Google Chose a Different Canonical: Compare These Three URLs

“Duplicate, Google chose different canonical than user” means Google did not adopt your canonical preference. The investigation needs three addresses: the URL you inspected, the canonical your site declared, and the canonical Google selected. Comparing only the tag with the inspected URL can miss the actual problem.

Sometimes Google’s choice is sensible and your declaration is wrong. Sometimes an important page has been grouped with a near-duplicate that should not represent it. Diagnose which situation you have before adding stronger-looking tags or changing URLs.

This differs from a missing canonical

When Search Console says Duplicate without user-selected canonical, Google did not find a preferred canonical for that URL. That separate issue is covered in our existing diagnostic guide.

Here a preference exists, but Google selected something else. Google’s Page indexing help specifically recommends reviewing the current page, your declaration and Google’s selection together.

Use this worksheet for an affected example:

B may be the same as A because the page declares itself canonical. C may be another path, a host variant or a URL on a different domain. Keep the exact addresses rather than shortening all of them to a page title.

First check whether you are comparing the same version

Search Console’s indexed view describes a recorded crawl and indexing decision. Your browser shows the current response. If the canonical changed after the recorded crawl, those are different versions of the page.

Run Test live URL to review today’s availability and tested HTML. Google says the live test cannot predict canonical selection, so a successful live test does not settle the disagreement.

For a public cross-check, our Canonical URL Checker reads initial HTML and HTTP canonical declarations, follows redirects and checks a bounded canonical chain. It does not render a JavaScript-only canonical or read Google’s index. Record any limit or unavailable result as incomplete evidence.

Compare the three pages in this order

1. Does your declared canonical represent the inspected content?

Read the main content on A and B, including the product, service, language and page purpose. A shared header is not evidence that two pages are duplicates. A category listing is usually a different thing from one product’s details.

If B is unrelated, repair the mapping that produced the declaration. A common implementation error is a template or SEO field copied from another page. Adding the same incorrect URL through an HTTP header would repeat the error.

2. What did Google receive from the preferred page?

Check B’s actual response and important resources:

• Does it deliver the expected page with a successful response, rather than redirect elsewhere?
• Is there a robots block, a noindex rule or an authentication challenge?
• Does the main content load for a signed-out visitor and in Google’s tested page?
• Are the HTML canonical and any HTTP canonical declaration consistent?

The Google Index Checker checks several of these public signals for one URL. It does not establish that B is indexed or explain Google’s entire duplicate cluster. Its canonical-target checks are bounded and same-host; inspect other destinations separately.

If an outage or deployment returned a generic holding page during Google’s last crawl, today’s normal page is only half the evidence. Compare the crawl date with deployment and server records before concluding that Google grouped the actual content incorrectly.

3. Why might Google have chosen C?

Now compare A with C. Does C contain the same main content? Is it a clean permanent page while A or B is a temporary or parameter version? Do your own navigation and sitemap favour C?

Google’s canonicalization troubleshooting guide covers incorrect CMS declarations, server configuration problems and unexpected cross-domain selection. An unfamiliar destination is a reason to preserve the evidence and investigate responses, rather than to assume a particular cause.

For an external C, the same review still applies. Check for an unwanted cross-domain canonical or redirect, copied content, and incorrect content being served by your host. A domain name alone does not prove hacking or establish that another site stole the page. Ask your developer or host to investigate with the exact URL and crawl date if you cannot explain the response.

Three worked decisions

These examples are illustrative. They show how to choose a correction; they are not claims about observed client indexing results.

A copied canonical field points at a category

The product details are different from the category listing. Google’s choice of the product itself is reasonable. Correct the product template to declare the product’s own stable URL, then inspect another product to learn whether the mapping error affects the whole template.

The goal is a correct declaration. There is no reason to remove a useful product page because Google declined its category canonical.

A sale path competes with the permanent product

An illustrative shop has the same jacket at /sale/blue-work-jacket/ and /products/blue-work-jacket/. Both declare themselves canonical, while prominent campaign links and the sitemap favour the sale path. Google selects the sale version.

If the permanent product is the intended representative, align the duplicate’s canonical, the preferred page’s self-reference, internal links and sitemap with that product URL. If the sale URL has no further purpose, a permanent redirect may be more appropriate. Check dependencies such as active campaigns before retiring it.

These are changes to express a chosen relationship. They do not compel Google to accept it.

Two pages should answer different needs

A repair business has /commercial-boiler-repair/ and /domestic-boiler-repair/. Each declares itself canonical, but their main content is almost identical. In the illustrative indexed record, Google selects one representative.

A different canonical tag cannot supply the missing distinction. Explain the actual differences where the business can substantiate them: equipment served, job scope, access requirements, relevant projects and the enquiry process. If both pages cover the same service, consider whether one page serves visitors better.

Do not add invented differences just to make the copy look unique. The decision is about useful separate services, rather than a target percentage of different words.

Make the technical signals agree with the decision

Once you have chosen the representative, review the affected template or URL pattern, not just one example:

Google’s canonical implementation guidance advises consistent references and discourages using robots.txt or noindex to select a canonical within the site. Blocking A can prevent Google from reading the very correction you want processed.

For translated pages, review the language relationship too. Do not automatically canonicalize every translation to English. Canonical preferences and language annotations need to reflect the actual equivalent pages.

What counts as resolved?

If C is the sensible permanent representative, accepting Google’s choice can be a valid outcome. Document it and align any incorrect declarations with your intended setup.

If you changed content or technical signals, record the release date, verify the public response and use an indexing request for important affected URLs where appropriate. Recheck the indexed record after Google crawls and processes the changes. Selection may still differ; there is no guaranteed deadline or tag that overrides Google’s judgement.

For a correctly recognised alternate, use the alternate page guide. For the underlying implementation choices, read canonical tags explained.

Google documentation and tool capabilities reviewed on 28 October 2026.