Canonicalisation

Canonicalisation is the process of telling Google which version of a URL is the one you want indexed and ranked when multiple URLs contain the same or very similar content. Without it, Google has to guess - and when it guesses wrong, the result is diluted rankings, wasted link equity, and pages competing against each other for the same keyword.

The mechanism is a single line of HTML in the <head> section of a page:

<link rel="canonical" href="https://example.com/preferred-url/" />

This tag is called a canonical tag, or rel=canonical. It is a recommendation to Google, not an instruction. Google will usually follow it, but will override it if the canonical you declare conflicts with other signals on the page or points to a URL that does not load correctly.

Why duplicate URLs arise naturally

Most duplicate content is not created deliberately. It emerges from the way websites work. Common sources include:

• HTTP and HTTPS versions of the same page (http://example.com/page vs https://example.com/page)
• www and non-www variants (www.example.com/page vs example.com/page)
• Trailing slash variants (/page vs /page/)
• URL parameters from tracking, filters, or session IDs (/page?utm_source=email or /products?colour=red&size=M)
• Ecommerce product pages accessible under multiple category paths
• Paginated archive pages that largely repeat the content of page one
• Print-friendly versions of articles

A site can have dozens or hundreds of these URL variants without anyone having intentionally created duplicate content. Each one is a separate URL to Google, and without canonical signals, Google must decide independently which to index and rank.

What happens without canonical tags

When Google encounters multiple URLs with the same content and no canonical signal, three problems follow. First, link equity is split: if external sites link to both /page and /page/, the authority from those links is divided rather than consolidated to one URL. Second, Google may choose to index the wrong version - the one you do not want appearing in search results. Third, multiple near-identical pages can compete for the same keyword, cannibalising each other's rankings rather than one strong page dominating the position.

Self-referencing canonicals

Google's John Mueller has stated: "I recommend doing this kind of self-referential rel=canonical because it really makes it clear for us which page you want to have indexed." A self-referencing canonical is a canonical tag on a page that points to its own URL - confirming to Google that this is the version you want indexed, even when there are no obvious duplicates.

Adding self-referencing canonicals to every page on your site is considered best practice. It prevents Google from treating minor URL variations as duplicates and gives every page a clear, unambiguous signal about its preferred URL.

Canonical vs 301 redirect

Both canonical tags and 301 redirects address duplicate content, but they work differently. A 301 redirect permanently forwards one URL to another - the original URL is no longer accessible. A canonical tag marks a preference but leaves both URLs accessible. Use a 301 redirect when you want to permanently retire a URL. Use a canonical tag when you need the duplicate URL to remain accessible (for example, a product page accessible under multiple category paths) while still consolidating its ranking signals to one preferred version.

Common mistakes

Pointing a canonical to a URL that redirects. According to Semrush's canonical guide, this creates conflicting signals: the canonical tag says "this page is preferred" while the redirect says "this page is gone." Google will likely ignore the canonical and make its own choice. Canonical tags must always point to live, final destination URLs.

Multiple canonical tags on one page. Having two canonical tags in the <head> sends contradictory signals. Google will ignore both and decide independently.

Canonical pointing to a noindex page. A page marked noindex but canonicalised to from another page creates a conflict: you are telling Google to index the canonical but not to index the canonical itself. The two tags must align.

Canonicalising unrelated pages. Canonical tags are for duplicate or near-duplicate content only. Using them to consolidate authority across genuinely different pages sends a misleading signal that Google will typically ignore.

How to audit canonical implementation

In Google Search Console, open the Pages report under Indexing. The "Duplicate without user-selected canonical" and "Google chose different canonical than user" categories identify pages where canonical signals are missing or being overridden. Both are worth investigating - the first means you have unaddressed duplicates, and the second means your canonical signals conflict with something Google is seeing on the page.

Next: Crawl budget