Canonical tags: how Google uses them and how to set them correctly

What a canonical URL is

Google describes it in one sentence: “Canonicalization is the process of selecting the representative –canonical– URL of a piece of content.” When the same content is available at several addresses, Google groups those URLs into a cluster and chooses one of them to show in search. The others stay known to Google but are normally not shown.

Duplicates appear on almost every site without anyone planning them: tracking parameters, sorting options, http and https, www and non-www, URLs with and without a trailing slash. Google is explicit that this is not a crime: “Some duplicate content on a site is normal and it's not a violation of Google's spam policies.” The goal is not to remove every duplicate, but to make sure Google picks the URL you want.

Why canonicals matter

Google lists three reasons to specify a canonical:

• “To specify which URL that you want people to see in search results.”
• “To consolidate signals for similar or duplicate pages.”
• “To avoid spending crawling time on duplicate pages.”

In practice the second reason matters most. Links and other signals that point to different versions of the same page are combined on the canonical. Without a clear canonical they can be split, and Google may show a URL you did not want, for example one with a tracking parameter.

What Google officially says

The same documentation lists what not to do. Each of these is a direct quote:

• “Don't use the robots.txt file for canonicalization purposes.”
• “Don't use the URL removal tool for canonicalization.”
• “Don't specify different URLs as canonical for the same page using different canonicalization techniques.”
• “Don't specify a URL fragment as canonical.”
• “We don't recommend using noindex to prevent selection of a canonical page within a single site.”

All the ways to signal a canonical

There is no single “canonical switch”. Google weighs several signals together, and they are not equally strong.

301 redirect

Google calls a redirect “A strong signal that the target of the redirect should become canonical.” Use it when users do not need the duplicate URL at all: old URLs after a migration, http to https, www to non-www, one trailing-slash format. After a redirect, visitors and Googlebot both land on the main URL.

rel="canonical" link element

For rel="canonical" Google uses the same words: “A strong signal that the specified URL should become canonical.” Use it when the duplicate has to stay available to users, as with filters, sorting, tracking links or printable versions. The tag goes in the HTML <head>; “The rel="canonical" link element is only accepted if it appears in the <head> section of the HTML.”

rel="canonical" HTTP header

Files without HTML, such as PDFs, cannot carry a tag, so the server sends the canonical as an HTTP header instead. A typical case is a PDF that duplicates an HTML page:

Link: <https://example.com/guide/>; rel="canonical"

XML sitemap

Listing a URL in your sitemap is “A weak signal that helps the URLs that are included in a sitemap become canonical.” On its own it settles nothing, but it should never contradict your other signals: list only canonical URLs.

Internal links

Google's advice is to “link to the canonical URL rather than a duplicate URL.” If your navigation links to /shoes while the canonical says /shoes/, you are voting for both. Internal links are one of the most common reasons Google picks a different canonical.

HTTPS

When everything else is equal, Google “prefers HTTPS pages over equivalent HTTP pages as canonical”. Redirect http to https anyway; do not leave it to Google's preference.

How to write a correct canonical tag

The tag itself is one line in the <head>:

<link rel="canonical" href="https://example.com/shoes/">

<head>
  <link rel="canonical"
        href="/shoes">
  <link rel="canonical"
        href="https://example.com/shoes/?sort=price">
</head>
<body>
  <link rel="canonical" …>

<head>
  <link rel="canonical"
        href="https://example.com/shoes/">
</head>

The rules, in full:

1. Absolute URL. “Use absolute paths rather than relative paths with the rel="canonical" link element.”
2. In the <head> only. A tag in the body is not accepted.
3. One canonical per page. A theme and an SEO plugin that both print a canonical are a classic cause of conflicting tags. When they disagree, Google is likely to ignore them.
4. Self-referencing on the main page. The canonical page should point to itself. That removes doubt when someone links to it with a parameter.
5. Exact format. The same protocol, host (www or not), trailing slash and letter case as the URL you actually serve.
6. A healthy target. The canonical URL returns 200, does not redirect, is not noindex, is not blocked in robots.txt and is listed in the sitemap.
7. In the server HTML. Do not add or change the canonical with JavaScript. If the raw HTML and the rendered page disagree, you are sending two signals.

