You added a ## FAQ section to 600 articles. Your ArticleLayout extracts it and emits FAQPage JSON-LD. You wait six weeks, the FAQ rich result never shows in SERP, and when you paste a URL into the Rich Results Test it says the FAQ feature is no longer supported. Before you spend a day debugging your extractor: in almost all cases as of June 2026 there is nothing wrong with your schema. Google deprecated the FAQ rich result.
TL;DR. Google stopped showing FAQ rich results in Search on May 7, 2026, for every site (it had already been restricted to a handful of well-known government and health sites since August 2023). FAQPage is still a valid schema.org type and Google says the markup can stay on your pages without causing problems — it just no longer produces a visible result in Google Search. So: confirm you are hitting the deprecation (not a real bug), then decide whether to leave the markup for AI/other-engine parsing or strip it. If your “FAQ” content is genuinely useful Q&A, keep the visible section either way.
First: are you sure it’s the deprecation and not a real bug?
Run through this once. If you match the first row, stop — the rest of the schema-debugging steps below do not apply to you.
| What you observe | Most likely cause | What to do |
|---|---|---|
| Rich Results Test says FAQ is “no longer supported” / no FAQ section in the report | Deprecation (May 7 2026) | Go to “Decide: keep or remove” below |
| Search Console FAQ Enhancements report disappeared | Report retired (June 2026) | Expected; nothing to fix |
| Rich Results Test says “No items detected” but your JSON-LD is in source | Extractor emitted empty / malformed JSON-LD | Validate the JSON-LD (Schema Markup Validator), see “If your JSON-LD is actually broken” |
view-source shows zero FAQPage block | Extractor never ran / regex didn’t match | See “If your JSON-LD is actually broken” |
Two or more FAQPage blocks in source | Duplicate emission (layout + plugin) | Dedupe, see below |
The point of this table: since the SERP feature is dead for everyone, “the rich result isn’t showing” is no longer diagnostic of a schema problem. Only use the source/validator checks to tell whether your markup is technically valid, not whether it will earn a SERP feature — it won’t.
What actually changed (timeline)
Google retired FAQ rich results in phases. As of June 2026:
- August 2023 — FAQ rich results restricted to “well-known, authoritative government and health websites” only. Most sites lost the feature here.
- May 7, 2026 — FAQ rich results stopped appearing in Google Search for everyone, including the gov/health sites that were spared in 2023.
- June 2026 — Google removes the FAQ search-appearance filter in the Performance report, the FAQ rich result report (Enhancements) in Search Console, and FAQ support in the Rich Results Test.
- August 2026 — FAQ rich-result data removed from the Search Console API.
Google’s official line: you can remove the FAQ structured data if you want, or leave it; it won’t cause problems, and other search engines and crawlers may still process it for their own purposes. See Google’s FAQ structured data documentation, which now carries the deprecation notice.
Decide: keep or remove the FAQ markup
There is no SEO penalty either way. Pick based on what you want from it.
Keep the JSON-LD if your Q&As are real, visible on the page, and useful. FAQPage is a clean question/answer structure that AI answer engines (and non-Google crawlers) can still parse, so leaving it costs nothing and may help AI citation / answer-engine optimization. Note: Google has not confirmed FAQ markup as an AI Overviews ranking signal, so treat AI upside as “plausible, unproven,” not a guarantee.
Remove the JSON-LD if it was only ever there to chase the SERP dropdown, the answers are thin/boilerplate, or you want to slim your <head> and reduce maintenance. Removing it does not hurt rankings.
Either way, keep the visible ## FAQ section if the questions are ones real readers ask. A useful on-page FAQ helps humans and gives AI answer engines extractable Q&A whether or not you emit JSON-LD.
If you choose to keep the markup: confirm it’s technically valid
You no longer get a SERP feature, but you still want valid JSON-LD so other parsers don’t choke. The Rich Results Test drops FAQ in June 2026, so validate with the Schema Markup Validator instead — it checks schema.org types directly and isn’t tied to Google’s SERP features.
A correct shape
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "Why does my image alt text matter?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Alt text is read by screen readers and used by image search engines..."
}
}
]
}Common structural mistakes that make it invalid FAQPage (even though it’s valid JSON): acceptedAnswer written as answer; the answer’s @type written as Text instead of Answer; mainEntity missing or written as a single object instead of an array.
