Methodology: How Everyday AEO Scores a Page

Scoring version 2. Last updated 2026-11-11.

Everyday AEO reads one web page, runs 99 documented rules over it and turns the results into a 0 to 100 score. This page explains every step so you can check our work. Every rule below has a stable ID you can link to.

How scoring works

Rules earn points. Each scored rule has a maximum number of points. A rule earns all, part or none of them, and records the evidence it saw on the page and, when it lost points, a plain-English fix.
Categories. A category score is 100 × points earned ÷ points available, counting only the scored rules that were measured.
Overall score. The overall score is the weighted average of the category scores, using the weights below. Only the final number is rounded.
Not measured is excluded, never a free pass. If we could not get the data for a rule (for example Google PageSpeed was unavailable or robots.txt could not be read), the rule is marked “not measured” and leaves both the points earned and the points available. If a whole category cannot be measured, the other weights are re-normalised. If nothing can be measured, the page is “Not scored”.
Caps are applied last. A few dealbreakers limit the final score. The report always shows both the capped score and the score before caps.

The 7 scoring categories. Weights sum to 100.
Category	Weight	What it covers
Content Structure	22	How the page's headings, sections, lists and metadata are organised so people and AI can scan and extract it.
Authority & Trust (E-E-A-T)	18	Experience, expertise, authoritativeness and trust signals: authorship, citations, sourcing and article markup.
AI Access & Readiness	15	Whether AI crawlers and answer engines are allowed to fetch, index and quote the page.
Technical Optimization	15	Behind-the-scenes markup and delivery: HTTPS, structured data, canonical URL, social tags and speed.
Content Clarity & Relevance	12	Writing that is clear, direct and easy for readers and AI systems to extract answers from.
Trust & Compliance	10	Disclosures, freshness dates and contact information that show the page is transparent and maintained.
User Experience	8	Mobile readiness, navigation and accessibility basics that make the page usable.
Total	100

There is no base score, no default value and no bonus for passing several rules at once. 70 rules are scored and 29 are information only.

Score caps

A cap is the highest score a page can receive when a dealbreaker is present. If several caps apply, the lowest one wins.

Dealbreaker caps
Cap	Maximum score	Applies when
`cap-noindex`	20	meta robots or X-Robots-Tag header says noindex (or none)
`cap-robots-blocked`	20	robots.txt (fetched OK) disallows the start page for EverydayAEO-Analyzer or *
`cap-empty-page`	30	under 10 words of main text and no other body content
`cap-thin-content`	49	under 100 words of main-content text
`cap-no-title`	69	no non-empty <title> element

A robots.txt file that cannot be reached, or that returns a server error, never triggers the robots cap.

Grades

Grade for the final (capped) score
Grade	Label	Score
A	Excellent	90 to 100
B	Good	80 and above
C	Fair	70 and above
D	Needs work	50 and above
F	Poor	0 and above

What we fetch

When you submit a URL our servers fetch it, not your browser. We identify ourselves with this user agent:

EverydayAEO-Analyzer/2.0 (+https://everydayaeo.com/methodology)

The page you submitted (HTTP or HTTPS, following up to 5 redirects).
/robots.txt and /sitemap.xml on the same site.
A handful of optional files used only for information: /llms.txt, /AGENTS.md and agent discovery files under /.well-known/.
Google PageSpeed Insights data for the URL, when that service is available.

We audit one page per request. We do not crawl your site, log in or submit forms. Requests to private or internal network addresses are refused. You can block our crawler in robots.txt with the token EverydayAEO-Analyzer.

AI crawler policy

Different AI crawlers do different jobs, so we treat them differently.

Citation and search crawlers are scored. These fetch pages so an AI product or search engine can find and cite them (for example OAI-SearchBot, Claude-SearchBot, PerplexityBot, Googlebot, Bingbot). Blocking one costs points.
Training crawlers are reported, not penalised. Crawlers that collect content to train models (for example GPTBot, ClaudeBot, CCBot, Google-Extended) are shown with their status. Opting out of model training is a legitimate choice and never lowers your score.
Speculative standards are informational. llms.txt, AGENTS.md, agent files under /.well-known/, Content-Signal directives and markdown responses are reported, but they do not move the score.

