Structured data is any data on a page that's marked up in a machine-readable way to explicitly describe what the content is — not just render it for humans. It's the layer that lets a search engine know "this $129.00 is a price, that 4.6 is a star rating, this string is the article author."
Format vs vocabulary
Two pieces:
- Format: how the data is encoded. Three valid formats: JSON-LD (recommended), microdata (legacy, embedded in HTML attributes), and RDFa (rare).
- Vocabulary: which types and properties are recognised. The dominant vocabulary is schema.org, used by Google, Bing, Yandex, and most others. There are smaller vocabularies for specific domains (Dublin Core for libraries, FOAF for social, etc.) but for SEO you want schema.org.
JSON-LD with schema.org is the modern stack. Microdata still works but is more verbose and harder to maintain.
What structured data does for SEO
Three real benefits in 2026:
- Eligibility for rich results: see Schema markup for the current list of what still surfaces.
- AI citations: AI Overview, Perplexity, ChatGPT search preferentially cite pages with clear structured data because the schema gives them confidence about what the page is.
- Voice and assistant integration: Siri, Alexa, Google Assistant all consume structured data for direct answers.
What it doesn't do
- Doesn't directly rank you higher. Schema is an eligibility input, not a ranking factor.
- Doesn't make bad content rank. Structured data on thin content still produces thin SERP results.
- Doesn't override page content. Schema must match what's visible on the page or Google will ignore it (and may issue a manual action for systematic mismatch).
Validate before shipping
Two free validators:
- Google Rich Results Test — checks eligibility for Google rich results specifically
- Schema.org validator — checks general schema.org correctness
Use the Schema Generator to produce valid JSON-LD for the most common types, then paste the output and validate.