Canonicalisation is the process of selecting a preferred URL when identical or substantially similar content is available at multiple addresses. A canonical URL helps clarify which version should represent the content. It is useful for duplicates, but cannot repair unrelated pages or replace a planned website migration.
Signals guide canonical selection
Google considers several signals, including redirects, canonical annotations, sitemap entries and links. A rel=canonical annotation communicates a preference rather than functioning as an unconditional directive. Search engines may select a different representative if the signals or content disagree. Keep the preferred version accessible and consistent across internal links and the sitemap. A canonical tag also does not move a visitor to another page; that behavioural change requires a redirect when a permanent move is intended.
Look for conflicting versions
Compare protocol, hostname, trailing slash and parameter variants of an important page. Inspect their content, response codes and canonical annotations, including the rendered result if scripts can modify metadata. Check the selected canonical in Search Console against the declared one. Look for circular references, non-existent targets and every category page accidentally pointing at the homepage. Fix template-level causes before editing hundreds of pages individually. Retest a representative set after the change and confirm that fresh sitemap entries use the same preferred format.
Review the preferred version’s own annotation as well as its duplicates. A suitable target can still point elsewhere through an inherited template rule, creating a contradictory chain of canonical preferences across equivalent URLs.
Separate duplicates from useful alternatives
Suppose a retailer exposes the same product through two category paths. A preferred product URL can make the intended representative clear while preserving useful browsing paths. A genuinely different regional service page is another situation: pointing it to a general national page may erase an intended distinction rather than solve duplication. Review whether the content and audience actually differ. Measure canonical outcomes at the URL level, and avoid using noindex as a general substitute for consolidating equivalent pages.
Canonical conflicts can be reviewed within technical SEO alongside links and redirects.
Common questions
Does rel=canonical guarantee Google's choice?
No. It is a canonicalisation signal. Conflicting content, redirects and internal links can lead Google to select a different representative URL.
Should every page canonicalise to the homepage?
No. That misrepresents distinct content. Useful unique pages normally identify their own preferred address unless a legitimate duplicate relationship exists.
Terms in this guide
- Canonical URL — The preferred URL for a page when identical or substantially similar content is accessible through multiple addresses. A canonical link communicates a preference rather than forcing a redirect.
- Redirect — A server or browser instruction that sends a request to another URL. Permanent URL migrations should avoid unnecessary chains and link directly to the intended destination.
- Sitemap — A structured list of relevant site URLs. XML sitemaps support discovery by search engines; an HTML sitemap helps people navigate the site.
- Internal link — A hyperlink between pages on the same website. Useful internal links connect related subjects, explain the site structure and give readers a relevant next step.
Related SEO guides
Apply this to your website
Explore Technical SEO, SEO Audits for help planning and implementing the relevant changes, or discuss your website with Refined IT.