How Good Is Your Page at SEO and GEO? Run the Check

SEO and GEO are two disciplines with one goal: Search engine optimization (SEO) makes your page visible in Google rankings, while generative engine optimization (GEO) makes it citable in AI answers from ChatGPT, Gemini, and Perplexity. Both build on the same foundation, a cleanly structured, fact-rich page; they differ in the outcome: A position in a results list versus a citation inside a generated answer. Our free check tests both in one pass: 29 criteria across six areas, and every rating is a verifiable finding rather than an estimate. This article uses exactly those criteria to explain how SEO and GEO work together and where your page can improve.

👉 Want the full analysis with prioritized actions instead of a self-test? Get the SEO audit

Run the check: test your page right here

Enter one page of your website, optionally with a focus keyword. The check loads the page the way crawlers see it and rates it across six areas with scores from 0 to 10. Every finding quotes the exact spot on your page, and every recommendation names the next step. The check is fully deterministic: Same page, same result. No login, no email gate. One note for international readers: The check interface is currently in German; the scores and measured values are self-explanatory, and an English interface is on our roadmap.

The following sections explain what each area measures and why these criteria decide your visibility. The criteria are also the clearest way to see where SEO ends and GEO begins.

SEO and GEO explained

SEO optimizes a website for the results lists of classic search engines. The outcome is measurable as a position: Spots 1 through 10 on the first page. GEO, short for generative engine optimization, optimizes the same website so that AI systems pick it as a source. The outcome is a citation: Your brand, your number, or your sentence appears inside the generated answer, ideally with a link.

The key point: GEO does not replace SEO, GEO builds on it. Google confirms this itself. According to Google’s own documentation, AI Overviews and AI Mode use the same core systems as classic search, meaning the same index and the same ranking signals. And in its official Search Central blog, Google wrote in May 2025 that its AI experiences come with no additional requirements; the foundation remains proven, user-focused search engine optimization. A page that is invisible to Google has poor odds in AI answers too. That is why we treat SEO and GEO as one package, not two separate projects.

The difference lies in the selection logic. Google sorts ten blue links by relevance; even position 7 can earn clicks. An AI system generates a single answer and selects only a few passages for it. It favors text that is short, fact-rich, and understandable without context. Exactly these properties are what areas C and D of the check measure. For the full comparison of the two disciplines, read our article GEO vs. SEO; this article focuses on how they work together and how you measure both on your own page.

Search demand shows how quickly the question is gaining traction: The query “seo and geo” is searched around 590 times per month in the US and has grown from 70 to 590 monthly searches within twelve months (SE Ranking, August 2026). The related query “geo audit” has climbed from 10 to 170 monthly searches over the same period.

The six score areas: how the check measures SEO and GEO

The check evaluates 29 individual criteria and condenses them into six scores. Two areas measure SEO, two measure GEO, plus the two overall values. The split is deliberate: It makes the difference between the disciplines visible. A page can be strong at SEO and weak at GEO, and the check shows you why.

This is how the 29 criteria are distributed across the four test areas:

A simple scale puts the values into context: From 8.5 upward, a page is in strong shape and only needs polish. Between 6.5 and 8.4, the foundation is there but gaps are clear. Below 6.5, the biggest levers are still unused. Criteria that cannot be tested never enter the average as zero; they are excluded from the calculation.

Area A: Content SEO

Area A tests whether your page’s content meets the fundamentals Google has rewarded for years. Six criteria contribute, three of them weighted at 20 percent each:

• Title and meta description: The title has 15 to 60 characters, contains the focus term, and does not string keywords together. The description has 50 to 160 characters.
• Exactly one H1 with a clean hierarchy: Two H1s or none costs at least 5 out of 10 points right away. An H2 is never followed directly by an H4.
• Answer above the fold: The core message sits within the first 100 words. An opening like “Welcome to our company” earns 3 points at most, because it answers nothing.
• Substance: Below 150 words of main content, the maximum is 2 points; from roughly 600 words, service and guide pages earn full points.
• Internal linking: More than 5 internal links in the main content earn the full score. Generic anchors like “click here” cost points.
• Alt texts: The share of images with a descriptive alt attribute is rated linearly.

None of this is secret knowledge. The value lies in the measurement: In our audits we regularly see pages whose owners know all these rules and still fail 2 of the 6 criteria, most often the answer above the fold and internal linking. Perceived implementation and measured implementation are two different things.

