Structured data: what it is, which schema.org types Google uses and how to validate JSON-LD

Structured data is a block of code, usually JSON-LD, that describes a page's content using the schema.org vocabulary: which company you are, which product you sell, which navigation path the page sits on. It makes a page eligible for rich results in Google, but does not guarantee them, and it has to match what is visible on the page.

To check it on your own site: SEO audit

It is a small part of SEO, but one of the parts that breaks most quietly: a price that changes on the page but not in the markup, a template that stops printing it, a type Google no longer shows. This guide sums up what Google documents as of October 2026, with a validated example.

What is structured data?

Structured data is a standardised format for giving Google explicit information about a page and classifying its content, such as the ingredients and cooking time of a recipe1. Most of it uses the schema.org vocabulary, a joint project started by Google, Microsoft, Yahoo and Yandex and now maintained by a W3C community group2.

Google supports three formats, and recommends the first1:

JSON-LD has another advantage: Google processes it even when it is generated with JavaScript, as long as it is in the DOM when Google renders the page3. But that advantage only applies to Google, as you will see below.

Which structured data types does Google use?

Google only uses the types in its gallery for rich results, not the whole of schema.org. Schema.org defines hundreds of types; Google's gallery, updated on 15 June 2026, lists only the ones its search supports4. These are the ones you will see most often on a company website, a shop or a blog:

Each type has required and recommended properties. Without the required ones, the item is not eligible8. The most common types have their own guides: Product schema for product pages, LocalBusiness and local SEO and author markup in E-E-A-T author signals.

Which rich results no longer exist?

Google retires features fairly often, and the old markup stays behind in templates. According to its documentation changelog9:

Google no longer showing a result does not turn the markup into an error: FAQPage is still a valid schema.org type. What changes is the priority. If keeping a FAQPage or HowTo block up to date costs you work on every template change, there is no longer a Google rich result to justify it.

A validated JSON-LD example

Two blocks almost any site can use. The first goes on the home page or the About page: it identifies the company, with a fixed @id so other blocks can refer to it.

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "@id": "https://www.example.com/#organisation",
  "name": "Example Kitchens",
  "url": "https://www.example.com/",
  "logo": "https://www.example.com/img/logo-512.png",
  "email": "hello@example.com",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "1 High Street",
    "addressLocality": "Manchester",
    "postalCode": "M1 1AA",
    "addressCountry": "GB"
  },
  "sameAs": ["https://www.linkedin.com/company/example-kitchens/"]
}

The second goes on an article and describes its navigation path. Google asks for at least two items, and the last one can leave out item: Google then uses the URL of the page itself6.

{
  "@context": "https://schema.org",
  "@type": "BreadcrumbList",
  "itemListElement": [
    { "@type": "ListItem", "position": 1, "name": "Blog", "item": "https://www.example.com/blog/" },
    { "@type": "ListItem", "position": 2, "name": "Kitchens", "item": "https://www.example.com/blog/kitchens/" },
    { "@type": "ListItem", "position": 3, "name": "How much does a new kitchen cost" }
  ]
}

We ran both blocks through the schema.org validator on 9 October 2026: zero errors and zero warnings. Three details that often go wrong: the Organization logo must be at least 112 × 112 pixels and crawlable5; the breadcrumb trail should reflect how a person navigates, not copy the URL structure6; and every piece of data in the markup must also be on the visible page.

How do you validate structured data?

There are two validators, and each answers a different question:

1. Google's Rich Results Test tells you whether the page is eligible for the rich results Google supports. If the markup is generated with JavaScript, use the URL option rather than the code option, because the code option has JavaScript limitations3.
2. The schema.org validator (validator.schema.org) checks that your JSON-LD, Microdata or RDFa is well formed according to schema.org. It does not check against the requirements of any particular service, Google included11.
3. URL Inspection in Search Console confirms that Google has found the markup on the live page1.
4. Rich result reports in Search Console count how many items are valid and how many are not, and separate critical issues, which block the rich result, from those that only affect how it appears12.

Passing the tests is not enough. Google is clear: markup enables a feature but does not guarantee it will appear, and its algorithm may show a plain text result depending on the search, location or device8.

Common structured data mistakes

These are the mistakes described in Google's general guidelines and its review documentation87:

The most serious consequence is a manual action for structured data. Google explains that it removes eligibility for rich results but does not affect how the page ranks in web search; you will find it in the Manual Actions report in Search Console8. Shops have one more nuance: Product markup generated with JavaScript can make Google Shopping crawls less frequent and less reliable, which matters when price and stock change often3.

Does structured data help you appear in AI answers?

For Google's AI, it is not a requirement. Google says so in its guide to generative AI features: structured data is not required for its generative AI search and there is no special schema.org markup to add. It does recommend continuing to use it for rich results in classic search13.

Outside Google, we have not found documentation from OpenAI, Anthropic or Perplexity explaining how they use schema.org markup. What is clear is the order of priorities: an assistant cites the text it can read, so the data that matters (price, opening hours, author, answer) has to be written on the page, not only in the JSON-LD. And if you generate the markup with JavaScript, remember that several AI crawlers do not run it, as the guide to JavaScript SEO and AI crawlers explains. What makes an assistant cite you is covered in AI search optimisation.

What SmoothSeen checks

Declaration of interest: this blog belongs to SmoothSeen, a web audit tool. In the SEO half of the analysis it checks whether the page has schema.org markup (JSON-LD or Microdata) and whether that markup contains errors. In the AI visibility half it looks at whether the page type carries the markup that suits it: authorship on an article, a breadcrumb trail on a category, business details on a local business. It does not replace Google's Rich Results Test, which is what tells you whether you are eligible.

What to do this week

Run your home page, one article and one product or service page through Google's Rich Results Test and note which types it finds and with what errors. If HowTo or FAQPage turns up, decide whether it is still worth maintaining. To see the markup reviewed alongside the rest of the page's technical analysis, run a free SEO audit.

Frequently asked questions

Does structured data improve rankings?

Google does not present it as a ranking factor, but as a way to make a page eligible for rich results, which can attract more clicks. Even a manual action for incorrect markup removes that eligibility without affecting the page's position in web search. Its value lies in how the result is displayed, not in where it ranks.

Which format is better, JSON-LD or Microdata?

Google supports all three formats, JSON-LD, Microdata and RDFa, as long as the markup is valid, but it recommends JSON-LD. It sits in a block separate from the visible text, so it can be changed without touching the design and is easy to generate from a template or CMS. Microdata only makes sense if your templates already use it and it works.

Should I remove FAQPage markup from my site?

There is no need. Google stopped showing FAQ rich results on 7 May 2026, but FAQPage is still a valid schema.org type and does not cause errors. The question is whether it is worth maintaining. What is worth keeping is the visible FAQ section itself, if it answers real questions from the people reading the page.

Why doesn't Google show my rich results when the test passes?

Because the test only tells you the page is eligible. Google decides on every search whether to show the rich result, depending on the query, location and device, and it may prefer a plain text result. It may also withhold it if the markup breaks its quality guidelines, for example with reviews that are not visible on the page.