What Is Schema Markup? A Plain-English Guide to Structured Data

Search engines read your page as a wall of text and try to guess what everything means. Is "4.8" a rating, a price, or a version number? Is "Dr Sarah Chen" the author, a person the article is about, or a product name? Schema markup removes the guessing. It is a small block of code that labels the meaning of your content in a language machines read perfectly, and in 2026 the machines reading it are not only Google. They are the AI engines answering questions in ChatGPT, Claude, Perplexity, and Google's AI Overviews. This guide covers what schema markup is, what it actually does, and how to add it without tripping the rules that get it ignored.

What is schema markup?

Schema markup is structured data that labels the meaning of your content for machines. You add a block of code to the page, and it states in a fixed vocabulary what each thing is: this is an Article, its headline is X, its author is a Person named Y, it was published on date Z. "Structured data" and "schema markup" are the same thing in practice; the vocabulary that defines the labels is schema.org, a shared standard that Google, Microsoft, and the major AI engines all understand. Without it, a search engine infers meaning from words and layout and gets it wrong often enough to matter. With it, the meaning is declared, and the engine that reads your declared entities is the one that can quote you back to a searcher.

What does schema markup do?

Schema markup does two jobs: it makes your page eligible for rich results, and it hands machines a clean version of your content to understand and quote. The first job is the visible one. Rich results are the enhanced listings in Google, the star ratings, the FAQ drop-downs, the product prices and stock status, the recipe cook times, the event dates. None of those appear unless the page carries the matching schema. The second job is the one that matters more every month. When an AI engine answers a question, it favours sources whose meaning is declared and unambiguous, because a labelled entity is safer to quote than a guess pulled from prose. Schema is how you make your content easy for a machine to trust, and trust is what decides which page an engine turns into an answer.

What is JSON-LD, and why is it the format to use?

JSON-LD is the schema format to use, because it sits in one block separate from your visible HTML. There are three ways to write structured data: JSON-LD, Microdata, and RDFa. Microdata and RDFa weave the labels through your page's HTML tags, which makes them fiddly to add and easy to break when the design changes. JSON-LD keeps the whole thing in a single script block, usually in the page head, decoupled from the layout. Google states a preference for JSON-LD, every major SEO team uses it, and it is the format our tools produce and check. If you inherit a page using Microdata, it still works; when you build something new, write it in JSON-LD and keep your structured data in one place you can maintain.

Which schema types matter most?

The types that matter are the ones that earn a visible result for your kind of page. Schema.org defines hundreds of types, but a handful cover almost every business. Match your page to the type that describes it, and you unlock the result that goes with it. The table below maps the common types to what they earn and who should use them.

Breadcrumb and Organization belong on nearly every site. The rest depend on what a page actually contains, which is the rule that decides everything else about schema.

Does schema markup help SEO?

Schema markup does not raise your rankings directly, and treating it as a ranking trick is the fastest way to waste it. Google has said plainly that structured data is not a ranking factor. What it changes is how your result looks and how machines understand you, and those move the numbers that matter: a rich result with stars and a price earns a higher click-through rate than a plain blue link in the same position, and a page an AI engine can parse cleanly is a page it can cite. So the honest answer is that schema helps SEO the way a good headline helps SEO. It does not move you up the page on its own; it makes the position you have work harder, and in AI search it decides whether you are quotable at all. That value only lands if the markup is correct, which is why the next thing to get right is how you add it.

How do you add schema markup to a page?

You add schema markup as one JSON-LD block in the page head, then validate it before you trust it. The reliable loop has three moves, and you do not need to hand-write JSON to run it. First, generate the markup: describe the page or fill a form and get valid JSON-LD out, with our free Schema Markup Generator or, for a whole page at once, the Schema Markup Generator skill inside Claude. Second, validate it: paste the code or a URL into the Schema Markup Validator and it flags missing required properties, weak recommended ones, and structural mistakes before Google ever sees them. Third, audit the whole page: hand the page to the Schema Markup Auditor skill and it grades every type, finds the schema you are missing, and rewrites it. Paste the final block before the closing head tag, confirm it in Google's Rich Results Test, and the loop is done.

What gets schema markup wrong?

The rule that gets schema wrong is marking up anything the page does not actually show. Structured data must match the visible content, and Google issues manual actions for markup that claims more than the page delivers. So the failures are consistent: adding AggregateRating with review counts the page has no reviews to support, adding FAQPage schema when there is no FAQ on the page, inventing a rating or a price to look better in the listing. Keep every value real and sourced from the page, mark up only what a visitor can see, and consolidate your types into one @graph rather than scattering competing blocks. Schema is a description, not a claim; describe the page you have, and the results follow.