Structured data explained: telling search engines what your page is

Structured data is a vocabulary for telling engines the truth the page already contains: it cannot make a thin page rank, it makes a true page legible.

Somewhere in every redesign conversation the term appears: structured data, usually offered as a spell, a small block of code that persuades search engines to treat the page kindly. The buyer hears a shortcut and pays for a shortcut. What it actually is sounds duller and is worth more: a description the page carries about itself, in a vocabulary machines can parse, saying this page is a product, this one an article, this block an address.

That is the honest mechanism. A search engine already reads the words; the markup repeats the page's own truth in a form the engine does not have to guess at. Legibility, not sorcery: it cannot make a thin page rank, because there is nothing underneath to be legible about. What it does is make a true page unambiguous, so the engine's understanding of what the page is rests on the page's own statement instead of inference drawn from prose.

The failures come from the gap between the description and the page. Markup describing a product on a page that has no product: the description and the visible content disagree, and engines are built to notice exactly this. Markup copied from a template, naming an organization that never matched the site, so every statement the page makes about itself is attributed to a company that does not exist. And the schema added once at launch and left behind through two redesigns, now describing a site that is no longer there.

The practice follows from what the thing is. Describe only what the page genuinely contains, no aspiration and no padding. Keep the entity descriptions, the name, the address, the opening facts, consistent with what the visible page shows, because the markup is a statement about the page and the page is the evidence. And re-check the markup whenever the site's structure changes: a redesign, a platform move, a template edit, any event that can quietly leave the description describing yesterday.

It helps to see where this sits among the page's other signals. The sitemap and the robots file are the map and the door policy. Breadcrumbs are the visible trail that a visitor and an engine can both follow. The canonical address settles which of two URLs counts when the same content lives at both. Structured data is none of those; it is the page telling the engine what it is, once they have all arrived.

We ship it as a standard part of the sites we build, and the rule we hold ourselves to is the dull one: nothing goes in the markup that the visitor cannot also see. Guidance, not guarantees: engines decide for themselves what they use and how they present it, and no markup buys a particular display; the details of each engine's handling live in that engine's own documentation and change without notice. Describe your pages truthfully and you have done the part that was ever yours to control.