Duplicate, Google chose different canonical than user: why it happens and how to fix it

“Duplicate, Google chose different canonical than user” means Google saw the canonical tag on your page and did not accept it: the page was placed in a duplicate cluster, and a different URL was chosen to represent that cluster instead of the one you declared. The content is in the index, but under the URL Google picked; your URL does not appear in search results. The fix is usually to make the other signals Google reads (redirects, internal links, the sitemap, the HTTPS version, hreflang) point at the same address.

Google says this plainly in its documentation: indicating a canonical preference is a hint, not a rule. Even if you explicitly designate a canonical page, Google might choose a different canonical for reasons such as the quality of the content or technical signals. The full list of reasons a page stays out of the index is covered as an ordered checklist in the diagnosis guide; the canonical conflict is one row of that list.

Three duplicate statuses, three different meanings

The Pages report has three rows that start with “Duplicate” or involve a canonical, and confusing them sends the fix to the wrong place. Google’s goal is a single one: every indexed page should be the canonical version, and duplicate or alternate pages should not be indexed.

Start by reading the two fields in URL Inspection

Open the affected URL in Search Console’s URL Inspection tool and look at two rows in the Page indexing section. “User-declared canonical” is the address your page declares through its canonical tag, an HTTP header or the sitemap. “Google-selected canonical” is the URL Google treats as authoritative when it finds similar pages on your site. In Google’s own words, this may be the user-declared canonical, but Google may sometimes pick a different URL it considers a better canonical example. The value of this field can lag the index by a few hours, so it does not reflect a fix you made minutes ago.

Putting the two fields side by side gives four pictures, and each one leads to a different job:

• The Google-selected URL is another version of your own page (https instead of http, the www version, a trailing slash, different letter case, a parameter): a signal conflict. Go to the list of causes below.
• The Google-selected URL is a different content page on your site: Google considers the two pages the same content. If they really differ, make the difference visible; if they do not, bind one to the other with a canonical or a 301 and accept it.
• The Google-selected URL is on another domain: syndication, a site that copied you, or a misconfigured server. That case has its own section below.
• The Google-selected URL and your declared URL look identical but the status is still listed: the report may be lagging. Check the last crawl date; if the page was crawled after your fix, start the process with “Validate fix”.

Why Google ignores your canonical

Google documents the signals it weighs besides rel=canonical when choosing a canonical URL: whether the page is served over HTTP or HTTPS, redirects, and the presence of the URL in a sitemap. The methods are not equally strong either: in Google’s guide to consolidating duplicate URLs, redirects are the strongest signal, rel=canonical comes second and the sitemap is the weakest. In most cases your canonical is ignored because one or more of these signals points somewhere else.

1. 01

   The canonical tag itself is wrong

   Use an absolute URL instead of a relative path such as /dresses/green/green-dress.html; that relative form is exactly the bad example Google gives. More than one canonical on the same page, a target that contains a fragment (#section), or a CMS plugin writing the home page into every page are common mistakes. Google notes that some CMSs and plugins can point to incorrect URLs; open the HTML in your browser’s developer tools, read the tag with your own eyes, and if the plugin is at fault, contact your CMS provider.

2. 02

   A redirect runs the other way

   Page A declares B as canonical, but B redirects with a 301 to A. Because the redirect is the stronger signal, Google picks A and ignores the canonical. Trace the chain from the command line; the redirect and the canonical must agree on one address. Google’s HTTPS example works the same way: if the HTTPS page redirects to HTTP, or the sitemap lists the HTTP version, Google may choose HTTP as canonical.

3. 03

   The sitemap lists another version

   A sitemap is a signal that the URLs it lists are canonical. If the canonical tag says https://www.example.com/page while the sitemap lists http://example.com/page/, the two signals conflict. A sitemap should contain only canonical URLs that return 200; how to generate one is covered in the sitemap guide.

4. 04

   Internal links go to the other version

   If the menu, the footer and in-content links go to a version with parameters, without www or with a different trailing slash, Google sees that version far more often. A canonical on its own may not outweigh that. Fix the links in the templates to use the final address; the internal linking guide explains how to plan this.

5. 05

   Different methods name different addresses

   Google asks you not to specify different URLs as canonical for the same page using different canonicalization techniques, and warns that using two methods at the same time is more error prone. The rel=canonical in the HTML and the Link: rel=canonical in the HTTP header must name the same address; picking one and removing the other is cleanest.

