Canonical URL
Canonical tags are a signal, not a command – I learned that the hard way when Google ignored my preferred URL and indexed a tracking-parameter mess instead.
Start here
- Run a full duplicate URL audit using a crawler before adding any canonical tags.
- For each group of near-identical pages, pick one canonical URL and stick to it.
- Never canonicalise a page to completely different content – Google will ignore it.
- Monitor Google Search Console's 'Page indexing' report to see which canonical Google actually selects.
- Add self-referencing canonicals on every page as a safety net against parameter duplication.
Plain-English take
I think of the canonical URL as a polite note you leave on your duplicate pages that says 'the real content lives over here, please go there.' It is not a redirect – the duplicate still exists – but it tells search engines which URL to index and which signals to consolidate. The standard implementation is the rel="canonical" link element in the <head> of a page. This is part of the broader [Directive](/directive/) system that guides crawler behaviour.
Why you need it: any site with multiple URLs serving the same content risks diluting ranking signals across those URLs. A canonical tag helps consolidate link equity and ranking power to one preferred version. Google defines canonicalisation as selecting the representative URL from a group of duplicate or near-duplicate pages. Without it, Google may pick a variant you do not intend – for example, a URL with tracking parameters instead of your clean base URL.
The key word is 'near-duplicate'. Canonical tags are not meant for pages that are merely similar in topic; they are for pages that are essentially the same content served at different URLs. E-commerce sites with [Duplicate Content](/duplicate-content/) from faceted filters are the classic case. You would canonicalise ?colour=red&size=m back to /category/ because the underlying product grid is identical. British spelling, by the way – I had to correct my own usage of 'colour' to 'colour' when writing examples.
When it actually matters
This is where canonical tags earn their keep. Three situations stand out.
E-commerce faceted navigation. Filter parameters like ?colour=red and ?size=large generate dozens of near-identical listings. Without a canonical pointing to the main category page, each filter combination can become a separately indexed page – a mess for crawl budget and ranking coherence. In a [Technical SEO Audit](/technical-seo-audit-service/) I ran last year, the client had 1,200 parameterised URLs for just 30 product categories. Adding a single canonical per category collapsed the indexable count to 30 and improved keyword rankings within six weeks.
Tracking parameters. ?utm_source=twitter and ?ref=newsletter should never create separate indexed pages. Use a canonical to the clean version of the URL. This is a quick win you can implement in any CMS or via your server-level redirects, but note that a canonical is not a substitute for a [301 Redirect](/301-redirect/) – both have their place.
WWW vs non-www and HTTPS vs HTTP. These are technical canonicalisation issues best handled at server level (via redirects). I prefer to set the preferred domain in Google Search Console and use a 301 from the non-preferred to the preferred. But even then, you should add a self-referencing canonical on the canonical version to reinforce the choice. This relates to [Website Structure](/website-structure/) best practices.
Content syndication. If you republish an article on a partner site, ask them to canonical back to your original. This is standard practice and avoids duplicate content penalties. However, you lose referral traffic because the partner page will not rank.
What I got wrong
I have made plenty of mistakes with canonicals, and admitting them saved me from repeating them.
Using canonicals as a lazy 301. Early in my career I had a page that should have been permanently moved. Instead of setting up a proper 301, I added a canonical to the new page. The result: the old page stayed in the index, diluted ranking signals, and confused users who landed on a dead URL. Now I know: a 301 redirect is for permanent moves; a canonical is for duplicate content that you keep available. This is a classic [SEO Audits](/seo-audits/) finding.
Canonicalising pages that were not duplicates. I once pointed a blog post about 'SEO tips' to a product page about 'SEO software' thinking it could pass relevance. Google ignored the canonical entirely, and the blog post lost its ranking. The lesson: canonicals are for duplicate content, not for consolidation of unrelated pages. If the content bodies differ, Google will treat the canonical as a suggestion and often ignore it.
Ignoring cross-domain pitfalls. I canonicalised a client's content to a partner site without realising the partner site set a different canonical. Google got confused and indexed a third URL. Cross-domain canonicals work only when both parties agree and the content is genuinely identical. Even then, test thoroughly.
Forgetting self-referencing canonicals. On many pages I simply omitted the canonical tag, leaving Google to guess. That often resulted in parameter-based duplicates being indexed as the primary. Now every page on my site has a self-referencing canonical, even if no duplicates exist. It is a safety net.
Next step
Quick answers
What happens if I canonicalise a page to a URL that returns a 404?
Google will ignore the canonical and may treat the page as orphaned or drop it from the index. Always ensure the target URL is indexable and returns a 200 status code. Check using the URL inspection tool in Google Search Console before finalising any canonical tag.
Can I use canonical tags across different domains?
Yes, cross-domain canonicals are supported by Google, but they are only effective if you own both domains and the content is essentially identical. If the content differs significantly, Google will ignore the canonical. Always test with the URL inspection tool before relying on them.
Should I use canonical tags on paginated pages?
For paginated series like blog archives, use rel="next" and rel="prev" rather than canonicalising to the first page. Canonicalising pagination suggests the first page contains all content, which is misleading and can cause Google to ignore the canonical in favour of the original series.
Sources
Primary documentation is linked directly. Anything commercial is marked nofollow.
- Google Search Central — Primary definition and official guidance on how Google treats canonical URLs.
- Google Search Console Help — Explains how duplicate groups are handled and how Google may override chosen canonicals.
- Conductor Academy — Clear SEO-focused explanation of canonical tags and their role in duplicate content management.
- Moz Learn SEO — Practical implementation advice and common pitfalls from a well-known SEO reference.
Notes from Callum Bennett.