Search

Schema markup

structured data · JSON-LD

Definition

Schema markup is structured data on a page — usually JSON-LD — that states claims in a vocabulary machines already know: Organization, Article, FAQPage, Product. It does not rank you by itself. It makes eligible rich results possible and helps entity disambiguation when the visible text already says the same thing.

Google’s docs: markup should match visible content. Fake review stars and invisible FAQs get ignored or punished.

How it works

Mark the organization (legal name, logo, sameAs profiles), the article (headline, date, author), and any FAQ that is actually on the page. Test with Rich Results and the URL Inspection tool. Keep sameAs URLs consistent with Wikidata and Wikipedia if they exist.

Worked example

A newsroom article about a study adds NewsArticle JSON-LD with the journalist’s byline, the datePublished, and the canonical URL. Search shows a headline and date. A fake FAQ of keywords is omitted on purpose.

How it differs

A knowledge graph is the store of entities. Schema is one input. llms.txt is an unofficial hint file, not schema. E-E-A-T is rater language, not a tag you add.

Common errors

Marking up reviews that are not on the page. Stale dates. Three conflicting legal names. Copy-pasting another brand’s Organization block.

Sources