Schema.org

Schema.org is a shared vocabulary that lets a website tell machines what its content actually is. A web page written in plain HTML shows text and images a person can read, but a search engine sees mostly unlabeled words and has to guess whether "$40" is a price, a rating, or a page number. Structured data fixes that by attaching explicit labels, and Schema.org is the agreed dictionary of labels everyone uses. It defines types — Product, Recipe, Event, Organization, Person, Article, and many more — and, for each type, the properties that describe it, like a Product's name, price, and availability. You add this vocabulary to your pages, most commonly with JSON-LD, a small block of code that states the facts in the shared terms. The markup does not change what a visitor sees; it adds a machine-readable layer underneath that says, in a language search engines already know, exactly what the page contains.

The vocabulary exists because it was a collaborative effort, not one company's invention. Schema.org was launched in 2011 by the major search engines — Google, Microsoft (Bing), Yahoo, and later Yandex — precisely so that publishers would not have to mark up their content four different ways for four different engines. One shared vocabulary means a site describes a recipe or a product once and every participating engine can read it. That is why Schema.org matters for anyone doing SEO: it is the standard route to being understood rather than guessed at. Well-formed markup can qualify a page for rich results — the star ratings, prices, FAQs, and event details that appear in search listings — and, more broadly, it feeds the knowledge graphs and answer systems that increasingly sit between a query and a click, all of which depend on structured facts they can trust.

Two distinctions keep people from confusing Schema.org with things it is not. First, Schema.org is the vocabulary — the set of types and properties, the meanings — while JSON-LD, Microdata, and RDFa are formats, the syntaxes you use to write that vocabulary into a page. You almost always want JSON-LD today, and search engines recommend it, but the choice of format is separate from the choice of vocabulary. Saying "we use JSON-LD" describes how the markup is written; "we use Schema.org" describes what it means. Second, structured data is a way to describe your content accurately, not a lever to inflate it. The markup must match what is genuinely on the page and visible to users. Marking up a review that does not exist, or a price the page does not show, is exactly the kind of misuse search engines penalize, and it can strip a site of rich results entirely.

It also helps to be clear about what Schema.org does and does not buy you. Adding markup does not, by itself, rank a page higher — it is not a ranking boost you sprinkle on. What it does is make a page eligible for rich results and easier for machines to understand, which can improve how the listing looks and how often it earns a click, and which increasingly feeds answer engines that quote or summarize content. Those are real, worthwhile gains, but they are downstream of accurate description, not of gaming a signal. The honest framing is that Schema.org is infrastructure for being understood correctly. Implemented faithfully, it helps the right page appear the right way for the right query; implemented dishonestly or sloppily, it invites errors, lost eligibility, or manual penalties, with none of the upside.

Match the markup to reality and to your goals. Pick the types that genuinely fit your content — Product for products, Article for articles, FAQPage only for real question-and-answer content — and populate the properties honestly with what the page actually shows. Prefer JSON-LD, which is the format search engines recommend and the easiest to maintain, and keep the structured facts in sync when the visible page changes, so the markup never drifts from the content. Validate what you publish with the testing tools search engines provide, because a small syntax error can void the whole block, and check which rich-result types are actually supported for your pages rather than marking up things no engine will surface. Done this way, Schema.org steadily improves how machines read your site and how your listings appear.

The failures are worth naming plainly. Marking up content that is not on the page, or that users cannot see, violates the guidelines and risks losing rich results or drawing a penalty — structured data must describe, not invent. Choosing the wrong type, or leaving required properties empty, produces markup that validators reject and engines ignore. Letting the markup fall out of sync with the visible page — a stale price, a rating that no longer matches — is both a quality problem and a trust problem. And treating Schema.org as a magic ranking trick leads to disappointment, because the value is eligibility and comprehension, not a direct boost. The discipline is simple to state and easy to neglect: describe your real content accurately, in the right types, keep it current, validate it, and let the honest description do its work.

Schema.org — launched in 2011 by Google, Microsoft, Yahoo, and Yandex — is a shared vocabulary of structured-data types that websites embed, usually as JSON-LD, so machines can understand page content.

Etymology: source.

Search interest for this term over the last five years:

View interest-over-time on Google Trends →

Areas of marketing where schema.org is a core concern: