Canonical tags: what canonical means in SEO, how to set one, and the mistakes that cost pages their index spot
The syntax, the signals Google weighs alongside your tag, when to use a canonical instead of a redirect or a noindex, and how to diagnose a canonical problem in Search Console.
Published by AI Knows Us · Updated 12 October 2026
A canonical tag is a line in the head of a page that names the URL you consider the master version of that content. In SEO, "the canonical" means the one URL a search engine chose to index and show when several addresses serve the same or nearly the same page. Your tag is one of the signals Google weighs, alongside redirects, internal links, which URL is in your sitemap, and which version is on HTTPS. It is a strong signal when every other signal agrees with it, and it is ignored when they contradict it. Google documents the whole mechanism in its guide to consolidating duplicate URLs.
This page covers the syntax, the decision of when to use a canonical, and how to diagnose one in Search Console. For removing a page from search entirely, that is a noindex job, and Google's note on blocking indexing covers it.
The syntax
<head>
<link rel="canonical" href="https://example.com/boots/leather-chelsea" />
</head>
Four rules get this right every time.
- Absolute URL, with the protocol and the hostname. A relative path resolves against the base URL and goes wrong the moment anything else changes.
- In the head, and only once. Two canonical tags on a page tell Google nothing, so it falls back to its own judgement.
- Exactly the URL you want shown, including the trailing slash or its absence, the case, and no tracking parameters.
- For non HTML files such as a PDF, send it as an HTTP header instead:
Link: <https://example.com/guide.pdf>; rel="canonical".
When to use a canonical, and when to use something else
| Situation | Use this | Why |
|---|---|---|
| Tracking, sorting or filter parameters on the same page | Canonical to the clean URL | Consolidates the duplicates while the parameter URLs stay usable |
| You moved a page permanently | A 301 redirect | A redirect is an instruction, not a hint, and it moves visitors too |
| www and non www, or http and https | 301 redirect to one version, plus a self referencing canonical | Belt and braces, and the redirect stops the problem recurring |
| A page you never want in search results | noindex | A canonical consolidates. It does not remove |
| Paginated archive pages | Self referencing canonical on each page | Page two is not a duplicate of page one. Pointing it at page one hides its content |
| Printer friendly or AMP style alternate versions | Canonical from the alternate to the main page | One version collects the signals |
| Your article republished on a partner site | Cross domain canonical on their copy, pointing at yours | Supported by Google, and the only clean way to syndicate |
| Same content in different languages | Self referencing canonicals plus hreflang | These are alternates for different audiences, not duplicates |
| Near identical product variants | Canonical the variants to a single product page, or write genuinely different pages | Pick one approach per catalogue and apply it consistently |
Add a self referencing canonical to every page as the default. It is not required, and it costs nothing, and it removes a whole class of accidental duplication when somebody shares your URL with a campaign parameter attached.
The mistakes that actually cause damage
- Every page canonicalised to the homepage. The single most destructive canonical error there is. It tells Google your whole site is one page, and the rest drop out of the index. It usually arrives through a theme setting or a misconfigured plugin.
- Staging canonicals on the live site. A launch where the tags still point at
staging.example.com. Check this on day one of every launch. - Canonical to a URL that redirects. Google has to follow a chain to work out what you meant. Point at the final destination.
- Canonical to a noindex page. Two contradictory instructions, so you get whichever one wins rather than the one you wanted.
- Canonical to a 404. The signal is discarded and Google chooses for itself.
- Two SEO plugins both emitting a canonical. Run one. View source and search for
rel="canonical", and if you find two, that is the bug. - Page two of a list canonicalised to page one. Everything beyond the first page becomes invisible, which matters on category pages and large blogs.
- Canonical injected by JavaScript after load, conflicting with one in the HTML. Serve one, in the source.
Diagnosing it in Search Console
The URL Inspection tool is the fastest answer. Paste the URL and read two lines: the user declared canonical, which is your tag, and the Google selected canonical, which is the decision. When they differ, Google found stronger evidence somewhere else, usually in your internal links or your sitemap.
In the Page indexing report, three statuses are canonical statuses, and each means something specific.
- Alternate page with proper canonical tag. Working as intended. The duplicate is recognised and the master is indexed.
- Duplicate, Google chose different canonical than user. Your tag was overruled. Make your other signals agree: link internally to the URL you want, put only that URL in the sitemap, and check for redirect chains.
- Duplicate without user selected canonical. You have duplicates and no tag at all. Add self referencing canonicals.
What gets asked about canonical tags
Is a canonical tag a directive or a hint?
Google treats it as a strong hint and weighs it against the other signals. The way to make it decisive is to make everything else agree with it: internal links, sitemap entries, redirects and the version you promote.
Does a canonical pass link value to the chosen URL?
Yes. Consolidating duplicates is the whole point, so links and other signals pointing at the duplicate are credited to the canonical version.
Do I need a self referencing canonical on every page?
It is not required, and it is good practice. It protects you against parameter duplicates you did not create, such as a URL shared with campaign tracking attached.
Canonical or redirect?
If both URLs need to stay reachable for a person, use a canonical. If one is finished with, use a 301 redirect. Redirect when you can, because it is the stronger and clearer signal.
Why is Google ignoring my canonical tag?
Something contradicts it. Work through this list: your internal links point at the other URL, the sitemap lists the other URL, the tag points at a redirecting or missing page, there are two tags on the page, or the two pages are different enough that Google reads them as separate content. Fix the contradiction and re-inspect.
Can I canonical to a page on another domain?
Yes, and this is the correct setup for syndicated articles. The partner's copy carries a canonical pointing at your original. Confirm it is live in their rendered HTML, because a syndication deal without that tag puts their copy in the results instead of yours.
How long until a canonical change takes effect?
Google has to recrawl both URLs first, so allow days to a few weeks. Speed it up by requesting indexing on the canonical URL in URL Inspection and by linking to it from pages that get crawled often.
Duplicates that no tag can merge
A canonical merges two URLs of the same page. Nothing merges two different pages that say the same thing: your guide, a rival's guide and the guide an assistant writes for the buyer. The reader gets the third, and it is a new document, not a copy of yours. The useful question is whether your page is a source for it. Ask, then open the citations. AI Knows Us does this for a whole list of questions.