Your extractor (single source of truth)
Keep one extractor in ArticleLayout and no plugin duplicating it. This is the shape the site already uses:
function extractFaq(bodyMdx: string) {
const items = [];
const re = /^### (.+\?)\s*\n+((?:(?!^### |^## ).+\n?)+)/gm;
let m;
while ((m = re.exec(bodyMdx))) {
items.push({
"@type": "Question",
name: m[1].trim(),
acceptedAnswer: {
"@type": "Answer",
text: m[2].trim().replace(/\n+/g, " "),
},
});
}
if (items.length === 0) return null;
return {
"@context": "https://schema.org",
"@type": "FAQPage",
mainEntity: items,
};
}Emit it once:
{faq && (
<script type="application/ld+json" set:html={JSON.stringify(faq)} />
)}Keep the JSON-LD text faithful to the visible answer (strip markdown, don’t truncate to one sentence) so the markup and the page agree:
function stripMarkdown(s: string) {
return s
.replace(/`([^`]+)`/g, "$1")
.replace(/\*\*([^*]+)\*\*/g, "$1")
.replace(/\[([^\]]+)\]\([^)]+\)/g, "$1");
}Don’t emit two FAQPage blocks
If a plugin and your layout both emit FAQ schema, you get duplicate blocks. Count them in built output:
curl -s https://site.com/en/articles/foo/ | grep -c '"@type": "FAQPage"'If the count is above 1, find every emitter and keep exactly one:
grep -rn '"FAQPage"' src/If your JSON-LD is actually broken (empty or missing)
This is the one scenario where you really do have a bug. Symptoms: view-source shows no FAQPage block, or it shows "mainEntity": [].
- Empty
mainEntity— your FAQ section structure doesn’t match the extractor’s regex. The example regex expects### Question?headings followed by paragraph text. If your articles use**Question?**followed by bullet lists, no question matches. Standardize the section format or broaden the regex. - No
FAQPageblock at all — the extractor returnednull(no questions found) or the layout branch never rendered. Confirm the article actually has a## FAQsection with at least one### Question?subheading, and thatextractFaqis called for that route.
Verify the page is also indexable while you’re in view-source — this matters for normal indexing even though it no longer gates a FAQ rich result:
curl -s https://site.com/en/articles/foo/ | grep -E 'robots|canonical'Robots must not be noindex; the canonical should point at the page itself, not elsewhere.
How to confirm it’s fixed
- Schema Markup Validator on a sample URL reports a valid
FAQPagewith the expected number ofQuestionitems (if you kept the markup). view-sourceshows exactly oneFAQPageblock per page (zero if you removed it).- You are no longer waiting on a SERP feature — that’s the headline. The “Eligible/Valid” status and the FAQ Enhancements report are gone as of June 2026; their absence is expected, not a regression.
FAQ
Did Google really kill FAQ rich results, or are they just down for my site?
It’s a real, permanent deprecation. FAQ rich results stopped appearing for everyone on May 7, 2026, after being limited to gov/health sites since August 2023. It’s not a per-site bug, not a manual action, and not something you can fix by tweaking schema.
Should I remove the FAQPage JSON-LD from my pages now?
It’s optional. Google says the markup can stay without causing problems, and other crawlers / AI answer engines may still use it. Remove it if it was only there to chase the SERP feature or if you want a leaner <head>; keep it if your Q&As are genuine and you want them parseable by non-Google systems. There’s no ranking penalty either way.
Why does the Rich Results Test no longer detect my FAQ schema?
Because Google removed FAQ support from the Rich Results Test in June 2026. To check that your markup is still structurally valid, use the Schema Markup Validator (validator.schema.org) instead, which validates schema.org types independently of Google’s SERP features.
My Search Console FAQ Enhancements report vanished — is that a problem?
No. The FAQ rich result report and the FAQ search-appearance filter were retired in June 2026, and the API data follows in August 2026. The report disappearing is the expected outcome of the deprecation, not lost data on your end.
Does keeping FAQ schema still help with AI search or AI Overviews?
Possibly, but it’s unproven. The question/answer structure is easy for AI answer engines to extract, and FAQ-formatted content tends to get cited, so keeping real Q&As is reasonable for answer-engine optimization. Google has not confirmed FAQ markup as an AI Overviews ranking signal, so treat any AI benefit as plausible rather than guaranteed.
Is FAQPage still valid schema.org markup?
Yes. Only Google’s visible rich-result feature was deprecated. FAQPage remains a valid schema.org type, so the markup itself is not an error and validators will still accept it.
Prevention
- One source of truth for FAQ JSON-LD (the layout’s extractor, no plugins) so you never ship duplicate blocks.
- Validate the extractor output against the Schema Markup Validator on template changes, not the Rich Results Test (FAQ support removed June 2026).
- Prebuild script counts emitted
FAQPageblocks per page and fails on more than 1. - Articles with a
## FAQsection auto-checked for at least one### Question?subheading so the extractor never silently emits an empty array. - Stop tracking the (now-removed) FAQ Enhancements report; if you still want a Q&A signal, monitor AI/referral traffic to pages with strong FAQ content instead.
Related
- Content Site Image Alt Missing Bulk
- Content Site Canonical Points to Self Wrong
- Content Site Hreflang Tags Misconfigured
- Search Console Low Value URLs
- Many Pages Few Impressions
- Content Site Translation Pages Mismatched
Tags: #Content ops #Site quality #Site audit #Troubleshooting #FAQ schema