What is schema markup?
Schema markup is a label on the page for machines. Humans can tell that a page shows a product with a price, a film with a running time or an article by a named author. A crawler sees text and tags. Schema markup spells it out in a shared vocabulary: this is a Product, its name is this, its price is this, its rating is this; this is a VideoObject, here is its thumbnail and duration; this is an Organization, here is its logo and its address. The vocabulary is maintained at schema.org, founded in 2011 by Google, Microsoft, Yahoo and Yandex so that every search engine would read the same labels.
The markup can be written three ways: JSON-LD, a block of structured text in the page head; microdata, attributes woven into the HTML; or RDFa, an older sibling. Google recommends JSON-LD, and almost all new markup uses it, because it sits apart from the visible HTML and can be generated by a template without touching the design.
Why it matters
Why schema markup matters
It is the difference between being understood and being guessed at. Search engines have become good at inferring what a page is about, but inference is probabilistic; markup is a statement. A page with correct schema markup is eligible for rich results, the listings with ratings, prices, images, FAQs and event details that take more space and earn more clicks. A page without it is a plain blue link, however good the content.
It also feeds the systems beyond the results page: knowledge panels, voice assistants reading an answer aloud, AI summaries citing a source. Each of those draws on structured data where it exists, and a brand that has described itself clearly in markup is a brand those systems can represent correctly.
In practice
How schema markup works in practice
- 01
Start with the types that match what the site has. Organization markup on the homepage, with the logo, the address and the social profiles. Article markup on posts, with the author and dates. Product markup on product pages. VideoObject markup wherever there is a film, with its thumbnail, its description and its duration. FAQPage markup where there are real questions and answers. Most content management systems generate the basics; the rest is templated so every page of a type gets the same structure with its own values.
- 02
Then validate and monitor. Google's Rich Results Test checks a page for eligibility and errors; the Schema Markup Validator at schema.org checks the syntax of any type. Search Console reports each markup type across the site, listing pages with errors and warnings. Markup is maintained like any code: when the template changes, the markup is re-tested, and when Google changes which types it displays, as it periodically does, the list is reviewed.
At RIOT
Schema markup at RIOT
This glossary is built with it. Each entry carries FAQ markup for the questions at its foot, so the questions and answers exist as structured data as well as text on the page. It is the practical version of a rule we apply to every site we design: the page should describe itself to a machine as clearly as it presents itself to a person.
Watch out
What people get wrong about schema markup
Marking up what is not there. Google's guidelines require structured data to reflect the visible content, and markup that claims reviews, prices or questions the visitor cannot see is treated as spam. The penalty is a manual action that removes rich results across the whole site.
The other mistake is adding every type that validates. A page that is an article does not need Product and Event markup because a plugin offered them. Pick the types that describe the page honestly, fill the required properties completely, and stop.
