How schema markup works
A page’s visible content is written for people. Schema markup is a parallel description of the same content written for machines. It uses the shared vocabulary published at schema.org — types such as Article, Product, LocalBusiness, FAQPage, Event and Recipe, each with its own properties — so that a search engine reads a labelled fact instead of guessing from the layout.
The recommended format is JSON-LD, a block of structured data placed in the page’s code and invisible to visitors. Two older formats, microdata and RDFa, wrap the visible HTML instead; both still work, but keeping the labels separate from the design is far easier to maintain.
Marking something up does not force anything to happen. It makes a page eligible for a rich result — review stars, an expandable FAQ, product price and availability, event dates, breadcrumbs — and the search engine still decides whether to show one.
Why schema markup matters
Its first job is removing ambiguity. A line of text holding a name, a street and a phone number is obvious to a reader and merely plausible to a machine; the same details inside LocalBusiness markup are unambiguous. Clear entity data is how a search engine connects a business, an author or a product across the web.
Its second job is presentation. A listing carrying extra detail occupies more space and answers more of the question before the click, which tends to attract more attention than a plain result beside it. That gain comes from the listing looking more useful, not from any ranking boost — structured data is not a ranking factor in itself.
It matters more as answers move into summaries. Assistants and AI features lean on clearly labelled facts, and a page that states its author, dates, prices and answers in machine-readable form is easier to quote correctly.
Common mistakes with schema markup
Marking up something the page does not show is the mistake that draws penalties. Guidelines require structured data to describe visible content, so FAQ markup for questions no visitor can see, or review stars assembled from nothing, risks a manual action against the site’s rich results.
Duplication is the quiet one. A theme adds its own markup, a plugin adds more, and a page ends up declaring two different articles or three organisations. Search engines then have to choose, and often choose nothing.
The rest is ordinary carelessness: markup left on a template so every page claims to be the same product, required properties missing so an item never qualifies, or effort spent on a rich result the search engine no longer displays.
How to act on it
Start with the types that describe what you genuinely are. Organization or LocalBusiness for the site, Article on posts, Product on product pages, FAQPage only where real questions and answers appear on the page. Resist adding types simply because they exist.
Then test. The Rich Results Test shows what Google can use from a live address, the Schema Markup Validator checks the syntax more broadly, and Search Console reports errors across the whole site once the markup is live. Where the code comes from a plugin, check what it actually emits rather than trusting the settings screen, because structured data across a site stays reliable only when each page has one clean source of truth.
