Schema types: which ones Google still shows

Schema.org defines hundreds of schema types, but Google turns only about two dozen of them into rich results, and that list keeps shrinking. As of its June 15, 2026 update, Google's Search Gallery lists 25 features, led by Article, Product, Review snippet, Recipe, Event, Video, Local business and Organization. FAQ, HowTo, Fact check, Course info, Estimated salary and a handful of others no longer show at all. Mark up what still pays, keep the types that describe your business, and stop adding the rest for SEO reasons.

Which schema types still get rich results

These schema types map to a feature in Google's Search Gallery today. The schema.org type is what you write in @type. The gallery names the feature, not the type, so the second column comes from each feature's own documentation.

The other gallery entries, such as Carousel, Speakable and Image metadata, add to these rather than standing alone. Book actions survived too, but only for onboarded partners who send Google a feed, per its documentation.

For most sites, the first eight rows are the whole list. Walkthroughs live in their own posts for Article schema, product schema, review schema, breadcrumb schema and LocalBusiness schema.

Schema types Google retired or limited

Google has removed ten rich results since 2023. Every date below comes from the Search Central changelog or the post it links.

The June 2025 announcement listed Book actions too, but Google later took the deprecation banner off that page because a feature still uses the markup. Note the overlaps. Course and Quiz each lost one feature and kept another, so check which one your markup was built for before you delete it.

Google said the June 2025 cuts "won't affect how pages are ranked". Retired markup is harmless dead weight. I leave it in place on existing pages and stop adding it to new templates. FAQ is the case people argue about most, and the FAQ schema post covers whether to keep it.

Do schema types matter for AI search?

Not as a requirement. Google's AI optimization guide says structured data isn't required for generative AI search and that no special markup exists for it. No AI company has published a rule that a given schema type gets you cited, and the schema markup guide covers what Microsoft has said on the subject.

So pick types for what they state, not for a display. A few are worth keeping on every site whatever happens to rich results:

• Organization on the home page. Name, logo, url and sameAs links tie your site to your company's profiles elsewhere. The Organization schema post lists the properties.
• Article with author and dates. It states who wrote a page and when it changed, which a reader otherwise has to infer from the layout.
• Product with Offer. Price, currency and availability in a fixed format, still a live Google rich result.
• LocalBusiness with address and hours. Facts people ask assistants for, and still a Google feature.

Whatever you add, put it in the HTML your server sends. Most AI crawlers do not run JavaScript, so markup injected by a tag manager may never reach them.

How to choose schema types for a page

Start from what the page visibly is, then check the gallery. Google's structured data policies say markup must describe content a reader can see on that page.

1. Name the page's main thing in one word, such as product, article or location.
2. Use the most specific schema type for it, such as Dentist instead of LocalBusiness.
3. Add BreadcrumbList on inner pages and Organization plus WebSite on the home page.
4. Skip any type from the retired table unless it serves a purpose outside Google Search.
5. Validate before you ship, then confirm eligibility with Google's Rich Results Test.

Check which schema types your page uses

The Schema Markup Checker fetches a public URL and reads every JSON-LD block in the HTML your server sends, or checks JSON-LD you paste. It lists the types already present, flags missing required and recommended properties against Google's documented rules, and notes types whose rich result is retired or limited. For a URL it also suggests documented types that fit the page and are missing, based on signals like product details, step lists, video and a breadcrumb trail.

Each run costs 8 credits. It does not run JavaScript, notes Microdata and RDFa without validating them, and cannot tell you whether Google will actually show a rich result.