Area B: Technical and semantic SEO

Area B measures whether search engines can reach, understand, and index your page at all. The heaviest single criterion in the area is indexability at 25 percent: A noindex sets the score to 0, a missing canonical caps it at 6 points, a canonical pointing to a different URL caps it at 3. The reason for the strict rating: This one line decides whether the other 28 criteria can have any effect at all.

Measurable basics follow: A final status 200 over HTTPS without a redirect chain (every hop costs 2 points), a readable URL, a time to first byte under 400 milliseconds for the full score, plus the language attribute, viewport, and Open Graph tags in the head.

One criterion in area B is also the most important GEO finding of the entire check: Content in the initial HTML. The check counts the words that sit in the delivered source code before any JavaScript runs. A measurement by Vercel across its own network logs shows why this matters: AI crawlers such as GPTBot and ClaudeBot fetch HTML but do not execute JavaScript. A page that renders its content in the browser is close to empty for these crawlers, no matter how good the visible text is. In that case the check shows a prominent warning, because rating the initial HTML is exactly the crawler’s view.

Area C: GEO content: does your page write in a citable way?

This is where GEO proper begins. Area C uses eight criteria to rate whether your text is phrased so that an AI system can quote it verbatim. The most important criterion carries 25 percent: The direct answer in the first paragraph. The check tests whether the first sentence contains a definition pattern, with the focus term as the subject at the start of the sentence, and whether the statement begins by word 8 at the latest. “A heat exchanger transfers heat between two media without mixing them” passes. “Welcome to our page all about heating technology” fails.

The word-8 rule sounds pedantic but has a measurable background: Insertions before the core statement (“A keyword, also known as a search term or query, is …”) dilute exactly the passage AI systems extract as the definition. We first counted this rule across our own SEO wiki with more than 90 glossary articles and then built it into the check.

The remaining criteria of the area at a glance:

• Questions as headings: From roughly 30 percent question share in H2 and H3, the score is full. People ask AI systems questions; pages that answer questions verbatim get cited more.
• A visible FAQ block: At least 3 questions with short answers. One to two sentences per answer is ideal; answers beyond 100 words cost points.
• Citable paragraphs: Over 70 percent of paragraphs stay under 65 words, average sentence length under 20 words. Vague openings like “This process takes two weeks” are downgraded because the paragraph is unreadable without context.
• Fact density: Concrete numbers, years, and units per 100 words, plus external source links. Vague plurals like “numerous benefits” without a number in the same sentence cost one point each.
• Citation patterns: The check counts five sentence patterns AI systems prefer to quote: Definition, comparison, numbered list, steps, and quantified statements. Three hits per 500 words earn the full score.
• Absence of fluff: Unproven superlatives like “leading”, “unique”, or “100% success” cost 2 points each. AI systems extract facts; claims without evidence get ignored.
• Freshness and author: A visible date and a recognizable author together earn the full score.

How to build texts around these patterns, with examples for all five citation patterns, is covered in our article on getting cited by AI. The check is the measuring instrument that goes with it.

Area D: GEO entity structure: does the AI understand who you are?

Area C rates what your page says. Area D rates what machines know about you. Nine criteria examine the structured data of your page, the JSON-LD that tells search engines and AI systems in machine-readable form who is behind the page, what it is about, and how everything connects.

The basic tests are quick to explain: Is JSON-LD present at all, does it parse without errors, and are all types real schema.org types? The classic among invented types is MarketingAgency; that type does not exist in schema.org, and the check names such findings concretely. After that it gets more demanding, and this is where standard separates from substance:

1. A fitting page type: A blog article needs a BlogPosting, a service page a Service, a glossary entry a DefinedTerm. The generic WebPage markup that SEO plugins generate automatically earns only 4 of 10 points, because it makes no statement about the page.
2. An organization entity with sameAs: AI systems must be able to recognize your brand as an unambiguous entity, verifiable through profiles. An organization with name, logo, description, and at least 3 sameAs references earns full points.
3. @id linking: The schema nodes should form a connected graph: The article references the author, the author references the organization, all via @id. Defining the same organization twice with different details is a contradiction risk and gets downgraded sharply.
4. FAQ markup matching the page: Every question in the markup must be visible on the page. Markup without a visible counterpart violates Google’s guidelines and earns 2 points.
5. speakable: The property marks the quotable core passages of a page via CSS selectors, and the check verifies that the selectors really exist in the DOM. Almost no website implements this; it is a genuine opportunity.
6. llms.txt: A curated briefing document for AI crawlers at /llms.txt that prioritizes and explains the most important pages. The check tests existence and format; full points require the tested URL to be linked in it. Our glossary entry on llms.txt explains the format.
7. Date and language hygiene: Dates in ISO 8601 format with a timezone offset, and the schema language matches the page’s language attribute.

