Canonical URLs explained: the tag that tells the search engine which address counts

Search engines do not punish duplicate pages, they guess which copy counts: the canonical tag is you replacing their guess with your answer.

A page earns its links over months, then a filter parameter quietly creates a second address for it, then a sorting option creates a third, and none of it looks like a problem because every address opens the same content. Weeks later the page sits lower than it should and nobody can say why. The site is not fighting a competitor. It is splitting its own case across copies.

The canonical URL is the fix, and it is smaller than the name suggests: one line in the page that names the address that should count. Where several addresses carry the same content, the tag tells the search engine which one is the original and which are echoes. The engine was going to pick one anyway. The tag is you picking instead.

It helps to know the duplicates are honest. The same article reachable from two paths, the print view, the parameters that track a campaign or hold a filter state: none of it is deception, and duplicate content is not a sin the engine punishes. What it does is guess. When several addresses carry one page, the engine has to decide which to index, which to show, and where the credit for every link goes. A guess made about your site is a decision made without you.

The failures follow a pattern. Every page canonicals to itself, including the ones drowning in duplicate variants, so the tag exists everywhere and helps nowhere. The tag points at an address that redirects somewhere else, sending the engine through a chain you built by accident. And the migration where everything moved but the old canonicals stayed behind, still nominating addresses on a domain the company no longer controls.

The practice: choose one address per piece of content, deliberately, and let it be the one every version agrees on. Let each page point to itself by default, and make the tag explicit only where a duplicate genuinely exists. Then, after any relaunch or migration, walk a sample of pages and check the tags against where the site actually lives now, because a tag is a statement about the future, and the future moved.

We check this line on every site we build and every one we inherit, because it is invisible when right and expensive when wrong. Guidance, not guarantees: how any particular search engine weighs the tag, and what your specific setup needs, belong to the engine's own documentation and to whoever audits your site. Pick the address. Say it once. Keep saying it after the move.