Skip to main content
All terms
AI Crawler Access & TechnicalUpdated

Structured data

Also known as: schema markup, JSON-LD, rich results markup

In short

Structured data is machine-readable markup — usually schema.org vocabulary in JSON-LD — that states explicitly what a page is about: an article, a product, a definition, an FAQ. Search engines use it for rich results and entity understanding rather than as a ranking factor.

What it is for

Structured data removes ambiguity. A page might be a product, a review of a product, or an article about a category of products, and the prose alone does not always make that obvious to a machine. Markup states it.

That has two payoffs: eligibility for rich results in search, and cleaner entity classification — which feeds the knowledge graph and helps systems attach facts to the right thing.

The honest position on AI citations

Structured data is routinely sold as an AI visibility tactic. The evidence does not support that framing: published analysis has found no measurable relationship between adding schema markup and being cited more often by AI engines.

The likely reason is mechanical. Engines retrieve and quote passages of text. Markup describes a page; it does not make a paragraph more quotable, and it is the paragraph that gets lifted.

Implement it, because it is cheap and it helps search and entity resolution. Do not expect it to move citations, and be sceptical of anyone who says it will — the same scepticism that applies to llms.txt.

Getting it right

  • Mark up what the page actually is. Misrepresentation is a policy violation with real penalties.
  • Keep it consistent with the visible content. Markup claiming a rating the page does not show is a common and penalised pattern.
  • Prefer JSON-LD, which is the format search engines recommend and the easiest to maintain separately from the page's markup.

Frequently asked questions

Does schema markup improve rankings?
Not directly. It can earn rich results that change how a listing looks and how often it is clicked, and it helps a system classify the page. Neither is a ranking boost, and treating markup as one leads to a lot of effort with no movement.
Does structured data help AI citations?
Published analysis has not found a measurable citation effect from adding schema markup. It remains worth implementing for search and entity clarity — just not as a lever for getting cited by AI engines.
Which schema types are worth implementing?
The ones that describe what your page genuinely is: Article, Product, FAQPage, DefinedTerm, BreadcrumbList, LocalBusiness. Marking a page up as something it is not is a policy violation, and adding types that do not apply achieves nothing.

Related terms

  • Knowledge graph

    A knowledge graph is a structured database of entities — people, companies, products, places — and the relationships between them. Search and AI systems use one to understand that a query is about a specific thing, and to retrieve facts about it rather than pages mentioning it.

  • Entity SEO

    Entity SEO is the practice of making search and AI systems recognise your brand as a distinct, well-defined thing rather than a string of characters. It covers consistent naming, unambiguous descriptions, and corroboration across independent sources so a system can confidently identify who you are.

  • Featured snippet

    A featured snippet is a passage Google extracts from a ranking page and displays at the top of results to answer a query directly, with attribution and a link. Unlike an AI Overview it quotes one source verbatim rather than synthesizing several.

  • llms.txt

    llms.txt is a proposed convention — a markdown file at your site root that summarizes your content for large language models. No major AI engine has been shown to read it, and publishing one has no measured effect on whether you get cited.

Related guides