Crawlers checked against robots.txt (list last reviewed 2026-11-10)
Token	Vendor	Purpose	Treatment
`OAI-SearchBot`	OpenAI	search	Scored
`ChatGPT-User`	OpenAI	user-action	Scored
`Claude-SearchBot`	Anthropic	search	Scored
`Claude-User`	Anthropic	user-action	Scored
`PerplexityBot`	Perplexity	search	Scored
`Perplexity-User`	Perplexity	user-action	Scored
`Googlebot`	Google	search	Scored
`Bingbot`	Microsoft	search	Scored
`Applebot`	Apple	search	Scored
`GPTBot`	OpenAI	training	Reported, never penalised
`ClaudeBot`	Anthropic	training	Reported, never penalised
`anthropic-ai`	Anthropic	training	Reported, never penalised
`Google-Extended`	Google	training	Reported, never penalised
`Applebot-Extended`	Apple	training	Reported, never penalised
`CCBot`	Common Crawl	training	Reported, never penalised
`Bytespider`	ByteDance	training	Reported, never penalised
`Meta-ExternalAgent`	Meta	training	Reported, never penalised
`cohere-ai`	Cohere	training	Reported, never penalised
`Amazonbot`	Amazon	training	Reported, never penalised

Healthcare add-on

If a page looks like health or life-sciences content (for example a drug product page, clinician content, patient education or a health organisation), we also run a labelled healthcare add-on. It checks for items such as important safety information, fair-balance language, a prescribing information link, an adverse-event reporting contact, a named medical reviewer, a recent “last reviewed” date, medical schema markup and links to primary sources. These rules are tagged “healthcare add-on” in the list below and are scored under Trust & Compliance and Authority & Trust.

On pages that are not health content, every add-on rule is marked not applicable, so non-health sites are never penalised.

Not an MLR review. The healthcare add-on is an automated check for the presence of common signals. It is not a medical, legal or regulatory (MLR) review, does not judge whether any claim is accurate or compliant, and does not replace your review process. Passing it does not mean content is approved for use.

Full rule list

99 rules, grouped by category. Each rule's anchor is its ID, for example #ai-agents-md. Rule IDs are stable: once published they are not renamed.

Content Structure weight 22, 10 rules, 47 scored points

How the page's headings, sections, lists and metadata are organised so people and AI can scan and extract it.

core-faq-section FAQ section4 pts#
If it fails: Add a short FAQ section answering the questions customers most often ask.
core-h1-present Page has a single H16 pts#
If it fails: Add one clear H1 that states the page topic. It anchors the content hierarchy for search engines and AI.
core-h1-title H1 and title describe the same topic3 pts#
If it fails: Make the H1 and the title describe the same topic (they may differ in wording, not in subject).
core-heading-count Enough headings to structure the page5 pts#
If it fails: Break the content into more sections with descriptive H2/H3 headings (aim for 6 or more).
core-heading-hierarchy Heading levels are not skipped4 pts#
If it fails: Fix the heading order so levels never skip (H1, then H2, then H3).
core-lists-tables Lists or tables for scannability4 pts#
If it fails: Add bulleted or numbered lists (or a table) for key points; AI often extracts them directly.
core-meta-description Meta description 70-160 characters5 pts#
If it fails: Add a meta description of 70-160 characters summarising the page.
core-question-headings Question-format headings6 pts#
If it fails: Rewrite key headings as questions people ask ("What is...?", "How does...work?"). AI assistants match them to queries.
core-title-length Title length 30-60 characters4 pts#
If it fails: Lengthen the title to 30-60 characters with the page topic and brand.
core-word-count Content depth (word count)6 pts#
If it fails: The page is thin. Expand it with substantive, useful detail (aim for 800+ words where the topic warrants it).