Typical canonical scenarios

Tracking and session parameters

?utm_source=…, ?gclid=… and session IDs never change the content. Canonicalise to the clean URL and never use these parameters in internal links.

Sorting and filters

A list sorted by price shows the same products as the default list, so canonicalise to the default. A filter that you deliberately built as a landing page, with its own title, text and search demand (for example “red running shoes”), should have a self-referencing canonical and be linked internally.

Product variants

If colours or sizes have no separate search demand and share one description, canonicalise the variants to the main product. If a variant is searched for on its own and has its own content, let it be its own canonical.

www / non-www, http / https

These variants are best fixed with site-wide 301 redirects. The canonical tag on the final URL then confirms the choice.

Trailing slash and letter case

/page and /page/, /Page and /page are different URLs. Choose one format, redirect the other, and use the same format in canonicals, links and sitemaps.

Pagination

Do not point page 2, 3 and so on to page 1. Each page in a series shows different items, so each one should have a self-referencing canonical. Google no longer uses rel="prev" and rel="next" as an indexing signal. What matters is that the pages are linked to each other with normal links.

Syndicated or republished content

When your article is republished on another site, the copy can carry a canonical pointing to your original (a cross-domain canonical). Like every canonical, it is a hint, so also ask for a visible link to the original.

Separate mobile URLs

With m.example.com, the mobile page has rel="canonical" to the desktop page and the desktop page has rel="alternate" to the mobile one. Responsive design avoids this setup entirely.

Language and regional versions

“Different language versions of a single page are considered duplicates only if the primary content is in the same language”. So the German and French versions are not duplicates of each other. Each one gets a self-referencing canonical, and hreflang links them together. Never point all languages to the English page. Google's rule: “make sure to specify a canonical page in the same language.”

How Google can ignore your canonical

Because the canonical is a hint, Google weighs it against everything else it sees. It is likely to choose differently when:

• The target is broken. It redirects, returns 404, is noindex or is blocked in robots.txt.
• Signals disagree. The canonical says A, but internal links, the sitemap or hreflang point to B.
• There is a chain. Page A points to B, and B points to C or redirects to C.
• The content is different. A canonical from a product page to its category, or from every post to the home page, connects pages that are not duplicates. Google usually ignores it. If it does not, the product pages drop out of search.
• There are several tags. Two different canonicals on one page, or one in the server HTML and another added by JavaScript.

Canonical statuses in Search Console

Three statuses in the Page indexing report are about canonicals:

In URL Inspection, the two lines to compare are User-declared canonical and Google-selected canonical. When they match, your setup works. When they differ, Google has seen a stronger signal pointing elsewhere.

Common mistakes

• ✕Relative or partial URLshref="/shoes" instead of the full https:// address.
• ✕Two canonicals on one pageUsually a theme and an SEO plugin both printing one.
• ✕Canonical to a URL that redirects or is noindexThe target has to be the final, indexable URL.
• ✕Everything canonicalised to one pageProducts to the category, posts to the home page, all languages to English.
• ✕Pagination pointing to page 1Each page in a series should point to itself.
• ✕Internal links to the duplicateMenus, breadcrumbs and product cards that use a different URL format than the canonical.
• ✕robots.txt on duplicatesA blocked URL is never fetched, so Google cannot see its canonical.
• ✕Canonical set by JavaScriptThe server HTML and the rendered page should say the same thing.

How to audit canonicals step by step

1. Crawl the site and export the canonical for every URL.
2. Find pages without a canonical. Indexable pages should have a self-referencing one.
3. Find canonicals pointing elsewhere and check that each pair really is a duplicate.
4. Check every target: status 200, no redirect, no noindex, not blocked, in the sitemap.
5. Find technical errors: relative URLs, several tags on one page, a tag in the body, a mismatched protocol, host or slash.
6. Compare raw and rendered HTML on key templates to make sure JavaScript does not change the canonical.
7. Check internal links and the sitemap: both should use canonical URLs only.
8. Check Search Console: the three statuses above, then URL Inspection on samples.

FAQ

Sources

• Google Search Central — What is canonicalization
• Google Search Central — How to specify a canonical URL
• Google Search Console Help — Page indexing report

Quotes checked against the official pages on 29 October 2026.