The same page can often be reached at several web addresses, and to a search engine those look like separate pages competing with each other. Canonical URLs are how you tell search engines which address is the real one, so their credit is combined rather than split. It is a small piece of markup that solves a surprisingly common problem.
How duplicate URLs happen
Duplicate addresses rarely appear on purpose. A product page might be reachable with and without a trailing slash, with tracking parameters appended by a campaign, through a session identifier, in a printer-friendly variant, or over both http and https. Each variation is a different URL to a crawler, even though a human sees one page. On a large site these multiply quickly, quietly scattering signals that should belong to a single page.
Why duplication is a problem
Search engines try to consolidate the signals a page earns — links, relevance, engagement — onto one URL. When the same content lives at five addresses, those signals are divided among them, and none of the five is as strong as the single page could have been. Search engines also have to guess which version to show, and their guess may not be the one you would choose. Canonicalisation removes the guesswork.
What a canonical tag does
A canonical tag is a line in the page’s head that names the preferred URL for that content. Placed on every duplicate, it points them all at one chosen address, telling search engines to treat that address as the definitive version and to consolidate signals there. It is a strong hint rather than an absolute command, but search engines generally respect it when the pages really are equivalent.
Choosing the canonical URL
The canonical should be the cleanest, most complete version of the page — usually the https, non-www or www form you have standardised on, without tracking parameters, and with a consistent trailing-slash style. Every duplicate, including the canonical page itself, should reference that same URL. Consistency is the whole point: mixed signals defeat the purpose.
Common mistakes
Several errors are easy to make. Pointing many genuinely different pages at one canonical can hide the others from search entirely. Canonicalising to a URL that redirects, or to a page that is blocked or missing, sends a confused signal. Setting a canonical to a non-equivalent page — say, all paginated pages pointing at page one — can suppress content you wanted indexed. The safe rule is that a canonical should point to a real, reachable page that is truly the same content.
Canonicals and other signals
Canonical tags work alongside, not instead of, other tools. Redirects physically move users and crawlers to one URL and are stronger when a page has genuinely moved. Robots directives control crawling and indexing separately. A canonical is the right choice when duplicate versions all need to exist but should be treated as one. Used consistently, it quietly keeps a site’s signals focused where they belong.