Authority & Trust (E-E-A-T) weight 18, 9 rules, 46 scored points

Experience, expertise, authoritativeness and trust signals: authorship, citations, sourcing and article markup.

auth-article-schema Article-type schema8 pts#
If it fails: If this is an article or post, add Article/BlogPosting JSON-LD with headline, author and dates.
auth-author-credentials Author byline or credentials12 pts#
If it fails: Add a visible author byline with role and credentials, linked to a short bio. AI systems weight authored content higher.
auth-citations Citations or references10 pts#
If it fails: Cite authoritative sources (research, official guidance, standards) with links.
auth-domain-signal Domain typeinfo only#
Reported for information. Never changes the score.
auth-external-links External sources linked10 pts#
If it fails: Link to more reputable external sources that support your claims; well-sourced pages are cited more.
pharma-last-reviewed Healthcare: Last reviewed date is recent1 ptshealthcare add-on#
If it fails: Show a visible "Last reviewed: date " line near the top of the content and refresh it every time a clinician re-checks the page.
pharma-medical-reviewer Healthcare: Medical or scientific reviewer named2 ptshealthcare add-on#
If it fails: Add a line such as "Medically reviewed by Jane Smith, MD" near the top, and add reviewedBy to your MedicalWebPage schema.
pharma-medical-schema Healthcare: Medical structured data1 ptshealthcare add-on#
If it fails: Add JSON-LD: MedicalWebPage (with lastReviewed and reviewedBy) for content pages, Drug for product pages, MedicalCondition for condition pages, MedicalOrganization for the organisation.
pharma-primary-sources Healthcare: Links to primary medical sources2 ptshealthcare add-on#
If it fails: Cite at least two primary sources (PubMed, FDA, CDC, NIH, WHO, ClinicalTrials.gov, DailyMed or a DOI link) for medical claims.

AI Access & Readiness weight 15, 34 rules, 100 scored points

Whether AI crawlers and answer engines are allowed to fetch, index and quote the page.

ai-agents-md AGENTS.mdinfo only#
Reported for information. Never changes the score.
ai-bot-amazonbot Amazonbot (Amazon) accessinfo only#
Reported for information. Never changes the score.
ai-bot-anthropic-ai anthropic-ai (Anthropic) accessinfo only#
Reported for information. Never changes the score.
ai-bot-applebot Applebot (Apple) access5 pts#
ai-bot-applebot-extended Applebot-Extended (Apple) accessinfo only#
Reported for information. Never changes the score.
ai-bot-bingbot Bingbot (Microsoft) access5 pts#
ai-bot-bytespider Bytespider (ByteDance) accessinfo only#
Reported for information. Never changes the score.
ai-bot-ccbot CCBot (Common Crawl) accessinfo only#
Reported for information. Never changes the score.
ai-bot-chatgpt-user ChatGPT-User (OpenAI) access5 pts#
ai-bot-claude-searchbot Claude-SearchBot (Anthropic) access5 pts#
ai-bot-claude-user Claude-User (Anthropic) access5 pts#
ai-bot-claudebot ClaudeBot (Anthropic) accessinfo only#
Reported for information. Never changes the score.
ai-bot-cohere-ai cohere-ai (Cohere) accessinfo only#
Reported for information. Never changes the score.
ai-bot-google-extended Google-Extended (Google) accessinfo only#
Reported for information. Never changes the score.
ai-bot-googlebot Googlebot (Google) access5 pts#
ai-bot-gptbot GPTBot (OpenAI) accessinfo only#
Reported for information. Never changes the score.
ai-bot-meta-externalagent Meta-ExternalAgent (Meta) accessinfo only#
Reported for information. Never changes the score.
ai-bot-oai-searchbot OAI-SearchBot (OpenAI) access5 pts#
ai-bot-perplexity-user Perplexity-User (Perplexity) access5 pts#
ai-bot-perplexitybot PerplexityBot (Perplexity) access5 pts#
ai-content-signal Content-Signal in robots.txtinfo only#
Reported for information. Never changes the score.
ai-llms-txt llms.txtinfo only#
Reported for information. Never changes the score.
ai-markdown-response Markdown response via content negotiationinfo only#
Reported for information. Never changes the score.
ai-meta-max-snippet max-snippet allows quoting3 pts#
If it fails: Raise max-snippet to at least 160 or use -1 (unlimited) so AI can quote meaningful excerpts.
ai-meta-noai noai / noimageai directivesinfo only#
Reported for information. Never changes the score.
ai-meta-noindex Page is indexable (meta robots / X-Robots-Tag)3 pts#
If it fails: Remove the noindex directive unless the page is intentionally private. It stops search engines and AI from indexing it.
ai-meta-nosnippet No nosnippet directive4 pts#
If it fails: Remove nosnippet, or use a generous max-snippet instead, so AI can show excerpts.
ai-robots-fetched robots.txt is retrievable and valid5 pts#
ai-schema-drift Structured data matches visible content15 pts#
If it fails: Make the JSON-LD describe what a visitor can actually read: headline, author name, FAQ questions and dates must appear on the page.
ai-semantic-html Semantic HTML makes content extractable15 pts#
If it fails: Wrap the primary content in <main>, use one <h1> with ordered <h2>/<h3> sections, and present key facts as lists, tables or definition lists so AI can extract them.
ai-sitemap XML sitemap is declared and reachable10 pts#
If it fails: Publish an XML sitemap and add a "Sitemap: https://your-site/sitemap.xml" line to robots.txt so crawlers can discover every page.
ai-well-known Agent discovery files (.well-known)info only#
Reported for information. Never changes the score.
render-content-gap Content visible without JavaScript vs after renderinginfo only#
Reported for information. Never changes the score.
render-js-required Main content does not depend on JavaScriptinfo only#
Reported for information. Never changes the score.

