Is FAQ Schema Dead in 2026? What Actually Changed

Since May, a claim has been circulating in SEO discussions that goes roughly like this: FAQ schema is dead, Google killed it, and leaving it on your pages will get you penalized.

The first half is an overstatement. The second half is simply false, and it's causing people to strip working markup off their sites for no reason.

Here's what actually happened, what Google actually said, and what the markup still does.

What changed

On May 7, 2026, Google added a deprecation notice to its FAQ structured data documentation. There was no blog post and no announcement, just a note on a developer page. Three things followed from it, confirmed across multiple outlets tracking the rollout:

• May 7, 2026: FAQ rich results stopped appearing in Google Search.
• June 2026: The FAQ search appearance filter, the rich result report, and FAQ support in the Rich Results Test are removed from Search Console.
• August 2026: Search Console API support for FAQ rich result data is removed. Anyone pulling that data into dashboards or BigQuery exports needs to update those calls or start silently receiving empty values.

That's the entire change. A search appearance was retired, along with the tooling that reported on it.

The part most articles are still getting wrong

A lot of posts on this topic, including ones published after May, say FAQ rich results are still available to authoritative government and health sites.

That was true between August 2023 and May 2026. It isn't true now. The 2023 change restricted FAQ rich results to well-known, authoritative government and health websites. The May 2026 change removed eligibility entirely, including for those sites. Nobody gets FAQ rich results anymore.

If you're reading advice that still describes the government-and-health carve-out as current, it was written before May 2026 or copied from something that was.

FAQPage went. Q&A didn't.

This distinction gets lost constantly, and losing it costs people rich results they're still eligible for.

FAQPage is for a page where the site publishes its own questions and its own answers. That's the one that lost its rich result.

Q&A is for a page where users submit multiple answers to a question, a forum thread, a community help page. It's a different schema type and it's still supported in Google's structured data gallery.

So are Education Q&A, for education questions and answers that surface as flashcards, and Discussion forum, for user-generated content with threaded discussion.

If your page genuinely has community-submitted answers, you may be marking it up as the wrong type and forfeiting an eligible rich result. Google still distinguishes FAQPage from QAPage in its documentation, this was never a blanket removal of question-oriented structured data.

Myth: Google penalizes FAQ schema

This is the claim worth killing outright, because it's the one causing actual damage.

Google deprecated a search feature. It did not deprecate the markup. FAQPage remains a valid schema.org type. Leaving it on your pages will not throw errors, will not trigger a manual action, and will not damage your rankings. Google's own documentation update specifically notes the markup can stay in place, and Google has said previously that unused structured data doesn't cause problems for Search.

Google has also stated that rankings aren't affected by the deprecation at all. This was a search-appearance change, not an algorithmic one. Pages that previously benefited from the larger FAQ dropdown may see click-through shift, because they now occupy less space in the results, but their ranking positions weren't adjusted because of it.

So: no penalty. No manual action. No ranking hit. If someone tells you otherwise, ask them to point at the documentation.

The one thing that can hurt you is unchanged from before: marking up questions and answers that aren't visible on the page. That was a policy violation in 2022 and it's a policy violation now. It has nothing to do with the deprecation.

What the markup still does

Google's SERP is one consumer of structured data. It isn't the only one.

Bing processes it independently. Bing continues to parse FAQPage markup, and Bing's Webmaster Guidelines still support FAQ rich results. DuckDuckGo, which runs on Bing's index, and Brave Search process it unchanged. Google removing its own feature was a Google decision, not an industry-wide one.

Microsoft has said structured data helps its AI. In March 2025, Fabrice Canel, Principal Product Manager at Microsoft Bing, confirmed that structured data helps Microsoft's large language models understand content for Copilot. That's one of very few on-record statements from any vendor about how structured data feeds AI systems, and it's specific.

AI crawlers read it. FAQPage markup remains crawlable by Bingbot, PerplexityBot, and the retrieval crawlers that feed answer engines. What weight each system gives it afterward isn't publicly documented (nobody outside those companies can tell you the exact weighting), but the markup is being read, and structured Q&A pairs are easier for a machine to extract cleanly than the same content buried in prose.

It removes ambiguity for machines. This is the underlying point that survives every deprecation. FAQPage markup explicitly labels one block of content as a question and another as its answer, rather than leaving a crawler to infer the relationship from formatting. That's the entire function of structured data: converting a visual convention that humans parse instantly into something a machine reads unambiguously. Google retiring a dropdown didn't change how the labelling works.

The honest limit: none of this is measurable the way the old dropdown was. You could see the FAQ accordion in Google or not see it. You cannot directly observe whether FAQPage markup made Perplexity more likely to quote your page. That uncertainty is real, and it's why we'd frame keeping the markup as a low-cost, sensible default rather than a guaranteed return.

Why Google retired it

Understanding the reasoning helps calibrate how much to worry, which is: not much.

Google hasn't published an exhaustive rationale, but the pattern across its recent retirements is consistent and the FAQ case fits it exactly. FAQ rich results were among the most heavily abused structured data features. Because the dropdown granted extra vertical space in the results, pushing competitors down, sites bolted FAQ sections onto pages that weren't really FAQs, stuffed them with keyword-bait questions, and marked up content that barely qualified. The feature stopped reliably signalling "this page genuinely answers these questions" and started signalling "this page wants more SERP real estate."

