What is schema markup?
Schema markup is standardised code — usually JSON-LD — that labels the content on a page so machines can identify it without guessing. It tells a search engine that one string is a price, another is an author, and another is a review score. Google uses some of these labels to generate rich results: the star ratings, prices and images that make a listing occupy more space in the SERP.
The critical thing to understand in 2026 is that the list of schema types that earn a visible rich result has been shrinking for three years, and two of the types most heavily promoted to marketers are now on it.
If you are adding FAQ or HowTo schema to earn a SERP feature, stop. Neither has produced one since May 2026 and September 2023 respectively. Here is the current state of play, and what to do instead.
What Google removed, and when
| Type | Status | Date |
|---|---|---|
| HowTo | Rich result removed from mobile, then desktop | Aug–Sept 2023 |
| FAQPage | Restricted to authoritative government and health sites | Aug 2023 |
| FAQPage | Rich result removed entirely, including those last eligible sites | 7 May 2026 |
| Book Actions, Course Info, ClaimReview, Estimated Salary, Learning Video, Special Announcement, Vehicle Listing | Rich results retired | June 2025 |
| Practice Problem | Removed from Search Console and Rich Results Test | Jan 2026 |
The FAQ change arrived as a banner on Google's FAQPage documentation rather than a blog post, which is why a lot of teams missed it. The rollout runs in three stages: the rich result stopped appearing on 7 May 2026, Search Console reporting and Rich Results Test support ended in June 2026, and Search Console API support ends in August 2026.
That last date is the operational one. If you pull FAQ rich result data into a dashboard or a BigQuery export, those calls start returning null. Check your reporting stack before the field goes quiet, because a dashboard that silently reports zero looks identical to a dashboard reporting a collapse.
Does this mean schema markup is dead?
No, and the two overcorrections doing the rounds are both wrong.
"Schema is dead" is wrong because roughly 25 types still produce rich results, several of which materially affect click-through.
"FAQ schema matters more than ever for AI" is also wrong, or at least unproven. Google's own guidance on AI features states that no special markup is required for AI Overviews or AI Mode, and that structured data should match the visible text on the page. Nobody has demonstrated that AI engines rely on the markup rather than the rendered text.
The accurate position is narrower and more useful: a rich result and a schema type are different things. Google retired the display feature. The FAQPage type is still valid schema.org, it still validates, it triggers no penalty, and Google has confirmed unused markup causes no problems — it simply has no visible effect.
Which schema types still earn rich results?
These are the ones worth your implementation time.
| Type | Rich result produced | Best for |
|---|---|---|
| Product | Price, availability, star rating, image | E-commerce. The highest-CTR-impact type on this list. |
| Review / AggregateRating | Star ratings | Products and services with genuine review data |
| Article / NewsArticle | Headline, image, publish date, byline | Editorial and blog content |
| LocalBusiness | Address, hours, phone, map presence | Any business with a physical location |
| BreadcrumbList | Breadcrumb trail replacing the raw URL | Every site with more than two levels of hierarchy |
| Organization | Knowledge panel inputs, logo, social profiles | Every site — this is your entity anchor |
| Event | Date, venue, ticket link | Anything with a date and a location |
| Recipe | Image, cook time, rating, calories | Food publishers |
| JobPosting | Title, employer, location, salary range | Careers pages |
| VideoObject | Thumbnail, duration, key moments | Any page with an embedded video |
If you implement four of these correctly — Organization, BreadcrumbList, Article and whichever of Product or LocalBusiness applies — you have captured most of the available value. Adding fourteen types does not compound.
Should you remove your existing FAQ schema?
Usually not. The decision is smaller than the headlines suggest.
Keep it where the questions and answers are genuine, visible on the page, and useful to a reader. It costs nothing to leave in place, it still labels a question-answer pair unambiguously for any machine reading the page, and removal carries engineering cost for no gain.
Remove or rewrite it where the FAQ block was written purely to capture the dropdown — generic questions nobody asked, answers padded with keywords, or markup describing content that isn't visible on the page. That last case was always a violation and is now pointless as well.
Stop recommending it for new builds as a visibility tactic. Add FAQ sections because your buyers ask those questions.
