Definition
Schema.org is a collaborative vocabulary, maintained as an open community project, for describing things
on the web: types such as Product, Article, Organization, Event and FAQPage, each with named
properties. Schema markup, or structured data, is the use of that vocabulary inside a page so that a
crawler can parse the page's meaning without inferring it from prose. Google's introduction to structured
data describes it as a standardised format for providing information about a page and classifying its
content, and says Google uses it to understand page content and to enable special search result features
and enhancements. The registry's one-line summary is JSON-LD markup describing entities like products,
articles and organisations, which is the common case; the vocabulary itself is broader than what any search
engine uses.
How it works
The markup lives in the page in one of the three formats Google documents: JSON-LD, a script block containing a JSON object and the format Google recommends; Microdata, attributes on the existing HTML elements; and RDFa, a similar attribute-based approach. The crawler extracts the data when it processes the page, validates it against the rules for each feature, and may use it to assemble a richer listing on the results page (SERPThe page of results a search engine shows.): review stars, prices and availability on a product, a recipe card, an event listing, breadcrumbs. Google's documentation is explicit that structured data is not a guarantee of a rich result, that markup must represent the visible content of the page and not describe things the visitor cannot see, and that pages violating the structured data guidelines can lose eligibility for rich results.
| Format | Where it sits | Google's position |
|---|---|---|
| JSON-LD | A script block, usually in the head, separate from the visible HTML | Recommended format |
| Microdata | Attributes on the HTML elements that display the content | Supported |
| RDFa | Attributes on HTML elements, also usable in non-HTML documents | Supported |
In practice
A site decides which of its templates match a documented rich result type, implements the required and recommended properties for that type, tests the page with Google's Rich Results Test, deploys, and then watches the enhancement reports in Search Console for errors and for the count of valid items. The structured data guide covers which types matter and how to implement JSON-LD at template level; the category and product page guide covers the product-specific case, where offer, price and availability properties need to stay in step with the visible page. The technical SEO hub places markup among the other crawl and index topics.
Product markup that stops matching the page
A store adds Product markup to every product template with Offer properties for price and
availability. For three months the Search Console report shows about 4,200 valid items and product listings
carry price and stock in the results. A pricing change is then deployed to the visible page but the markup
still reads from an old cache field, so 1,150 products show one price on the page and another in the
markup. Within weeks the enhancement report flags the mismatch, those listings lose the price enhancement,
and click-through on the affected pages falls by roughly a fifth until the markup is wired to the same
source as the displayed price.
What it is not
Schema markup is not a ranking factor in its own right; Google documents it as a way to make content eligible for features, not as a signal that lifts position. It is not a substitute for visible content, and marking up reviews, prices or questions that are not on the page breaches the guidelines. It is not the same as canonicalisation: markup describes a page, while the canonical tag (Canonical TagTells search engines the preferred URL.) says which of several duplicate pages should be indexed, and the two should agree on the URL. And it is not limited to Google: the vocabulary is shared, but each consumer documents its own supported subset, so the feature list on one search engine does not transfer to another.
Related terms
Search engine optimisation (SEOImproving visibility in unpaid search results.) is the discipline in which markup is one technical task among many, and the results page (SERPThe page of results a search engine shows.) is where its effect, if any, is visible as a rich result. The canonical tag (Canonical TagTells search engines the preferred URL.) sits alongside markup in the page head and the two must describe the same URL, which is a common failure on sites with parameter variants. Related ideas that are not schema.org include Open Graph tags for social previews, which use a different vocabulary and serve a different consumer, and the meta robots directives that control indexing; both are covered in the technical guide rather than here. Markup for video, jobs, events and local businesses follows the same pattern of a documented type with required and recommended properties.
Common mistakes
- Marking up content that is not on the page, which breaches the guidelines and can remove eligibility for every rich result on the site, not only the offending type.
- Letting markup drift from the visible content when prices, stock or authors change in one place but not the other, which is the most common way a working implementation fails later.
- Expecting a rich result because the markup validates; validity is necessary, and Google decides the rest based on the page, the query and the device.
- Implementing every type the vocabulary offers instead of the handful with documented search features for the page's content.
- Testing once and never monitoring the Search Console enhancement reports after releases, so errors accumulate unseen.