6. 06

   Trying to manage canonicals with robots.txt or noindex

   Google says not to use robots.txt for canonicalization purposes: blocked URLs can still be indexed, and Google cannot read a canonical on a page it cannot fetch. Using noindex to prevent a page from being selected as canonical within a single site is not recommended either. The way to keep duplicates out of the index is consolidation: a canonical or a 301.

7. 07

   Language and region variants without annotations

   Pages in the same language that differ only by region, such as separate English sites for the United States, the United Kingdom and Australia, fall into a duplicate cluster without hreflang annotations, and Google picks one. Connect the versions with hreflang. Pages where only the header and footer are translated while the main content stays in the same language are treated as duplicates by Google in any case.

8. 08

   The canonical page does not point to itself

   Google recommends including a rel=canonical link on the canonical page itself. If A points to B but B carries no canonical, or points to C, the representative of the cluster becomes unclear. Do not build chains: every page in the cluster should point straight to B, and B should point to itself.

Google may be right: are the pages really different?

If the technical signals are consistent and Google still keeps choosing another page, the reason may be content quality: Google asks you to make sure the pages being clustered are sufficiently different from each other. Colour and size variants of a product, listings that carry the same supplier text, and region pages generated for the same city with small changes are typical examples. Decide: if the pages deserve to stay separate, give each its own title, description and body text; if they do not, accept the page Google selected as canonical and bind the others to it. The “Crawled, currently not indexed” guide walks through that consolidation decision with examples.

Re-evaluation takes time. Google writes that it may keep a page in a duplicate cluster for up to two weeks; checking the report every day during that window adds no information. Once the signals are fixed you can request indexing from URL Inspection; Google notes that this request has a quota and should be saved for the most important URLs.

When the selected canonical is on another site

Google’s canonicalization troubleshooting document lists the cases where an external domain is selected one by one. Syndication, publishing your content on partner sites by agreement, is one of the most common; Google does not recommend the canonical element here, because the pages are usually very different. The recommended solution is for the partners to keep their copies out of the index. If a site that copied you without permission was selected, you can approach the host or file a copyright (DMCA) takedown. For misconfigured servers, such as one domain returning another domain’s content, go to the hosting provider. Injected 3xx redirects to another domain or cross-domain canonical links are a sign of a hacked site; in that case Google may select the malicious URL instead of your own.

After the fix: the validation sequence

• The canonical tag is an absolute URL, appears once, has no fragment (#) and the canonical page points to itself
• Redirects and the canonical lead to the same address; HTTP → HTTPS, the www preference and the trailing slash follow one form
• The sitemap contains only canonical URLs; the other versions were removed
• Template and in-content links go to the final address, not to a parameter version
• The HTML canonical and the HTTP header name the same address, or only one of them is used
• Region versions are connected with hreflang
• Pages you do not want treated as duplicates really have different content
• “Validate fix” was clicked once under the relevant reason in the Pages report

Fix every instance of the problem before starting validation; Google says the validation process usually takes about two weeks, and the list of example URLs in the report is capped at 1,000 rows. On large sites you will need to find the URLs missing from that list with your own crawler and apply the same fix to them.

To see which address the fixed pages were indexed under, the bulk index checker queries each URL separately, so you can read from the list whether the address you declared or Google’s old choice is in the index. If you have many canonical URLs that need to be crawled again, WeRankLab’s URL indexing tool checks every address before sending it: a URL with a canonical pointing elsewhere, a noindex tag, a robots.txt block or a redirect is not sent and not charged. That is why the list you submit should consist of canonical addresses only.

Frequently asked questions

Does this status hurt my rankings?

The content stays in the index, shown under the address Google selected. If that address is on your own site, search traffic goes to that page; the loss is mostly about reporting and link consistency. If the selected address is on another site, you do not receive the traffic, and that is a real loss.

I added the canonical tag again. Why has nothing changed?

The tag had already been read; the problem is not the tag but the other signals that contradict it. Check the redirects, the sitemap, internal links and the HTTPS version. After the fix, Google’s re-evaluation can take up to two weeks.

Can I just accept the page Google selected?

Yes, and it is often the shortest route. If the pages really carry the same content, point your canonical tag at the address Google selected and adjust internal links and the sitemap to match. On the next crawl the status moves to “Alternate page with proper canonical tag”.

Will blocking the duplicate in robots.txt end the problem?

No. Google says not to use robots.txt for canonicalization; a blocked URL cannot be crawled, so its canonical cannot be read, and it can still end up indexed. Consolidate the duplicate with a canonical or a 301 instead.