SEO & Technology

Structured data & schema markup: what helps for Google and AI?

Structured data is a standardised description of what is on a page, in a form that machines can read without guesswork. With schema markup, usually written in JSON-LD, you tell search engines and AI systems who the publisher is, what a page describes and how the parts of your website fit together.

At a glance

What is structured data? The short answer

A visitor immediately sees what a title, a price or an author is on a page. A machine has to infer that from the markup. Structured data removes that guesswork: you state explicitly “this is a service”, “this is the provider” or “this is the author of this article”.

What helps is markup that matches what is visible on the page and that connects your entities: the organisation or person, the services, the articles and the place of a page within the website. What does not help is markup for the sake of markup: types that do not fit the content, or data that a visitor cannot find anywhere.

Structured data is not a button for higher rankings. It makes your content easier to understand, and some types make you eligible for a richer display in Google. Whether that happens is up to Google.

Structured data – example of schema markup in JSON-LD

The terms

Structured data, schema markup and JSON-LD: what is the difference?

The three terms are used interchangeably, but they do not mean the same thing. Structured data, as the name suggests, is the principle: information in fixed fields rather than in running text.

Schema markup is structured data based on the schema.org vocabulary. That vocabulary defines which types exist, such as Person, Organization, Service or Article, and which properties belong to them. The search engines set up that vocabulary together, so you do not need a different notation for each search engine.

JSON-LD is the notation: a small block of code in the HTML of the page, separate from the visible layout. Other notations exist too, such as microdata, but Google, in its introduction to structured data , recommends JSON-LD because it is the easiest to maintain.

Example

What does schema markup look like?

This example describes an organisation and one service it offers. The service uses an @id to refer to the organisation, instead of repeating its details.

JSON-LD in the HTML of the page
<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@type": "Organization",
      "@id": "https://www.example.com/#organisatie",
      "name": "Voorbeeld BV",
      "url": "https://www.example.com/"
    },
    {
      "@type": "Service",
      "@id": "https://www.example.com/dakwerken/#dienst",
      "name": "Dakwerken",
      "provider": { "@id": "https://www.example.com/#organisatie" }
    }
  ]
}
</script>

Two things stand out. Each part has its own fixed address: the @id. And the service does not repeat the organisation’s name, but points to it. This creates one coherent whole, a graph, rather than separate blocks that can contradict each other.

The name in the markup must be the same as the name a visitor reads on the page. That applies to everything you include.

Types

Which schema types does a business website need?

Schema.org has hundreds of types. An ordinary business website only needs a handful. These seven cover most situations.

Schema types for a business website
TypeDescribesUse it whenWatch out
PersonA person: name, job title, expertiseThe brand is one person, or authors play a roleOne Person per real person, with a fixed @id
OrganizationA business or organisationThere is a business behind the websiteName, logo and contact details as they appear elsewhere
ArticleAn article or blog postYou publish knowledge articles or newsAuthor and date must also be visible on the page
ServiceA service and who provides itA page describes one serviceDo not add a price or rating that is not on the page
ProductA product with price and stockYou sell a product, for example in an online shopDo not use it for a service to get stars in Google
BreadcrumbListThe place of a page within the websiteThe website has more than one levelMust match the visible breadcrumb trail and the URL structure
WebSite and WebPageThe website itself and the pageAlways: they form the framework for the restRefer from here to the organisation or person

Person and Organization are the foundation: they say who the publisher is. All other types refer to them. You can read why that matters so much in Entity SEO: how Google and AI recognise your brand.

Effect

What does structured data do for Google and for AI systems?

Structured data has three possible effects. They are not equally certain, and it is only fair to make that distinction.

01

Understand

Search engines use structured data to understand what a page is about and how pages relate to each other. That effect is always there, even when you see nothing of it in the search results.

02

Rich results

For a number of types, Google can show a richer search result, with a breadcrumb trail or product details, for example. Valid markup makes you eligible, but does not guarantee it. Google also regularly adjusts which types it still shows.

03

AI answers

Google’s AI features rely on the same index and the same signals as the regular search results. How other AI systems handle structured data varies by system and is not always documented. So do not count on a direct effect, but do count on a website that states unambiguously who you are.

Cohesion

One graph instead of separate blocks

Your website gains most not from more types, but from markup that comes from one source and says the same thing everywhere. On my own website, for example, the services and the articles all refer to the same Person.

