Canonicalisation identifies a preferred representative for duplicate or substantially similar content. It should not erase the identity of genuinely different pages.
Identify true duplicates
Review parameter variants, alternate hostnames and repeated content before choosing the preferred URL. Distinct services or translated pages usually need their own identity. Do not point every article at a category page merely because they share a layout. Document why each canonical relationship exists.
Align the supporting signals
Use the preferred URL in internal links and the sitemap. Redirect obsolete variants when that fits the user journey. Ensure the canonical destination is accessible and represents the content. Avoid chains or pointing at a missing page. Review generated templates for accidental site-wide canonical values.
Inspect the selected result
Compare the declared canonical with the search engine’s selected version in available inspection tools. A canonical is a signal, not an absolute command. If the result differs, investigate content similarity, redirects and conflicting links. Fix the cause before repeating the declaration.