What is a canonical tag, and when do you need one?
A signal telling search engines which version of a page is the preferred one when the same or very similar content is reachable at several URLs — so that ranking signals consolidate on one address rather than splitting.
The problem it solves. Duplicate content is usually accidental. The same page can exist at many URLs through:
Tracking parameters — ?utm_source=... appended to shared links creates a distinct URL for every campaign.
Sorting and filtering on category pages — the same products in a different order.
Pagination and session identifiers.
HTTP and HTTPS, www and non-www, and trailing slashes.
Print versions and AMP variants.
Syndicated content republished elsewhere with permission.
Why it matters. Links and other signals pointing at different versions of the same page are divided between them, so no single version accumulates the strength it should. It also wastes crawl budget on near-identical pages.
How the tag works. A rel="canonical" link element in the page head names the preferred URL. It is a hint, not a directive — search engines consider it alongside other signals and may choose a different canonical if the evidence points elsewhere, which surprises people who treat it as an instruction.
Self-referencing canonicals — a page naming itself — are standard practice and prevent parameterised versions being treated as separate.
Cross-domain canonicals are used for syndication, pointing to the original publisher.
Common mistakes:
Canonicalising every page to the homepage, which tells search engines the rest of the site is duplicate and removes it from results entirely. This is a genuinely catastrophic and not-uncommon error.
Canonical pointing to a page that redirects, or to a noindexed page, which sends contradictory signals.
Using canonical for pagination, pointing every page to the first — the second page is not a duplicate of the first, and this loses the rest.
Conflicting signals — canonical saying one thing while sitemaps, internal links and redirects say another. Consistency matters more than the tag alone.
Redirects are stronger where a page genuinely should not exist separately.