Why SEO and GEO belong together

The four areas reveal the pattern: Not a single GEO criterion contradicts an SEO criterion. A direct answer in the first paragraph improves the snippet in Google and the citability in ChatGPT. Short, fact-rich paragraphs help readers, rankings, and AI answers alike. Clean structured data feeds Google’s rich results and the entity recognition of AI systems. If you do SEO seriously today, 70 percent of GEO is already done; the remaining 30 percent is deliberate work on citability and entity structure.

The reverse also holds: GEO without an SEO foundation does not work. A page with noindex, without indexable content in the initial HTML, or with a 2-second load time can be phrased as citably as you like; it enters no index and therefore no answer. That is why we treat GEO as a fixed building block of ongoing SEO services, next to strategy, content, internal structure, technology, and structured data, not as a separate product with its own contract.

For prioritization this means: First bring the SEO foundation to a solid 6.5 or better, then build out the GEO areas systematically. The check hands you the order, because it names the three criteria with the largest weighted point gap as your top levers.

From score to action: the most common findings and their fixes

A score is only worth as much as the action that follows. These five findings appear most often in client projects, each with the fitting first step:

1. Step 1: The opening answers nothing. The first paragraph greets instead of answering. Fix: Phrase the page’s core message as the first sentence, with the topic as the subject at the start.
2. Step 2: Plugin schema only. The page carries the generic WebPage markup of the SEO plugin but no fitting page type. Fix: Add BlogPosting, Service, or FAQPage and connect it to the organization via @id.
3. Step 3: Organization without sameAs. The brand exists in the schema but is linked to no external profile. Fix: Add at least 3 verifiable profile URLs as sameAs.
4. Step 4: Vague instead of concrete. “Numerous happy clients” instead of a number, “years of experience” instead of a year. Fix: Replace every vague phrase with the concrete value that almost always exists behind it.
5. Step 5: Content only after JavaScript. The initial HTML contains almost no text. Fix: Server-side rendering or pre-rendering, so crawlers see the content without executing scripts.

The check tests a single page; that is its strength and its limit. Findings such as redirect chains across the whole domain, orphaned pages, duplicate content between language versions, or a contradictory robots setup only show up in a full crawl. That is what the SEO audit is for: Every URL of the website, prioritized by impact, with an action plan.

FAQ: SEO and GEO

What is the difference between SEO and GEO?

SEO optimizes a website for positions in search engine results lists, while GEO optimizes it for citations in generated answers from AI systems such as ChatGPT, Gemini, and Perplexity. Both work on the same content and technical foundation but differ in the outcome: Ranking position versus answer citation.

Does GEO replace classic SEO?

No. GEO builds on SEO: According to Google, its AI features use the same index and the same core systems as classic search. A page that is not indexable or has weak content appears neither in rankings nor in AI answers.

How can I measure the AI visibility of my website?

Two levels are measurable: The prerequisites (citable content, structured data, crawler accessibility), which can be tested deterministically on any page, and the actual mentions in AI answers, which specialized monitoring tools capture through repeated prompt queries. A page check like the one in this article measures the first level.

What is a good score in an SEO and GEO check?

From 8.5 out of 10, a page is in strong shape; between 6.5 and 8.4 it is solid with clear gaps; below 6.5 the biggest levers are unused. The individual findings matter more than the total: A single noindex can make an otherwise perfect page completely invisible.

How often should I check my pages?

Check a page after every content or technical change, and your most important pages at least once per quarter. Relaunches, theme updates, and plugin changes are the most common triggers for unnoticed regressions in indexability and structured data.

Sources

• Google Search Central: “AI features and your website“, developers.google.com, 2025.
• Google Search Central Blog: “Top ways to ensure your content performs well in Google’s AI experiences on Search“, developers.google.com, May 2025.
• Vercel: “The rise of the AI crawler“, vercel.com, December 2024.