Because Google switched them off. Google's Search Central documentation changelog says the FAQ rich result stopped appearing in Google Search starting 7 May 2026, and Google removed the FAQ documentation entirely on 15 June 2026. The markup on your page did not break. The feature it fed no longer exists.
We get asked about this a lot, usually by someone who paid for FAQ schema to be added a year or two ago and now cannot find any trace of it in the results. Nothing is broken on their site. The search feature was retired.
The interesting question is what to do next. Do you strip the markup out, leave it alone, or spend that effort somewhere else? Our answer is not the one most SEO checklists still give.
FAQ schema is structured data using the schema.org FAQPage type. You add it to a page as JSON-LD, listing each question and its answer in a format machines can read. For years, Google used it to show expandable question dropdowns directly under your listing in the search results.
The value was always about search real estate. A normal listing takes up one slot. A listing with four FAQ dropdowns underneath it took up far more of the screen, which pushed competitors down the page.
That is why FAQ schema spread so fast. It was one of the few pieces of markup that visibly changed how your result looked, and it did not require any change to the page a human sees.
Google narrowed the feature first, then killed it. Its documentation limited FAQ rich results to well-known, authoritative government and health websites back in 2023, so most sites stopped seeing them then. The final step came in 2026, when Google's changelog confirmed the feature would no longer appear in Search from 7 May 2026.
The 8 May 2026 changelog entry added a deprecation notice to the FAQ documentation. The 15 June 2026 entry then removed that documentation altogether, noting that the FAQ rich result is no longer shown in Google Search results.
So there were two separate moments, roughly three years apart. Most sites lost the feature in 2023 and did not notice, because it faded rather than vanished. The 2026 change is the one that made it official for everyone, including the government and health sites that had kept it.
If you have an SEO audit or a template that still recommends FAQ schema for rich results, it is out of date. There is no version of Google Search in 2026 that renders those dropdowns.
Not useless, but no longer a search feature play. FAQPage is still a valid schema.org type. It still describes your content accurately to anything that reads structured data. What it no longer does is change how your listing looks in Google, which was the entire reason most people added it.
We think that distinction matters because a lot of teams treat structured data as a slot machine. You add markup, Google gives you a visual reward. That model was never the point of schema.org, and Google has now removed three of the rewards people chased hardest.
The healthier way to think about structured data is as a description of what your page contains. Our fuller take on how schema markup actually helps your SEO goes deeper on that, but the short version is that markup should mirror the page, not decorate it.
No, and this is the practical answer most people want. Leaving valid FAQPage markup in place costs you nothing. It is a small block of JSON-LD, it does not slow the page in any meaningful way, and Google does not penalise markup for a retired feature. Removing it across hundreds of pages costs time you could spend better.
There is one exception we do act on. If the FAQ markup does not match visible content on the page, take it out. Google's structured data guidelines have always required markup to reflect what a user can actually see, and invisible FAQ blocks added purely for the rich result now have zero upside and a real policy risk.
We audit for that specific mismatch on every migration we run. Hidden question and answer pairs that exist only in the JSON-LD are the ones worth deleting. Everything else can stay.
HowTo went first. Google's documentation changelog records that on 14 September 2023 it removed the How-to structured data documentation, because the rich result was no longer shown in search results on desktop or mobile. FAQ then followed the same path three years later. Two features, same ending.
Look at the pattern rather than the individual features. Both were markup types that gave sites extra visual space without giving searchers much extra value. Both got narrowed, then retired. Google has been steadily simplifying its results page, and features that mostly served publishers were the first to go.
That gives you a reasonable filter for the future. If a rich result exists mainly to make your listing bigger rather than to answer a question better, treat it as temporary. Build your content so it still works when the feature disappears.
The ones that describe things Google needs to understand, not things Google decorates. In practice that means Organization, WebSite, BreadcrumbList, LocalBusiness, Product, and Article. These describe your business, your site structure, your navigation, and your content, and none of them depend on a single fragile search feature.
The HTTP Archive's 2024 Web Almanac gives a sense of what the web actually uses. In its structured data chapter, WebSite appeared on 12.73 percent of mobile pages, Organization on 7.16 percent, BreadcrumbList on 5.66 percent, LocalBusiness on 3.97 percent, and ItemList on 2.44 percent.
Against that, the same Web Almanac chapter put FAQPage adoption on desktop at 0.6 percent in 2024, up from 0.2 percent in 2022. So FAQ markup was never as widespread as SEO advice made it sound, even at its peak.
Our own priority order on client builds is Organization and WebSite first, because they help Google connect a site to a real business entity, then BreadcrumbList, then whatever content type fits the page. That set has survived every results page change we have watched.
Nobody has proven they do, and we will not pretend otherwise. ChatGPT, Perplexity, Claude, and Google's AI Overviews do not publish ranking or citation criteria, so any confident claim that FAQPage markup gets you cited is guesswork. What we can say is that clean, well-structured content is easier for any machine to parse.
Here is how we reason about it instead. Language models read the rendered page. A question written as a heading, followed by a direct answer in the first sentence, is legible whether or not there is JSON-LD around it. The structure of the visible page is doing the work.
So we optimise the page, not the markup, when the goal is citation. That is the approach behind our guide to getting your content cited by AI search engines, and it does not depend on any single schema type staying supported.
If you keep FAQPage markup because you think it might help with AI answers, that is a defensible bet at a very low cost. Just do not build a content plan on it.
Keep it, and make it better. The FAQ rich result is gone, but the reason FAQ sections exist has not changed. People arrive with objections, and answering them on the page shortens the sales conversation. That value was never coming from the markup.
What we would change is the quality bar. A lot of FAQ sections were written to feed the schema, so they are full of thin, keyword-shaped questions nobody asks. Now that there is no rich result to win, there is no reason to keep those.
Write the questions your sales conversations actually contain. Price, timeline, what happens if it goes wrong, who owns the work at the end. Our notes on designing an FAQ section that answers real questions cover how to structure that on the page.
In our experience the FAQ block is one of the highest value sections on a services page, and it got worse for years because teams were writing it for a Google feature instead of for a buyer.
Three things, in order. First, stop adding FAQPage markup for rich result reasons, because there is no rich result. Second, leave existing valid markup alone unless it describes content a visitor cannot see, in which case remove it. Third, move that effort to Organization and BreadcrumbList markup.
Then go and read your own FAQ section as a buyer rather than as a marketer. If the questions were reverse engineered from a keyword tool, rewrite them. That single edit usually does more for conversions than any structured data ever did for clicks.
The wider lesson we take from this is to be careful about building on borrowed features. A rich result belongs to Google. The clarity of your page belongs to you, and it keeps working when the feature list changes.
If you want a second opinion on what structured data your site should actually be running, or you want your FAQ content rewritten so it earns its place, we are happy to walk through it. Reach out at phoenix.studio.
Tell us where you want to go. We'll tell you how we'd get you there.