FAQ schema in 2026

Google removed FAQ rich results from most search listings in 2023, which led a lot of people to strip FAQPage markup out entirely. That was a mistake. The rich result is gone; the machine-readable question-answer pairing is still read by answer engines, and it is one of the cleanest ways to hand a model an extractable answer.

What changed and what didn't

What changed: FAQ rich results — the expandable questions under a search listing — now appear only for a small set of authoritative government and health sites. For everyone else, adding FAQPage markup will not change how the listing looks.

What did not change: the markup still explicitly pairs a question with its answer in a structure that retrieval systems parse trivially. That is exactly the shape a generative engine wants when it is looking for a passage to quote.

How to write FAQs worth marking up

Use real questions. The ones from sales calls, support tickets and the 'People also ask' box beat invented ones every time, because they match how people actually phrase things.

Answer in two to four sentences, completely, with a checkable fact where one exists. An answer that says 'it depends on your needs' is worse than no answer, because it teaches the model that your page does not resolve the question.

Include the uncomfortable questions — pricing, limitations, who the product is not for. Pages that address objections are cited disproportionately often, because models favour sources that read as balanced.

Implementation rules

The markup must match visible page content exactly. Marking up questions that a visitor cannot see is a guidelines violation and gets the markup ignored.

One FAQPage block per page. Do not duplicate the same questions across many pages — pick the page that owns each question and link the others to it.

When to skip it

Skip it when the questions are padding. Three genuine questions beat twelve manufactured ones, and a page stuffed with thin Q&A reads as low quality to both people and models.

Skip it on pages where the whole page is already the answer to one question — there, answer-first structure does the job and an FAQ block just repeats it.

What we measured on real domains

Everything below comes from ThoughtForge runs against live domains, not from a study we read. Each run scores three things out of 100: identity (can an engine say who you are without contradicting itself), durability (does your brand survive a multi-turn conversation), and retrieval (can your pages be chunked into something quotable).

Across every domain we have scanned, one number is consistently the worst, and it is not the one people optimise for. Durability sits at or near zero almost everywhere. Brands that get named in turn one disappear by turn three, because nothing on the site gives the model a reason to keep holding onto them.

  • thoughtforge.app, July: identity 16, durability 16, retrieval 49, overall 27.
  • thoughtforge.app, September, after entity and schema work: identity 52, durability 0, retrieval 48, overall 33.
  • A services domain (megap•••••••••••.com.au): identity 28, durability 0, retrieval 47, overall 25.
  • An industry association (propertyasso•••••••.com.au): identity 4, durability 0, retrieval 50, overall 18.
  • A consultancy (obse•••••••••.com): identity 40 in July, 52 in September; overall 30 to 35.

What FAQ markup is and is not worth

Google retired most FAQ rich results, so the search-listing payoff is largely gone. The reason to keep the markup is different: it pairs a question with a self-contained answer in a machine-readable block, which is exactly the shape an answer engine wants to lift.

It will not save a weak page. The association we scanned at identity 4 has perfectly valid markup on several pages. Markup describes structure; it cannot supply a clear identity that the site never states.

Write the questions people actually ask

Source them from support tickets, sales calls and search queries, not from a keyword tool. A question nobody asks produces an answer nobody quotes.

Keep each answer self-contained and under about 80 words, with the specifics inside it. The value of the block is that question and answer travel together in a form an engine can lift whole.

Where FAQ markup goes wrong

Marking up questions that are not on the page, burying answers in accordions that never render, and padding a thin page with ten invented questions instead of writing the one thorough answer the topic needed.

If a page needs FAQ markup to feel substantial, the page is the problem.

Frequently asked questions

Should I remove existing FAQPage schema?

No. Removing valid markup gains nothing and loses the machine-readable pairing that answer engines use.

Do FAQs still win featured snippets?

Yes. The question-and-direct-answer pattern is still the most common source of paragraph snippets, independent of the schema.

How many questions is right?

Three to eight genuine ones. Beyond that you are usually padding, and padding is visible.

Related guides

  • Answer-first content structure — The answer-first structure that gets pages lifted into AI answers and featured snippets, with the exact block pattern to use.
  • Entity clarity and schema — How to make an AI model certain about what your brand is, using Organization and Person schema, sameAs links and consistent naming.
  • AEO vs SEO vs GEO — SEO wins rankings, AEO wins the answer, GEO wins the citation. What each term actually means, how they overlap, and how to measure all three.

All guides · Blog · Case studies