Technical Optimization weight 15, 23 rules, 61 scored points

Behind-the-scenes markup and delivery: HTTPS, structured data, canonical URL, social tags and speed.

tech-canonical Canonical URL valid5 pts#
If it fails: Add a link rel="canonical" href="https://..." tag with the absolute preferred URL to consolidate duplicate URLs.
tech-cwv-cls Core Web Vitals: Cumulative Layout Shift (field)2 pts#
tech-cwv-inp Core Web Vitals: Interaction to Next Paint (field)3 pts#
tech-cwv-lcp Core Web Vitals: Largest Contentful Paint (field)3 pts#
tech-faq-schema FAQPage schema3 pts#
If it fails: Mark up your FAQ content with FAQPage JSON-LD so AI can extract question/answer pairs.
tech-howto-schema HowTo schemainfo only#
Reported for information. Never changes the score.
tech-hreflang hreflang annotations are saneinfo only#
Reported for information. Never changes the score.
tech-https Served over HTTPS6 pts#
tech-jsonld-valid JSON-LD blocks parse as valid JSON5 pts#
If it fails: Fix the JSON syntax of the invalid JSON-LD block(s) (trailing commas, missing brackets or quotes). Validate with the Schema Markup Validator.
tech-lab-cls Lab CLS (Lighthouse)info only#
Reported for information. Never changes the score.
tech-lab-lcp Lab LCP (Lighthouse)info only#
Reported for information. Never changes the score.
tech-lab-tbt Lab Total Blocking Time (Lighthouse)info only#
Reported for information. Never changes the score.
tech-lang HTML lang attribute3 pts#
If it fails: Add lang="en" (or the right language code) to the html element.
tech-mixed-content No mixed content (http:// subresources)3 pts#
If it fails: Load every image, script, stylesheet and frame over https://. Browsers block or warn on mixed content.
tech-open-graph Open Graph tags3 pts#
If it fails: Add og:title, og:description and og:image so shared links render well.
tech-page-complexity Page is simple enough to analyse safelyinfo only#
If it fails: Reduce the size and nesting of the HTML (unclosed or deeply nested elements, extremely large generated pages). Very large DOMs also slow browsers and AI crawlers.
tech-pagespeed-performance PageSpeed performance6 pts#
tech-robots-consistency Meta robots and X-Robots-Tag agreeinfo only#
Reported for information. Never changes the score.
tech-schema-present Structured data present5 pts#
If it fails: Add valid JSON-LD structured data (Article, FAQPage, HowTo, Organization or Product as relevant).
tech-schema-required-props Schema has required properties6 pts#
If it fails: Add the missing required properties to each schema item (for example Article needs headline, author and datePublished).
tech-schema-richness Structured data variety3 pts#
If it fails: Add valid JSON-LD with @type values describing the page.
tech-twitter-cards Twitter/X card tags2 pts#
If it fails: Add twitter:card (and optionally twitter:title/description/image) so links render well on X.
tech-x-robots-tag No noindex in X-Robots-Tag header3 pts#

Content Clarity & Relevance weight 12, 5 rules, 44 scored points

Writing that is clear, direct and easy for readers and AI systems to extract answers from.

content-direct-answers Direct answers after question headings12 pts#
If it fails: Put a 1-3 sentence direct answer (10-80 words) right after each question heading; this is what AI extracts.
content-lorem-ipsum No placeholder filler text6 pts#
If it fails: Replace all placeholder text with real content before publishing.
content-paragraph-length Paragraph length8 pts#
content-readability Readability10 pts#
If it fails: Simplify the language: shorter sentences and plainer words.
content-sentence-length Sentence length8 pts#
If it fails: Sentences are very long. Split them; aim for an average under 20 words.

