FAQPage (schema)

Technically, an FAQPage consists of a list of Question objects, each of which is linked via acceptedAnswer to exactly one Answer. Today, the whole thing is ideally delivered as JSON-LD in the <head> of the page. The distinction from a related type is important: FAQPage is intended for the site operator's own questions and answers, where the user has no way of answering. A community forum in which users answer themselves would use the QAPage type instead. This distinction is not a formality: anyone who chooses the wrong type risks search engines ignoring the markup or treating it as a guideline violation.

How FAQPage markup is structured

At its core, a minimal FAQPage object in JSON-LD looks like this:

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [{
    "@type": "Question",
    "name": "How long does shipping take?",
    "acceptedAnswer": {
      "@type": "Answer",
      "text": "Shipping within Germany takes 1 to 3 working days."
    }
  }]
}

The central building blocks at a glance:

• @type: FAQPage – declares the page or block as a collection of FAQs.
• mainEntity – the list of all questions on the page.
• Question.name – the exact question text as the user would ask it.
• acceptedAnswer.text – the complete answer; it must reflect the content visible on the page.

The text property of the answer may contain limited HTML, such as paragraphs, links or lists, so that longer, structured answers can also be represented cleanly. One basic rule is not negotiable: the markup must reflect the visible page content. Questions and answers that only appear in the JSON-LD but not on the page violate the guidelines and can lead to a manual action. Promotional questions or questions built purely for search that no real user would ask are equally taboo.

JSON-LD, Microdata or RDFa?

schema.org can be expressed in three syntaxes: JSON-LD, Microdata and RDFa. For FAQPage, as for structured data in general, JSON-LD is today the clearly recommended variant. The reason: JSON-LD sits as a self-contained block in the <head> or at the end of the page and is decoupled from the visible HTML. This makes maintenance more robust and avoids the scattered, error-prone attributes that Microdata writes into the middle of the body. Modern shop systems therefore rely consistently on JSON-LD.

What benefits FAQPage brings

The benefit lies on two levels. In classic search, FAQPage was able in the past to generate so-called rich results: expandable question-and-answer blocks directly in the search result. In August 2023, Google initially restricted the display of these FAQ rich snippets to well-known, authoritative government and health websites and abolished them completely on 7 May 2026: since then, FAQ rich results no longer appear in Google Search for any website. Google posted the deprecation notice on 8 May 2026 and removed the documentation on 15 June 2026. The markup is therefore no longer a route to a special display in classic search; however, Google has not asked for existing markup to be removed. Its value today lies in the machine-readable structure, and that is precisely where it has increased.

Because on the second level, AI visibility, FAQPage plays to its strengths. Answer engines such as Google AI Overviews, ChatGPT or Perplexity work on the basis of questions. A page that explicitly marks up its questions and answers provides these systems with precise, extractable answer building blocks, exactly the format that generative search needs for its answers. Clean FAQPage markup is therefore a building block of GEO (generative engine optimisation) in e-commerce.

The reason for this suitability lies in the nature of the matter: a good FAQ answer is already formulated so that it stands on its own. It answers a clearly defined question in a few sentences without needing the context of the rest of the page. It is precisely this property, self-containment and extractability, that an answer engine looks for when it assembles a building block for its answer. FAQPage additionally makes these already well-cut morsels machine-readable.

Real-world example: FAQ on a Shopware product page

A Shopware retailer selling espresso machines adds an FAQ block to the product page with questions such as "Is the machine suitable for ground coffee and pods?" and "How often do I have to descale?". If these pairs are marked up as FAQPage in JSON-LD, an answer engine immediately understands that the page answers exactly these purchase questions and can cite the shop as a source when a user asks ChatGPT or Google's AI. Without markup, the system would have to laboriously reconstruct the same information from the description text, which succeeds less often and less reliably.

The business effect is concrete: anyone who cleanly answers and marks up their customers' real pre-purchase questions increases the chance that their shop will appear as a source in the AI answer to precisely these questions, at the point in the customer journey where the purchase decision is made.

Implementing FAQPage in Shopware

Since version 6.7.9, Shopware has output structured data natively as JSON-LD, provided the corresponding feature flag is active. However, an FAQPage node is only created if question-and-answer content has actually been maintained; the flag alone does not create any questions. In practice, this means storing FAQ content on the product or in shopping experiences and ensuring that the output is valid. For types that the core does not cover, or for targeted population, retailers turn to a plugin or a custom extension.

