What Is Structured Data (Schema) and How to Set It Up

Two recipe sites: in one's result, stars, cooking time and calories show; in the other's, a dry blue link. The content quality may be the same; the difference is which language the page speaks with Google. The first speaks the machine's language too.

Schema markup is the code layer that explains your page's content to machines in structured form: "this is an article, this is its author, this its date; this is a question, this its answer." Google uses it for rich results, AI systems to understand the content. This article focuses on the practical side: which schemas give real benefit and how to attach them.

The format question: JSON-LD and nothing else

Three technical formats exist (JSON-LD, Microdata, RDFa); the real answer is one: JSON-LD. The format Google recommends, that does not touch the HTML, and that can be read and written. It looks like this: one script block placed in the page's head/body:

<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","headline":"...","author":{...}}</script>

The important principle: schema must describe the content VISIBLE on the page. Writing a star rating into the schema that does not exist on the page is a rule violation and a penalty topic; schema is translation, not decoration.

The schema types that give real benefit

The priority rule: start with the 2–3 types fitting your field. For a blog the Article + FAQPage pair, for a local service LocalBusiness + FAQPage, for a store Product is the standard start.

The application routes: from no-code to code

• WordPress: the SEO plugins (Rank Math, Yoast) attach the Article, Breadcrumb and FAQ blocks via settings; in most cases enough.
• Builders: some have a ready section; if not, the JSON-LD block can go into the custom code field by hand.
• Writing by hand: the schema.org documentation + adapting from an example; telling ChatGPT "write FAQPage JSON-LD for this page" also gives a fast draft — on condition of checking.
• Google Tag Manager: injecting without touching the code; the middle road for sites without a technical team.

Testing: is it attached, does it work?

Two official tools: the Rich Results Test (search.google.com/test/rich-results; which rich results this page qualifies for) and the Schema.org validator (the syntax check). The workflow: publish the page → run it through the Rich Results Test → fix the errors/warnings → track long-term in Search Console's Enhancements section. A note: the "eligible" status is not a display guarantee; Google shows the rich result when it deems it deserved.

Schema's new role in the AI search era

In classic SEO, schema was a "rich appearance" topic; in the AI era its weight has grown: for ChatGPT-type systems and Google's AI overviews, structured data is a clear signal while processing content; who wrote it, when, which questions and answers exist. This is precisely one of the technical foundations of the GEO/AEO strategy: presenting your answers in a form machines can cite. FAQPage is, from this angle, the most "profitable" schema: the question-answer format is the native language of AI answers.

Sources and further reading

Where to verify the source

The supported-types list gets changed by Google; check the current state in the official document.

• Google — structured data introduction: the official rules
• Schema.org: the dictionary of all types
• Rich Results Test: the checking tool

Continuing the topic

Schema is one part of the technical layer; the whole picture:

• The technical SEO checklist
• AEO and GEO: being visible in AI search
• On-page SEO: the full checklist
• The Search Console guide
• Other articles on this topic

The starting task: run your most important page through the Rich Results Test. If the result is empty, attach your first schema this week; do not be late to the conversation with the machines.