A canonical URL is the representative address for a page among duplicate or very similar versions. You can declare a preferred URL with rel="canonical", redirects and other signals, but Google makes its own selection. The canonical URL is the address; the canonical link element is one way to express your preference.
What is a canonical URL?
A product page might be accessible through a clean URL and an email-tracking version. If both show the same content, they can belong to the same duplicate group. Canonicalization identifies a representative URL for that group. Google’s canonicalization overview distinguishes the URL you prefer from the one its systems select.
This applies to duplicate or substantially similar content. Two services do not become duplicates simply because their pages use the same template. Do not canonicalize distinct pages together to avoid writing useful descriptions.
Why canonical URLs matter for SEO
Canonicalization helps Google consolidate signals for duplicate pages and show a representative version in search. Google generally crawls that representative more often and duplicates less often. A canonical annotation is not a crawl prohibition, a ranking guarantee or a promise that an alternate URL can never appear.
Normal duplicate content is not automatically a spam-policy violation. The practical problem is often inconsistent signals: the sitemap favors one URL, internal links favor another, and the page annotation points somewhere else.
How to declare a canonical preference
For HTML, place the link element in the document’s head. This is an illustrative tracking-URL example, not an instruction to change this page’s own canonical:
<!-- Both /guide/ and /guide/?ref=email show the same guide. -->
<link rel="canonical" href="https://example.com/guide/">Google supports a canonical HTTP Link header for eligible non-HTML documents as well. Use absolute URLs to avoid deployment ambiguity. Relative paths are supported, but Google recommends absolute paths. These are implementation details in its canonical annotation guidance, not a reason to assume every relative tag is invalid.
Canonicalization and other indexing controls
| Control | Purpose | Visitor behavior |
|---|---|---|
| rel="canonical" | Declare the preferred representative of duplicate content. | Both URLs can remain accessible. |
| Permanent redirect | Move requests to a replacement URL. | The visitor follows the new address. |
| noindex | Exclude the fetched page from search results. | The page can remain accessible. |
| hreflang | Describe language or regional alternatives. | It does not redirect the visitor. |
Google describes both redirects and canonical annotations as strong canonicalization signals; sitemap inclusion is weaker. Hreflang expresses a different relationship. It is not a replacement for duplicate selection or a reason to point every translation at an English page.
Examples that need different canonical decisions
Tracking parameters: if an email parameter changes attribution but not the guide’s content, declare the clean guide URL as preferred. Use the same clean address in contextual links.
Pagination: page two of a product collection can show different products from page one. Google recommends giving each page its own canonical URL rather than pointing the sequence at the first page. See its pagination guidance.
Filtered pages: decide whether the filter creates genuinely different, useful content. A parameter is not proof of duplication. Do not discard a meaningful filtered collection merely because its URL contains a query string.
Canonical URL or 301 redirect?
Use a 301 redirect when the old address should send visitors to its replacement. Use a canonical annotation when duplicate versions still need to be available. Choose from the visitor’s task first, then make the search signals consistent.
Neither method makes unrelated pages interchangeable. A destination should represent the source’s content. RFC 6596 defines the canonical link relation for duplicative content and warns against declarations that lose the source’s information.
Checks before changing a canonical URL
- Compare the content. Confirm that the URLs are duplicates or sufficiently similar, rather than distinct pages worth retaining.
- Check the destination. Prefer the intended live, indexable URL, not a missing page or an intermediate redirect.
- Check the output. Inspect the HTML head or response header. Resolve conflicting declarations rather than adding another tag.
- Align references. Use the chosen URL in internal links and the sitemap. Check language annotations separately where applicable.
- Check Google’s selection. Inspect the indexed URL in Search Console and compare the user-declared and Google-selected canonicals.
Common mistakes and their limits
A homepage canonical repeated across distinct articles is not a sensible duplicate map. A target that returns a 404 error, contains no useful main content or conflicts with an indexing rule also warrants investigation. Correct the actual target and page state rather than assuming another tag will repair them.
RFC 6596 allows a target on a different domain. However, Google does not recommend canonical links as the way to keep syndicated copies out of search. For that goal, its current syndication guidance recommends that partners block indexing of the copies. Agree on the outcome and implementation with the publisher.
Checking canonical URLs on a small business site
Your CMS or SEO plugin may generate the annotation. Verify the output rather than assuming every platform has identical defaults. Check a normal service page, a tracking URL and any paginated collection you use. Record which template or setting controls an incorrect declaration so the correction is maintainable.
Search Console can show a different Google-selected canonical even when a tag exists. A live test does not evaluate all canonical conditions. Use the indexed inspection details and compare the pages, following the Page indexing report guidance.
Frequently asked questions
Must every page have a canonical tag?
Google can select a canonical without an explicit tag. A self-referencing annotation is a useful recommended practice on a canonical page, especially when duplicate variants exist; it is not a requirement for every page to be indexed.
Can Google ignore the canonical URL I declare?
Yes. A declaration is a preference, not a command. Google can select a different representative when the content or technical signals indicate another URL. Compare the user-declared and Google-selected canonicals in indexed URL inspection details.
Does a canonical tag stop crawling or replace noindex?
No. Canonicalization concerns which duplicate represents the content. It is not a crawl block or an indexing-exclusion rule. If exclusion is the intended outcome, use an appropriate noindex rule that Google can fetch and read.
Related glossary terms
Sources and editorial note
This entry was revised with AI assistance using the primary references above. URL examples are illustrative. It does not claim platform testing, customer results or named expert approval; the original publication date remains unchanged.