Before going live, a test with the Schema.org validator is worthwhile. It checks whether the JSON-LD is syntactically correct and complete. This reveals before publication if, for example, an answer is missing or a field has been named incorrectly, errors that would otherwise lead search engines and AI systems to silently ignore the markup. Google's Rich Results Test no longer helps here: it has not supported FAQPage since June 2026, nor have the FAQ report and the FAQ filter in Search Console; support in the Search Console API ended in August 2026.

Typical mistakes and how to avoid them

In practice, FAQPage rarely fails because of the technology and often because of the implementation. The recurring pitfalls:

• Markup without visible content – questions only appear in the JSON-LD, not on the page. A clear guideline violation.
• Promotional pseudo-questions – "Why is product X the best?" is not a real user question and undermines credibility.
• Wrong type – use QAPage instead of FAQPage for user-generated answers.
• Duplicate markup – the same FAQ several times in the source code, which creates contradictory nodes.

FAQPage in combination with other schema types

FAQPage rarely stands alone. On a product page it complements the Product schema, on an advice page the Article schema, and for step-by-step instructions the related type HowTo comes into play, which, however, shares the same fate in Google Search: Google restricted HowTo rich results to desktop in August 2023 and abolished them completely in September 2023; as a structural type for AI systems, it remains useful. The distinction is simple: FAQPage answers self-contained individual questions, HowTo guides through a sequence of steps towards a goal. A question such as "How do I descale the machine?" can be either, depending on the level of detail: as a brief answer an FAQ entry, as an illustrated guide with several steps a HowTo. Anyone who separates the two cleanly gives search engines and AI systems a more precise picture of the page.

It is important to keep the various schema types on a page free of contradictions. Several JSON-LD blocks on one page are permitted and common, but they must not contradict each other. A clean set-up uses exactly the right type for each content type and avoids duplication, such as the same question once as FAQPage and once accidentally in a second block.

FAQPage and trust (E-E-A-T)

Good FAQ content also contributes to what Google summarises as experience, expertise, authoritativeness and trustworthiness (E-E-A-T). Anyone who answers their customers' real questions honestly, specifically and completely signals expertise and closeness to the customer. The markup alone does not create this trust, but it makes the existing quality of the content visible and attributable for machines. In a world in which AI answer engines increasingly decide which source to cite, precisely this combination of genuine substance and clean markup is the lever that decides visibility.

What makes a good, citable FAQ answer

The best markup is useless if the answer behind it is weak. A citable FAQ answer begins with the statement itself, names concrete facts and remains understandable even when extracted from the page. The difference is shown most quickly with an example.

Weak: "The delivery time depends on various factors and will be shown to you during the ordering process." This answer carries no information that an AI could cite; it postpones the answer to another point in time.

Strong: "Within Germany we deliver in 1 to 3 working days, to other European countries in 3 to 5 working days. Orders placed by 2 pm are shipped the same day." This answer is self-contained, factual and exactly what an answer engine takes over as a reliable building block.

In practice, this means: before marking up questions with FAQPage, you should sharpen the answers themselves. Every answer should name its topic in the first sentence, contain at least one concrete figure, deadline or condition and be able to stand without the surrounding page context. Only this content quality in combination with clean markup turns an FAQ block into a genuine GEO asset.

A final note on maintenance: FAQ content becomes outdated. Shipping times, return periods or compatibility information change, and an incorrect but cleanly marked-up answer is taken over by machines just as reliably as a correct one. Regularly checking the stored questions and answers is therefore part of content maintenance, precisely because FAQPage answers flow so effectively into AI answers.

Frequently asked questions about FAQPage

What is the difference between FAQPage and QAPage?
FAQPage represents questions and answers specified by the site operator. QAPage is intended for pages on which users ask a question and the community can give several answers.

Will I get rich snippets with FAQPage?
No. Google discontinued FAQ rich results completely on 7 May 2026, for all websites, including the government and health portals that had been the last exception since 2023. The benefit today lies in the machine-readable structure for AI answer engines.

Does the markup have to reflect the visible text?
Yes, without exception. Questions and answers in the JSON-LD must correspond to what the user sees on the page. Hidden markup violates the guidelines.

How many questions should an FAQPage contain?
There is no fixed limit, but quality comes before quantity. Genuine, frequently asked purchase questions with complete, self-contained answers make sense, not artificially inflated lists.

Does FAQPage replace a good description text?
No. It additionally structures existing content. Citable running text and valid schema markup complement each other.

In which format should I deliver FAQPage?
As JSON-LD. It is recommended by Google, more robust to maintain than Microdata and decoupled from the visible HTML.