  • Fixed addresses Give each entity one @id and reuse it on every page.
  • Refer, do not repeat A service or article points to the organisation or person instead of copying the details.
  • One source Have the markup generated by one system. Three plugins that each output their own Organization contradict each other.
  • Matching the page What is in the markup is also visible on the page, with the same words and figures.
  • In the HTML Put the markup in the HTML that the server sends, so that a crawler can read it without executing JavaScript.
  • Kept up to date If a name, an address or a service changes, the markup changes with it.

Structured data is therefore part of technical SEO, and a building block of GEO & AI SEO: GEO (Generative Engine Optimization) starts with a website that machines can read correctly.

Avoid

Which markup is best avoided?

  • Made-up reviews and ratings. Stars without real, visible reviews are misleading and go against Google’s guidelines.
  • Markup for invisible content. Data that a visitor cannot find anywhere on the page does not belong in the structured data.
  • The wrong type. Marking up a service as a Product, or an ordinary text page as an Article, to get a more eye-catching search result.
  • FAQ markup on everything. Adding questions and answers to pages where they are only there for the markup.
  • Duplicate entities. Two organisations or two people with the same name and different details on one page.
  • An incorrect address. Listing a location where customers cannot find you.

The rule of thumb is simple: only mark up what is true and what a visitor can also see. Misleading markup can lead Google to stop showing rich results for your website.

Oversight

How do you validate structured data?

Markup that is syntactically correct can still be wrong in substance. So check both, and repeat that after every major change to the website.

01

Test the syntax

The Schema Markup Validator from schema.org shows whether the code is valid and which types and properties it contains.

02

Test eligibility

Google’s Rich Results Test shows which displays a page is eligible for, and which fields are missing.

03

Compare with the page

Read the markup alongside the visible content. Are the name, date, author and price correct?

04

Check the rendered HTML

Check that the markup is in the source code the server sends, and does not only appear after scripts have loaded.

05

Monitor Search Console

Google reports errors and warnings there per type, across the whole website.

06

Repeat after changes

A new plugin, a new theme or a new release can duplicate markup or make it disappear.

Implementation

Structured data in WordPress and in Next.js

In WordPress, an SEO plugin usually creates the basics: the website, the page, the breadcrumb trail and the article. What a plugin does not know is how your services, your organisation and your authors are connected. You define that connection in the settings or with a bit of custom code, so that everything points to the same entity. Watch out for themes and page builders that add their own markup without being asked.

In an application built with React and Next.js, you write the markup yourself, per page type. The advantage is full control. The point to watch is rendering: the JSON-LD must be in the HTML that the server sends, not added later in the browser.

In both cases, the same order applies: first determine which entities the website has, and only then write the markup. Working the other way round leaves you with a lot of code and little coherence.

FAQ

Frequently asked questions about structured data and schema markup

Schema markup is code in your page that uses the schema.org vocabulary to describe what is on the page: a person, an organisation, a service, an article or a product. Search engines read that code to interpret the content correctly.

JSON is a general notation for exchanging data. JSON-LD is JSON with fixed conventions for recording meaning: it does not just say “name: Example BV”, but also that this is the name of an organisation according to schema.org.

No. A page can be indexed and perform well without structured data. JSON-LD is, however, the notation Google recommends when you add structured data, and valid markup is a prerequisite for rich results.

In WordPress, an SEO plugin does most of the work, supplemented with settings or custom code for your services and entities. In a bespoke build, you write the JSON-LD into the HTML for each page type. Then test each type of page with a validator.

Not directly. Structured data helps Google understand your content and can result in a richer display. That display can bring more clicks, but the ranking itself depends on the content, the technology and the authority of your website.

It is not a guarantee and not a direct lever. It does help to state unambiguously who you are and what you offer, and that is the foundation search engines and AI systems rely on. Without clear, visible content, markup changes nothing.

More on this topic

Further reading

The related service

Technical SEO as a service

Structured data is one part of the technical foundation, alongside crawlability, indexing, rendering and performance. On the service page, you can read how I investigate these and, where needed, implement the work myself.

Dries Nijs, Organic Web Specialist and author of this article

About the author

Dries Nijs, Organic Web Specialist

I am Dries Nijs, a developer and SEO consultant from Maasmechelen. I have been building for the web since the 1990s and write structured data myself, in WordPress and in React and Next.js. MadeByDries is the personal brand under which I work.

Not sure whether your markup is correct?

Tell me which platform your website runs on. Then we will look at what is in it today and what is missing.

Discuss your project

Tell me what you want to build or improve

You do not need a technical brief yet. Tell me where you are now and what you want to achieve, and I will help you find the right approach.

Choose at least one topic. You can select more than one.

I only use your details to respond to your enquiry.

Prefer to email me directly?

Location
Maasmechelen, Belgium