations on sites with Open Graph titles.