Definition:
The rel canonical attribute is used in a link element to indicate the preferred URL among several duplicate or closely similar pages.
The reference is usually placed in the document’s head and points to a canonical URL. Search engines interpret it as a signal rather than a binding directive, so they may select another version when they find conflicting evidence.
Table of contents
What rel canonical is used for
Canonicalisation helps consolidate signals associated with several URLs that display the same content or equivalent versions, such as pages with parameters, tracking variants or alternative paths. It also guides the search engine towards the URL that should appear in results and avoids unnecessarily dividing links and other signals between duplicates.
A canonical tag does not remove alternative URLs or prevent them from being crawled. Nor does it guarantee the indexability of the chosen URL. When an address has changed permanently and does not need to remain available, a redirect may be more appropriate. If the page needs to exist but should not appear in results, canonical and noindex perform different functions.
Search engines can select a canonical even when the website does not declare one. The tag expresses a preference and helps reduce problems caused by duplicate content, but it should be consistent with internal links, the sitemap, redirects and other technical signals.
Canonical tag syntax
The usual form inside the head of an HTML page is:
<link rel="canonical" href="https://www.example.com/page/">
The canonical value identifies the relationship and href contains the preferred URL. An absolute URL including the protocol and domain should be used. For non-HTML documents such as certain PDFs, the canonical relationship can also be communicated through an HTTP Link header.
Canonical tag fundamentals
A consistent implementation should take the following points into account:
- Use absolute URLs: Including the full protocol, domain and path reduces ambiguity.
- Point to a valid page: The canonical URL should respond correctly and be accessible and indexable.
- Include a self-referencing canonical: The preferred page can declare itself canonical to reinforce the selected version.
- Declare one consistent canonical: HTML, HTTP headers and plugins should not send different destinations.
- Align the signals: Internal links, the sitemap, redirects and language annotations should favour the same URL where appropriate.
The exact form of the path, including letter case, trailing slashes and parameters, should match the version actually used by the server. Converting every address to lowercase is not sufficient if doing so changes its meaning or response.
Errors when implementing a canonical tag
Common problems arise from incompatible signals or from selecting a destination that cannot act as the main version:
- Pointing to an error or redirecting URL: The canonical should lead directly to the final, stable version.
- Selecting a non-indexable or blocked page: A destination with
noindexor one that is inaccessible to the search engine creates conflicting instructions. - Publishing several canonicals: Duplicate tags or different destinations may cause the search engine to disregard those indications.
- Creating canonicalisation chains: Each variant should point directly to the preferred URL rather than pass through intermediate pages.
- Grouping non-equivalent content: The tag should not consolidate pages with substantially different intents, languages or content.
After implementation, the rendered HTML and headers should be reviewed, variants crawled and the URL recognised by the search engine as canonical checked. The website’s declaration and the search engine’s selection may differ.