FAQ schema is one of the few pieces of structured data whose purpose changed completely without most people noticing. In August 2023 Google announced that FAQ rich results would be limited to well-known authoritative government and health websites, which removed the expandable question list from search results for essentially every ordinary business site. A lot of teams responded by ripping the markup out. That was the wrong lesson.
Does FAQ schema still do anything?
Yes, but not the thing it used to do. It no longer earns the expandable rich result for ordinary sites, and it will not add pixels to your listing. What it still does is state, in machine-readable form, that a specific block of text on your page is the answer to a specific question — which is precisely the pairing a retrieval system is looking for when it decides what to quote.
Think of it as labelling rather than decoration. The markup does not create value; it makes existing value legible. If the answer underneath is vague, the schema faithfully labels vagueness. If the answer is sharp and self-contained, the schema tells every parser exactly where the boundaries of that answer are, so nothing gets clipped mid-thought.
What makes an FAQ answer quotable?
An answer is quotable when a stranger could read it in isolation, understand it completely, and act on it. Most FAQ blocks fail that test in the first sentence, because they open with a throat-clearing preamble instead of the answer.
| Question | Answer that never gets lifted | Answer that does |
|---|---|---|
| Do I need GST registration to sell online in India? | GST rules can be complicated and it is best to consult a professional about your specific situation. | Selling through a marketplace that collects tax at source generally triggers a registration requirement that does not depend on your turnover, whereas selling only through your own website follows the ordinary turnover thresholds. This is general information rather than tax advice — confirm your own case with a chartered accountant, as thresholds differ by state and by whether you supply goods or services. |
| How long does a website redesign take? | Every project is different, so timelines vary depending on scope and requirements. | The build itself is rarely the long pole. Content approval, brand sign-off and third-party integrations set the real timeline, so a redesign with copy already written moves several times faster than one where the copy is still being argued about. |
| What is the difference between a chatbot and a voice agent? | Both use AI to help customers and each has its own advantages and disadvantages. | A chatbot handles typed conversations where the user can see previous messages and correct themselves; a voice agent handles spoken calls where nothing is visible and interruptions have to be handled in real time. The voice case needs turn-taking, latency and accent handling that the chat case never has to solve. |
The pattern in the right-hand column is consistent: the first clause commits to a position, the specifics are named, and the caveat comes last rather than first. Caveats at the top destroy quotability, because a summariser reading the opening sentence finds nothing to summarise.
Where should the FAQs live on the page?
Inside the visible page, in the order a reader would reach them, and marked up exactly as they appear. Google's guidance on FAQPage is explicit that the questions and answers must be visible content on the page, written by the site rather than submitted by users, and not used to carry promotional text.
Two placements work well in practice. Mid-article FAQs handle objections at the point they arise, which helps human readers more than a block at the bottom does. End-of-article FAQs catch the adjacent questions the piece did not cover, and are the natural place to link onward to the pages that do cover them. Either way, the markup follows the visible text — never the reverse.
What does correct FAQ markup look like?
Minimal, and matching the page word for word. Here is the shape, with entities escaped for display:
<script type="application/ld+json">
{ "@context": "https://schema.org", "@type": "FAQPage", "mainEntity": [{ "@type": "Question", "name": "Does FAQ schema still earn rich results?", "acceptedAnswer": { "@type": "Answer", "text": "No. Google limited FAQ rich results to well-known government and health sites in 2023, so ordinary sites no longer see the expandable listing." } }]
}
</script>The answer text field accepts limited HTML, and unescaped tags are a common reason a block fails validation. If your CMS injects markup into that field, sanitise it at render time — a five-minute fix for whoever handles your website build, and one that quietly breaks otherwise correct implementations.
Free consultation
Have your FAQ markup checked before it costs you
We will validate your structured data, flag the answers that no engine will ever quote, and rewrite two of them as examples. No obligation.
What are the most common FAQ schema mistakes?
- Markup that does not match visible text. The classic version is a marketing team editing the page copy while the JSON-LD stays frozen at whatever the developer pasted in a year ago.
- The same FAQ block on every page. Identical questions sitewide tell a parser nothing about any individual page, and they create the duplicate-content pattern covered in our post on duplicate content.
- Promotional answers. An answer that turns into a pitch by its second sentence is both against Google's stated guidance and useless to lift.
- Questions nobody asks. "Why choose us for the best services in India?" is not a question. Take real ones from your sales inbox, your call notes and your Search Console query report.
- FAQs stapled to category and product pages. On e-commerce templates this tends to produce hundreds of near-identical blocks; the effort belongs in the product content itself, as we set out in our e-commerce SEO work.
How many FAQs should a page have?
As many as the page genuinely raises, which is usually between four and eight. Beyond that, the questions start being invented to fill the block, and invented questions are obvious in a way that damages the credibility of the real ones.
A more useful discipline than counting is auditing: every quarter, pull the questions your sales team actually got asked and compare them against what is on the page. Questions that appear repeatedly in real conversations and nowhere on your site are the highest-value content you can write, and they usually deserve their own page rather than a line in an accordion.
How this fits the wider answer-engine picture
FAQ schema is one tactic inside a larger shift: search surfaces increasingly want bounded, attributable answers rather than pages to browse. The same writing discipline that makes an FAQ quotable makes a whole article quotable, which is why we treat it as one workstream in our SEO engagements rather than a structured-data task. The wider mechanics, including what Google grounds its summaries in, are covered in how AI Overviews pick their sources and in our answer engine optimisation guide.
None of this is legal, tax or compliance advice — where an FAQ touches regulation, say plainly that it is general information and point the reader to a qualified professional. That sentence costs you nothing and protects both the reader and you.
Free consultation
Rewrite your FAQs around the questions buyers actually ask
Send us your sales inbox questions and your current FAQ block. We will show you which ones are worth a full page and which just need a better answer. No obligation.