Structured Data
Structured data is machine-readable information added to a web page, usually following the shared vocabulary at schema.org, that describes what the page is about (a product, a recipe, a review) in a fixed format so search engines and AI systems can read it without guessing.
The term is often used interchangeably with schema markup, but they are not the same thing. Structured data is the broad idea of describing a page in a defined, machine-readable shape; schema markup is the specific implementation using the schema.org vocabulary, which is by far the most common one on the open web. In practice, when people say structured data for SEO they almost always mean schema.org types expressed in a supported format. The point is to remove ambiguity: rather than letting a crawler infer that "£24.00" is a price and "4.7 out of 5" is a rating, you state both as named properties (price, ratingValue) that any compliant parser reads the same way.
There are three formats Google supports: JSON-LD, Microdata, and RDFa. JSON-LD is the recommended one because it sits in a script block separate from the visible HTML, so it is easier to add, read, and maintain. The same data can power a rich result in Google (star ratings, prices, FAQ accordions) and give AI assistants a clean source to quote, rather than parsing it out of page copy.
Consider a Shopify store selling merino base layers. The product template renders the title, price, availability, and a block of customer reviews. On its own, that page is readable to a shopper but loosely defined to a machine. Adding a Product schema with nested offers and aggregateRating tells a crawler, in plain terms, that the item costs £58, is in stock, and holds a 4.6 average across 213 reviews. With that in place the listing becomes eligible for review stars in the search snippet, and the same block of properties is the cleanest thing a model can lift when it summarises the product elsewhere.
That second use matters more each year. Answer engines such as ChatGPT, Perplexity, and Google AI Overviews tend to favour facts they can extract cleanly and attribute with confidence. A model that has to scrape a price out of prose may get it wrong or skip the page; a model handed a labelled offers block has a far better chance of citing your store accurately. Structured data does not buy you a citation, but it lowers the cost of being understood, and that is often the deciding factor in whether you are quoted at all.
The honest caveat: structured data does not guarantee a rich result or a citation, and it must accurately reflect what is visible on the page. Marking up ratings or reviews that a shopper cannot actually see is against Google policy and can trigger a manual action. The data has to be true and corroborated, not decorative. Treat it as a faithful description of the page, never as a lever to claim something the page does not show.