AI Trust AuditBETA
    Retour à l'aperçu
    Savoir SEO/GEO

    Google AI Ranking Factors: 5 Insights from Major GEO Expert Survey

    GEO Expert28. septembre 20268 min de lecture
    Google AI Ranking Factors: 5 Insights from Major GEO Expert Survey

    What factors determine whether Google cites your website in AI Overviews and in AI Mode? Cyrus Shepard surveyed over 130 SEO professionals and had them evaluate 100 potential influencing factors. He published the results in the study "Google AI Ranking Factors" at Zyppy Signal. We summarize the five most important lessons and show what they specifically mean for online shops.

    The Study at a Glance

    FeatureValue
    PublisherCyrus Shepard, Zyppy Signal
    Respondents131 SEO and GEO experts
    Evaluated Factors100
    Scale−3 (strongly negative) to +3 (strongly positive)
    MethodExpert consensus, no correlation or causal analysis

    Important for context: The study measures what experienced practitioners believe to be effective. It is not mathematical proof, but it aligns well with the known functionality of generative search: retrieve content, weight by trust, formulate answer.

    1. Crawlability is the Admission Ticket

    The highest score (+2.20) was given to access for AI crawlers and suitability for snippets. What Google cannot retrieve, render, or use as a snippet cannot be cited.

    What this means for your shop:

    • Do not accidentally block AI crawlers in the robots.txt.
    • No general nosnippet or max-snippet:0 directives.
    • Deliver important content server-side in HTML, rather than reloading it via JavaScript.

    The AI Trust Audit checks exactly this foundation in the Discoverability stage – from crawler access to content that only becomes visible via JavaScript.

    2. The Answer Must Match the Question

    Following with +2.15 is the match between the search query and the answer. AI systems prefer passages that answer a question directly, completely, and without long-windedness.

    What this means for your shop:

    • Place the core statement at the beginning of a section.
    • Use real customer questions as subheadings.
    • Mark up FAQ content as visible HTML, rather than hiding it in purely client-side accordions.

    3. Brand and Entity Count More Than Any Trick

    In third place (+2.08) is the anchoring of the brand in the "memory" of the language model. The more clearly a company is recognizable as an independent entity on the web, the more likely it is to be mentioned.

    What this means for your shop:

    • A meaningful About Us page with team, location, history, and trust seals.
    • Organization schema with sameAs links to official profiles.
    • Consistent brand name, consistent address, and consistent contact details everywhere.

    4. Citable Facts and Original Data

    Experts rated concrete, verifiable facts and primary data almost as highly at +2.07. An AI prefers to cite "delivery in 24 hours, 4.8 out of 5 stars from 12,000 reviews" rather than "fast shipping and satisfied customers."

    What this means for your shop:

    • State figures, prices, delivery times, and warranties precisely.
    • Publish your own analyses or surveys.
    • Integrate customer reviews in a machine-readable way so that AI systems can capture them.

    5. Mass Content Hurts, Speculation Doesn't Help

    On the opposite side is scaled, AI-generated text: with −1.82, it is considered one of the most harmful factors. Much-touted individual measures like an isolated llms.txt file or artificial "chunking" of texts were also rated as having little effect by the experts.

    What this means for your shop:

    • Focus on a few well-founded pages rather than many interchangeable ones.
    • Do not invest energy in measures whose effect is unproven.
    • Quality and trust over quantity.

    What the Study Means for the AI Trust Audit

    The results confirm our three-stage model of discoverability, citability, and semantic fit. Checks such as extractable answers, structured content formats, entity linking via sameAs, and the quality of the About Us page directly cover the top factors. We deliberately do not weight the llms.txt highly – which the study also supports.

    You can find more on the assessment logic in our Methodology and in the overview of all Checks.

    FAQ

    Is the Zyppy study scientifically reliable?

    It is an expert survey with 131 participants and therefore a qualified consensus, not proof of causality. However, it is very informative for strategic priorities.

    What is the most important factor for Google AI Overviews?

    According to the study, access for AI crawlers and snippet suitability. Without retrievable content, all other measures are ineffective.

    Do I need an llms.txt?

    It doesn't hurt, but according to the study, it achieves little on its own. Crawlability, clear answers, and a strong brand are significantly more important.

    How do I check if my website is prepared?

    With the free AI Trust Audit, you will receive an evaluation of the most important factors for AI visibility in about one minute.