Canonical URL

Definition

When one document is reachable at several URLs (with and without tracking parameters, on http and https, under www and the bare host), search engines group the duplicates and keep one address as the representative. That address is the canonical URL. The others are treated as duplicates: they may be crawled, but they are generally not shown in results, and signals such as links pointing to them are consolidated onto the canonical.

Google's documentation lists the ways a site can state a preference, in order of strength: redirects (strong), rel="canonical" link annotations (strong) and sitemap inclusion (weak). The methods stack, and none is mandatory; without them Google picks a canonical on its own.

In practice

A rel="canonical" element must sit in the <head> of the page, and Google recommends absolute URLs and a self-referencing canonical on the preferred page itself. For PDFs and other non-HTML files the same hint can be sent as a Link HTTP header.

Common mistakes, per Google's guidance:

• Using robots.txt to canonicalize. A blocked URL can still be indexed without its content.
• Declaring one URL in the sitemap and a different one in rel="canonical".
• Using noindex to choose between duplicates, which removes the page from Search entirely.
• Pointing canonicals at URL fragments, which Google generally ignores.

Search Console's URL Inspection tool shows both the user-declared canonical and the Google-selected canonical, which can differ.

Relation to backlink indexing

Canonicalization matters on the linking side of a backlink. If a guest post is reachable at /post?ref=feed and that variant is what was submitted, Google may crawl it, file it as a duplicate and index only the clean URL. An index check on the parameter version then reports "not indexed" even though the content is in Google. The remedy is to submit and check the canonical address.

A page that declares a canonical on another domain (common with syndicated articles) consolidates toward that other URL, and the copy carrying the link may never be indexed on its own. A link indexing service can get Googlebot to fetch the copy, but it cannot override a canonical decision. The IndexChex backlink monitor records the canonical URL of each source page for this reason.

Related terms

• XML sitemap, a weak canonical signal
• noindex, which removes rather than consolidates
• Crawled, currently not indexed, a status duplicates often fall into
• robots.txt, which should not be used for canonicalization

Return to the glossary index.