In short. Microdata embeds structured data in HTML through attributes that identify items, their types and their properties.

Definition: Microdata is an HTML syntax for embedding structured data, meaning information labelled for machine processing, in page elements. The WHATWG HTML Standard defines it as groups of name-value pairs called items, whose properties can contain text or nested items.

Key facts

The WHATWG Microdata specification uses itemscope to create an item and itemprop to label its properties.
Schema.org's Getting Started guide combines Microdata with a vocabulary of types and properties, including Movie and name.
As of 17 October 2026, Google's structured data introduction lists Microdata among its supported formats and generally recommends JSON-LD, a notation embedded in an HTML script element.

How do the HTML attributes work?

In the WHATWG model, itemtype supplies the item's type as a URL on the element carrying itemscope. That type establishes the vocabulary used to interpret property names. Adding itemscope to an element with itemprop makes that property's value another item.

Values depend on the HTML element. A link can supply its href, an image its src, and a time element its datetime. The optional itemref attribute associates properties elsewhere in the document through element IDs. The optional itemid supplies a global identifier when the chosen vocabulary supports one.

What does Microdata look like?

This shortened example follows Schema.org's Movie example:

<div itemscope itemtype="https://schema.org/Movie">
  <h1 itemprop="name">Avatar</h1>
</div>

The div encloses an item of type Movie. The heading supplies its name property with the value Avatar. The example illustrates syntax; it is a reduced extract of the guide's markup.

The same guide shows a nested Person item for the movie's director, with its own name property. Nesting therefore assigns the person's name to the director item. Schema.org also explains that additional span elements can attach properties to the appropriate text without changing how that text renders.

How should a website check the markup?

Google's structured data introduction treats its feature documentation as definitive for Google Search behavior. Each feature specifies required, recommended and optional properties. The document recommends the Rich Results Test during development and rich result status reports after deployment, when templating or serving issues can break markup.

Google's general guidelines require structured data to represent the page's visible content accurately. All required properties for the intended rich result, a search appearance enhanced with structured information, must be present. The guidelines also require pages with structured data to be accessible to Googlebot. Technical testing and checking the marked-up information against the visible page address different parts of those requirements.

What is its role in Google AI search?

Google's AI features guidance says that AI Overviews and AI Mode require no special Schema.org markup. It recommends matching structured data to the visible text. A page must be indexed and eligible to show a text preview in Search before it can qualify as a supporting link.

FAQ

Can I keep Microdata instead of switching to JSON-LD?

Google's format guidance accepts all its supported formats when valid and implemented according to the feature documentation. It recommends choosing the format easiest to implement and maintain, usually JSON-LD.

Can one item contain repeated properties?

Yes. The WHATWG specification allows properties with the same name and different values. Its ice cream example repeats flavor for lemon sorbet and apricot sorbet.

What should I check beyond syntax?

Google's quality guidelines require relevant, current information that represents the page accurately. Check that the marked-up information is visible to readers and follows the intended feature's additional content policies.

Sources

Check how a page reads to answer engines. Open the GEO score analyzer