One distinction people get wrong constantly: FAQPage is for questions where the site owner supplies the single official answer. QAPage is for a question with multiple user-submitted answers — forums and community Q&A. QAPage remains supported. Using the wrong one gets the markup ignored.
Why the deprecations happened
Google has not published a reason for the 2026 removal, but the pattern is consistent with its 2023 explanation for restricting FAQ and HowTo: a cleaner, more consistent results page.
The mechanism was abuse. Two sites with FAQ schema on the same SERP could each occupy several hundred vertical pixels, pushing genuine results below the fold, and the questions were frequently scraped, duplicated or invented. Google's response was to remove the display feature rather than police individual implementations at scale.
The lesson generalises, and it's worth internalising before the next feature disappears: any schema type whose primary value is decorative is a candidate for removal. Types that describe genuine entities — a product with a real price, a business with a real address, an article with a real author — have survived every round of pruning. Types that only expanded your listing have not.
How to test your schema markup
Three tools, three different jobs:
- Google's Rich Results Test — checks whether your markup is eligible for a *Google rich result specifically*. Since June 2026 it no longer tests FAQ. This is the tool that answers "will this show up in Google?"
- Schema.org validator (validator.schema.org) — checks whether your markup is valid schema.org, regardless of what Google displays. This is the tool that answers "is this correct?" FAQPage still passes here.
- Search Console's rich result reports — shows what Google is actually finding on your live site at scale, which is different from what one test URL returns.
Use all three. A common failure is markup that validates perfectly in isolation but never appears in Search Console because it's injected by JavaScript that Googlebot doesn't execute in time.
Common schema markup mistakes
Marking up content that isn't visible on the page. Google's guidance is explicit that structured data should match the visible text. Invisible markup is a policy violation, not a shortcut.
Adding every type you can. Service, Course and Event blocks in a site footer with arbitrary properties do nothing except create maintenance debt.
Injecting JSON-LD via a tag manager and never verifying. It often works. It also often doesn't, and nobody checks until a quarterly report shows zero enhancements.
Inconsistent Organization data. Your Organization block is the anchor that ties your site to your entity across Google's Knowledge Graph and, increasingly, how AI engines resolve who you are. If the name, logo URL and social profiles differ from what appears on your LinkedIn and Crunchbase entries, you are weakening the strongest signal you control.
Chasing types before fixing the page. Schema describes content. It does not create it. A thin page with perfect markup is still a thin page.
Frequently asked questions
Is FAQ schema dead? The rich result is. The schema type is not. FAQPage remains valid schema.org markup, it causes no errors and no penalty, but as of 7 May 2026 it produces no visible feature in Google Search.
Should I remove FAQPage schema from my site? Only where the questions were written for the SERP rather than for readers. Genuine, visible FAQ content with matching markup can stay — it costs nothing and still labels the Q&A pairing for any machine reading the page.
Does schema markup help with AI Overviews? Google states no special markup is required for AI Overviews or AI Mode, and that structured data should match visible text. Clean markup on real entities is sound practice; treating it as an AI citation shortcut is not supported.
What is the difference between FAQPage and QAPage? FAQPage is for questions the site owner answers officially. QAPage is for a question with multiple user-submitted answers, such as a forum thread. QAPage is still supported by Google.
Does schema markup improve rankings? No. Rich results change how a listing looks at its existing position; they do not move the position. Google has confirmed rich results are not a ranking input.
Which schema types should a service business implement? Organization, BreadcrumbList, Article on guides and blog posts, and LocalBusiness if you have a physical location. Four types, implemented correctly, beat fourteen implemented loosely.
What happens to my FAQ data in Search Console? Reporting and Rich Results Test support ended in June 2026. Search Console API support ends in August 2026, after which those calls return null. Export anything you need before then.
Get your schema markup implemented properly
TapasSEO's Structured Data Implementation delivers validated JSON-LD across your key templates — Organization, BreadcrumbList, Article, and Product or LocalBusiness as applicable — tested in all three validators, with the deprecated types removed rather than left in your codebase to accumulate.
Structured Data Implementation — $697 →
If the goal is having AI engines correctly identify who your business is rather than decorating a SERP listing, Entity & Schema Optimization aligns your Organization markup with your external entity footprint.