Skip to content

Schema markup

Schema markup is structured data describing a page to search engines. Which formats Google recommends, which types earn rich results, and which do not.

Definition

Schema markup is structured data added to a page that describes its content in a vocabulary search engines understand, defined at schema.org. Google recommends JSON-LD as the format, and supports a specific list of types that can produce rich results.

Illustration representing schema markup

Why it matters

Schema markup is how a page states what it is rather than leaving it to be inferred. For some content types it changes how the result looks: star ratings, prices, event dates and breadcrumbs all come from structured data.

It is also where a lot of effort gets wasted, because the list of types that actually change a search result is much shorter than the list of types that exist.

How it works

The vocabulary is defined at schema.org, a collaborative standard. The format Google recommends is JSON-LD, a block of JSON in a script tag, which keeps the markup separate from the visible HTML:

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "BreadcrumbList",
  "itemListElement": [
    { "@type": "ListItem", "position": 1, "name": "Home", "item": "https://example.com/" },
    { "@type": "ListItem", "position": 2, "name": "Glossary", "item": "https://example.com/glossary" }
  ]
}
</script>

The crucial distinction: schema.org defines hundreds of types, and Google supports a specific subset for rich results. Google's search gallery lists the ones that produce a visible feature, around 29 of them, including Breadcrumb, Product, Review snippet, Event, FAQ-era types that have since been retired, Job posting and Video.

Types outside that list are still valid and still worth adding where they describe the page accurately. They just will not change how the result looks. DefinedTerm, which every page in this glossary carries, is one of those: it states that the page defines a term, which helps machine understanding, and it earns no rich result.

What has been retired

Support changes. The FAQ rich result was removed from Google Search in May 2026, and the documentation for it withdrawn. Markup added for a feature that no longer exists is maintenance cost with no return, which is why the supported list is worth re-checking rather than trusting a guide written two years ago.

Worked example

The three that earn their place on most business sites:

  • Organization. once, sitewide: name, logo, contact, profiles
  • BreadcrumbList. on every page with a hierarchy
  • Product. on ecommerce pages: price, availability, reviews
An open printer's type case of blank sorts, with exactly three lifted into a composing stick and their compartments empty
The gallery is full of types. The honest filter is to set the ones that are accurate, not the ones that exist.

Everything else depends on what the site publishes. The honest filter is to check the gallery first and add a type because it is accurate, not because it is available.

Common mistakes

  • Marking up content that is not on the page.
    Structured data has to match what a visitor sees. A rating in the markup and no rating on the page is a spam signal.
  • Adding types for features that do not exist.
    FAQPage is the current example.
  • Expecting ranking benefit.
    Rich results affect appearance and click-through, not position.
  • Letting markup drift from the page.
    A price in JSON-LD that no longer matches the page is worse than no markup.
  • Forgetting to validate.
    The Rich Results Test and the enhancement reports in Search Console both catch errors before they matter.

Does it help AI search

Indirectly, and less than it is sold as. Google's guidance on AI features says there is no markup that makes a page eligible for inclusion. Accurate structured data helps a machine understand what a page is about, which is worth having, and it is not a route into generated answers. The content and its structure do that work, as our technical SEO checks assume.

Let's build

Got something to build?

Tell us what you're making. We'll tell you - honestly - if we're the right team, and how we'd do it.