Structured data is machine-readable markup describing what a page contains. It does not make you rank higher. It changes how you can appear in results and how confidently a machine can understand you, which are different and increasingly more useful things.
How it works
The vocabulary is schema.org, a shared set of types and properties. The format Google recommends is JSON-LD, a block of JSON dropped into the page rather than woven through the HTML.
You describe the entities on the page: an article with an author and a publication date, a product with a price and stock status, a recipe with a cooking time. A rough example:
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Article",
"headline": "What structured data actually does",
"datePublished": "2026-08-17",
"author": {
"@type": "Person",
"name": "James Burfield",
"jobTitle": "Senior SEO consultant"
}
}
</script>The two properties worth understanding beyond the basics are @type, which says what kind of thing this is, and @id, which gives an entity a stable identifier so several blocks can describe the same thing without creating duplicates.
What it gets you
Rich results, when Google chooses
Certain types make a page eligible for enhanced appearances: review stars, product prices, recipe cards, event listings, breadcrumbs, job postings. Eligible is doing real work in that sentence. Google decides whether to show them, and it has withdrawn rich result types before, as it did when it cut FAQ and HowTo results for most sites in 2023.
So treat rich results as an upside rather than a plan. Anything built on the assumption that a specific enhancement will keep appearing is built on Google’s discretion.
Being understood without interpretation
This is the part that has quietly grown in importance. Structured data states facts about your content in a form a machine reads directly rather than infers from prose. As more surfaces are assembled by systems reading pages rather than people scanning them, saying plainly who wrote something, when it was checked and what it relates to is worth more than it was.
It’s why every entry in our research library carries a citation property naming the original researcher. The markup describes what the page actually is: our summary of someone else’s work, with the credit machine-readable.
The types most sites actually need
| Type | Use it for | Worth the effort? |
|---|---|---|
| Article / BlogPosting | Editorial content, with author and dates | Yes, and pair it with a real author entity |
| Person | The author, with jobTitle, knowsAbout, sameAs | Yes, this is where E-E-A-T becomes machine-readable |
| Organization | The site owner, with logo and contact | Yes, once, sitewide |
| BreadcrumbList | Navigation hierarchy | Yes, cheap and often displayed |
| Product | Ecommerce, with price and availability | Yes if you sell things |
| FAQPage | Question and answer sections | Marginal since Google cut most FAQ rich results |
| HowTo | Step-by-step instructions | Marginal, same reason |
| LocalBusiness | Physical premises | Yes for local, alongside Google Business Profile |
| DefinedTerm | Glossary entries | Yes for a glossary; this site uses it |
Common mistakes
- Marking up content that isn’t on the page. A direct structured data policy violation and it can earn a manual action. If the markup says there are 47 reviews, 47 reviews need to be visible.
- Self-serving review markup. Reviews you wrote about yourself are ineligible, and Google has tightened this repeatedly. Genuine reviews live on a third-party platform.
- Treating it as a ranking factor. It affects eligibility and understanding. Google has been clear it isn’t a ranking signal in itself.
- Leaving plugin defaults unexamined. SEO plugins generate a lot of schema automatically, and some of it will be wrong for your site. Check what yours actually outputs.
- Conflicting blocks. Two plugins each emitting their own Organization or Article markup produces contradictions. Where possible, enrich the existing block through a filter rather than adding a second one describing the same entity.
- Validating once and never again. Schema breaks silently when templates change. Google’s Rich Results Test and Schema.org’s validator both take seconds.
Related
EEAT covers why authorship markup matters and what it can’t do on its own. Technical SEO places this in the wider job.