What Is FAQ Schema? Markup, Eligibility, and Worth

Jul 16, 2026 · SEO Glossary

An honest article about FAQ schema has to lead with the 2023 plot twist, because most advice you'll find predates it. In August 2023, Google restricted FAQ rich results — the expandable Q&A dropdowns under a search listing — to "well-known, authoritative government and health websites." For everyone else, the visual reward that made this markup famous is gone. The markup itself, though, still does things. So let's cover what it is, what changed, and the sober case for still using it.

What is FAQ schema? It's structured data — machine-readable annotation, almost always JSON-LD — that identifies question-and-answer pairs in your page content, using the FAQPage type from schema.org. It tells parsers "these exact strings are questions, and these are their accepted answers," removing any need to infer that from your HTML layout.

What FAQ schema markup looks like

A minimal, valid example — a script block in the page's HTML:

{"@context": "https://schema.org", "@type": "FAQPage", "mainEntity": [{"@type": "Question", "name": "Do you ship to Ireland?", "acceptedAnswer": {"@type": "Answer", "text": "Yes, standard shipping to Ireland takes 3-5 business days and costs 8.95."}}]}

Wrapped in <script type="application/ld+json">, with one Question object per Q&A pair in the mainEntity array. The structural rules that trip people up:

• The questions and answers must be visible on the page. Markup describing content users can't see violates the guidelines — schema annotates content, it doesn't replace it.
• FAQPage is for pages where the site asks and answers its own questions. If users submit answers (a forum thread, a product Q&A section), the correct type is QAPage instead.
• One FAQPage per page, however many questions it holds.
• Answers can carry limited HTML (links, lists, emphasis) inside the text field, but keep them self-contained — each answer should stand alone when extracted.

What the 2023 change did and didn't remove

What it removed: the rich result. Sites that had enjoyed the expanded SERP real estate lost it in one update. If your analytics show a step-drop in organic CTR around August 2023 on FAQ-marked pages, this is why, and no amount of markup tuning brings it back — eligibility is now about who you are, not how you've marked up.

What it didn't remove:

• The parsing benefit. Structured Q&A pairs remain unambiguous machine-readable statements about your content. Search systems — and increasingly, AI-powered answer engines that synthesize responses from crawled pages — consume exactly this kind of clean, attributed question-answer data. That's not a promised placement, but it's cheap legibility.
• The content discipline. Writing real FAQ content forces you to collect actual customer questions and answer them concisely. That content can also surface in ordinary ways: as regular snippets, and occasionally in featured snippets or People Also Ask when an answer happens to match a query well.
• Other schema types' rich results. Product, review, recipe, event, and how-to-adjacent markup kept their treatments (with their own eligibility rules). The 2023 change was specifically about FAQ and HowTo display, not structured data generally.

So should you still add it?

A cost-benefit that's actually honest: the cost is near zero if your platform generates it (most CMS SEO plugins emit FAQPage automatically from FAQ blocks), and small if you hand-roll JSON-LD. The benefit is machine legibility plus whatever answer-engine visibility that buys, with no rich-result payoff for ordinary sites. Verdict:

• Add it when the page genuinely has FAQ content serving users — the marginal cost rounds to zero.
• Don't manufacture FAQ sections just to carry markup. Bolting three filler questions onto every product page adds noise for users and nothing measurable for rankings — the markup was never a ranking factor even when the rich result existed.
• Government and health-authority sites: the rich result is still yours; implement carefully and it still pays in SERP real estate.

Validation, and the errors that actually occur

Test with Google's Rich Results Test (eligibility view) and the schema.org validator (pure syntax view). The recurring failures, from years of audits: invalid JSON from an unescaped quote inside an answer string — one typo silently kills the whole block; markup declaring Q&As that aren't in the visible content, usually after a redesign removed the FAQ section but left the plugin emitting the old data; two plugins both outputting FAQPage on the same URL; and QAPage content mislabeled as FAQPage. Because these break silently — the page looks fine, the JSON just stops parsing — structured data errors are worth catching in routine monitoring rather than by surprise; running your pages through a scheduled site audit alongside Search Console's enhancement reports keeps a plugin update from quietly invalidating every marked-up page on the site.

Search Console's side of that monitoring deserves a specific mention: under Enhancements, sites Google considers eligible get an FAQ report listing valid items, items with warnings, and invalid ones — with the detected markup shown per URL. Since the 2023 change, many ordinary sites no longer see the report at all, which itself confirms what the restriction means for you in practice. Where the report does exist, a sudden drop in valid items after a deploy is your earliest signal that a template change broke the JSON — typically days before anyone would notice by hand.