Google's typical response to that arc is to narrow eligibility, then remove the feature. FAQ followed it precisely: restricted to authoritative government and health sites in 2023, removed entirely in 2026. It's the same arc HowTo schema went through: rich results retired from every surface entirely by 2026 after years of narrowing eligibility first. The markup specification was never the problem; the gaming of the visual feature was.

That history matters for the penalty myth. Google didn't retire FAQ rich results to punish anyone, it retired a feature whose signal had degraded and whose space it wanted back for its own AI answers. Reading "Google removed the feature" as "Google is now hostile to FAQ markup" gets the motivation backwards. The tags are fine. The dropdown outlived its usefulness to Google.

The honest counter-argument

Some experienced practitioners now argue you shouldn't bother, that FAQPage markup earns nothing visible in Google, that the effect on AI systems is unmeasured, and that effort is better spent elsewhere. Some go further and recommend removing it from commercial sites entirely.

That position is reasonable. If your FAQ section existed only to win extra pixels in the SERP, that play is genuinely over, and no amount of reframing brings it back.

We should also be upfront about our own position: we make a schema generation plugin. "Keep using schema" is exactly what you'd expect us to say, and you should weigh our argument accordingly.

Here's the most useful thing anyone said during this whole episode, and it cuts against our commercial interest: the schema was never doing the work: the content was. A vague FAQ section stuffed with keyword-bait questions didn't earn citations because it had markup. Markup made an already-good answer easier for a machine to identify. If your FAQ content is weak, adding FAQPage to it accomplishes nothing, and it accomplished nothing before May 2026 either. The deprecation just removed the cosmetic cover.

What to actually do

Leave existing FAQPage markup in place. Removing it gains you nothing, costs you time, and forfeits whatever value non-Google systems derive from it.

Audit it for accuracy. Confirm every marked-up question and answer appears visibly on the page. Remove orphaned markup left behind on pages whose FAQ sections were deleted. This is the one genuine risk area, and it predates the deprecation.

Export your historical data before the deadlines. If you want a before-and-after baseline, pull FAQ rich result data out of Search Console before the reporting disappears, and update any API or BigQuery pipelines before August 2026 so they don't start returning empty values without anyone noticing.

Change what you use the Rich Results Test for. It's no longer a "will this trigger the dropdown" check for FAQs. Treat FAQPage validation as ordinary structured data QA: valid syntax, matches visible content, no manufactured questions.

Stop writing FAQs for the SERP. Write them because people genuinely ask those questions and a clear, self-contained answer serves them. That's what makes content citable by an answer engine, with or without markup.

Keep your effort proportionate. For most sites this is a hygiene exercise, not a strategy pivot. Meanwhile, Google's structured data gallery still supports dozens of other features, including Product, Local business, Organization, Article, Recipe, Review snippet, Breadcrumb, Video, Event, Job posting, Q&A, and Course list. If you want visible SERP features, that's where the return is. We've mapped the complete current list in Every Schema Type Google Supports in 2026, or browse the full schema types directory to see what applies to your site.

How to audit your FAQ markup now

If this post prompts one action, make it an accuracy audit rather than a removal. The deprecation didn't create a risk, but it's a good moment to check for one that predates it: markup describing content that isn't visibly on the page.

Find every page with FAQPage markup. A site-wide structured data crawl is fastest; for a small site, a site: search combined with viewing source works. You want the full list before you start.

Check each marked-up Q&A against the rendered page. Every question and answer in your JSON-LD should correspond to a question and answer a visitor can actually see. The common failure: an FAQ section was edited or removed from the visible page, but the markup was left behind. That orphaned markup is the one genuine policy risk, and it existed before May 2026 too.

Confirm you're using FAQPage, not QAPage, correctly. FAQPage is for questions you answer yourself. If your "FAQ" is actually community-submitted questions and answers, QAPage is the right type, and it still produces a rich result. Using the wrong one either forfeits an eligible result or misrepresents your content. We cover the distinction in detail separately, but the quick test is: did you write the answers, or did your users?

Update your reporting expectations. Stop treating the Rich Results Test as a "will this trigger the dropdown" check for FAQs, that dropdown is gone. Treat FAQPage validation as ordinary structured-data QA: valid syntax, matches visible content, no manufactured questions. And if you pulled FAQ rich-result data from Search Console into a dashboard, update that pipeline before the API support ends, or it will quietly start returning nothing.

Decide about new FAQ sections on their own merits. Write FAQs because real people ask those questions and a clear, self-contained answer serves them, which also happens to be what makes content extractable by an answer engine. Don't write them to chase a Google feature that no longer exists, and don't bolt a Q&A block onto a page that isn't genuinely a FAQ just to have the markup.

This is a half-hour of checking on most sites, and it leaves you with markup that's accurate, correctly typed, and doing whatever useful work it can across the systems that still read it.

The short version

FAQ schema isn't dead. A Google search feature is dead.

The markup is still valid, still unpenalized, still parsed by Bing and by AI crawlers, and still the clearest way to tell a machine which text is a question and which is its answer. What's gone is the expandable dropdown under your listing. And if that was the only reason you were using it, then for you, the honest answer is that this particular play is over.

For everyone else: leave it, keep it accurate, and put your energy into answers worth citing.

AI Schema Gen generates valid FAQPage markup on every blog post, along with 530+ other schema types, directly from your page content, and validates it before publish. Start free or see plans and pricing.