Schema Markup (Structured Data)
Definition
Schema markup is structured data vocabulary from Schema.org that you add to a page so search engines can identify what its content describes — a product, an event, a business, a recipe. Implemented in JSON-LD, Google's recommended format, it makes a page eligible for rich results such as star ratings, prices, breadcrumbs, and business details. Eligibility is not a guarantee: Google decides per query whether to show them. Schema is also the clearest way to state entity relationships, which is why Organization and Person markup underpin Knowledge Graph work. Validate with Google's Rich Results Test, which renders JavaScript that a raw source view would miss.
Which schema types still produce rich results?
The list shrinks over time. Product, Review, Recipe, Event, Breadcrumb, Video, and LocalBusiness remain well supported. Two prominent types are retired: HowTo rich results were removed in 2023, and FAQ rich results were withdrawn for all sites in June 2026. Markup for retired types is harmless to leave in place but earns nothing in search — check Google's structured data documentation before building for a specific result type.
Does schema markup improve rankings?
Not directly. Google has repeatedly said structured data is not a ranking factor; it affects how a result can be displayed, not where it places. The indirect benefit is real but narrower than commonly claimed — a richer result can earn more clicks at the same position. Treat schema as presentation and entity clarity, not as leverage on position.
JSON-LD, Microdata, or RDFa?
Use JSON-LD. Google explicitly recommends it, and it is the only format that lives in a self-contained script block rather than being interleaved with your markup — which means changing your layout cannot silently break your structured data. Microdata and RDFa are still parsed and are worth leaving alone if already in place, but no new implementation has a reason to choose them.
What makes structured data invalid?
The common failures are marking up content that is not visible on the page, omitting required properties for the type, and letting values drift from what the page says — a stale aggregateRating is the classic case. Google's guidelines treat markup that misrepresents page content as spam, and it is one of the few structured-data issues that can trigger a manual action rather than being quietly ignored.
Does schema help with AI search visibility?
Less than commonly claimed. A controlled test across roughly 1,900 pages found no measurable lift in AI citations from adding schema. The models retrieve from page text. Schema remains worth implementing for rich results and for entity disambiguation across Google's Knowledge Graph — just not as a lever for getting quoted by ChatGPT or Perplexity.
Related Tools
Run a free SEO audit to see how your site performs.
Try the SEO Checker