Trust & Compliance weight 10, 13 rules, 53 scored points

Disclosures, freshness dates and contact information that show the page is transparent and maintained.

pharma-adverse-event-reporting Healthcare: Adverse event reporting contact2 ptshealthcare add-on#
If it fails: Add a line telling people how to report side effects: the company contact (phone or email) and FDA MedWatch at www.fda.gov/medwatch or 1-800-FDA-1088.
pharma-audience-gating-signal Healthcare: Audience pathways (patient vs healthcare professional)info onlyhealthcare add-on#
Reported for information. Never changes the score.
pharma-disclaimer Healthcare: Medical disclaimer1 ptshealthcare add-on#
If it fails: Add a plain disclaimer such as "This information is not a substitute for professional medical advice, diagnosis or treatment. Talk to your healthcare provider."
pharma-fair-balance Healthcare: Risk information near efficacy claims (fair balance)2 ptshealthcare add-on#
pharma-health-context Healthcare: Healthcare checksinfo onlyhealthcare add-on#
Reported for information. Never changes the score.
pharma-isi Healthcare: Important Safety Information (ISI)3 ptshealthcare add-on#
pharma-prescribing-info-link Healthcare: Prescribing Information link1 ptshealthcare add-on#
If it fails: Add a visible link to the full Prescribing Information (and Medication Guide where one exists) near the safety information and in the footer.
pharma-promotional-claims-info Healthcare: Promotional or absolute claims for compliance reviewinfo onlyhealthcare add-on#
Reported for information. Never changes the score.
trust-contact-info Contact information8 pts#
If it fails: Add visible contact details (email, phone or a contact page link).
trust-dates Publication / update dates10 pts#
If it fails: Show publication and last-updated dates (use time datetime="..." and schema datePublished/dateModified).
trust-disclaimers Disclaimers or disclosures10 pts#
If it fails: Add the disclaimers, editorial policy or disclosures appropriate to your industry (for example a medical or financial disclaimer).
trust-freshness Content is recently updated8 pts#
If it fails: Review and update the content, then show a new "Last updated" date. Stale pages are cited less.
trust-policy-links Privacy, terms or editorial policy linked8 pts#
If it fails: Link your privacy policy, terms and (for editorial sites) editorial policy from the footer.

User Experience weight 8, 5 rules, 36 scored points

Mobile readiness, navigation and accessibility basics that make the page usable.

ux-accessibility-score Lighthouse accessibility score4 pts#
ux-best-practices Lighthouse best-practices score2 pts#
ux-image-alt Images have alt text12 pts#
If it fails: Add descriptive alt text to your images (use alt="" only for purely decorative images).
ux-navigation Navigation element8 pts#
If it fails: Add a nav element with the main site links.
ux-viewport Mobile viewport meta tag10 pts#
If it fails: Add meta name="viewport" content="width=device-width, initial-scale=1" so the page works on mobile.

Limitations

One page at a time. A score describes the single URL you submitted, not your whole site.
Heuristics, not verdicts. Rules look for patterns in the HTML and related files. They can miss things or flag things incorrectly, especially on unusual layouts.
Static HTML. By default we read the HTML the server returns. Content added later by scripts may not be seen.
No guarantees. A higher score does not guarantee rankings, traffic or citations by any AI product. Vendors change how their crawlers and answer engines work, and we review the crawler list regularly.
Third-party data. Speed metrics come from Google PageSpeed Insights and are shown as not measured when that service is unavailable.
Not professional advice. Reports are not legal, medical or regulatory advice.

Changelog

Scoring version 2 (current): Server-side and rule-based. Every rule has a stable ID, points, evidence and a fix. Seven weighted categories, score caps for dealbreakers, “not measured” excluded from the score, per-crawler AI access checks and a healthcare add-on. Scores can differ from version 1 for the same page.
Scoring version 1: Browser-side. The score was calculated in your browser using hash-based and heuristic checks, with fixed starting values for some signals.

Last updated 2026-11-11.