A canonical URL is a link element — <link rel="canonical" href="..."> — that tells search engines which version of a page is the preferred, authoritative copy when the same or near-duplicate content is reachable through multiple URLs. It consolidates PageRank, crawl budget, and indexing signals to one URL, preventing duplicate content from splitting your ranking strength.
Every URL variant a page can be reached through — ?utm=..., ?sort=price, http vs https, trailing slash vs not — is a fresh candidate for Google's index. Without a canonical, your ranking strength splinters across those variants. Set one correctly and every backlink, every internal link, every share signal collects on the same URL.
What is a canonical URL?
A canonical URL is the version of a page you want search engines to treat as the source of truth. It's declared in the HTML head using <link rel="canonical" href="https://example.com/preferred-page/">, in the HTTP Link header, or as a signal inferred from your sitemap and internal links.
Google, Bing, and Yahoo jointly introduced the tag in February 2009 to solve one of the messiest problems on the web: the same page being crawlable through dozens of URLs.
- Protocol variants —
http://vshttps:// - Subdomain variants —
www.example.comvsexample.com - Trailing slash —
/pagevs/page/ - Tracking parameters —
?utm_source=twitter,?gclid=xyz - Filter and sort parameters —
?color=blue&sort=price - Uppercase and lowercase paths —
/Productsvs/products - Session IDs —
?sessionid=abc123
Duplicate content is not a manual penalty. But when the same content sits at multiple URLs, Google picks one to index and ignores the rest — and it may not pick the URL you want. A canonical tells Google which version to keep.
Why canonical URLs matter for SEO
Duplicate URLs quietly drain ranking strength. Three reasons every SEO cares about canonicals:
- Signal consolidation. If 20 backlinks point to
/page?utm=x, 30 to/page/, and 10 to/PAGE, they behave like three separate pages. A canonical merges those 60 links into one ranking signal. - Crawl budget. Googlebot has a finite crawl budget per site. Every duplicate URL it crawls is a real URL it doesn't. Canonicals tell it which variants to skip.
- Clean analytics and reporting. When traffic and rankings resolve to one URL, your GA4, Search Console, and rank-tracker data reflect reality instead of fragmenting across variants.
How a canonical URL actually works
The canonical is declared in one of three places, in order of signal strength.
<link rel="canonical" href="https://example.com/product/blue-widget/">
# 2. HTTP response header (for PDFs, images, non-HTML)
Link: <https://example.com/downloads/report.pdf>; rel="canonical"
<!-- 3. Sitemap.xml — implicit canonical signal -->
<url><loc>https://example.com/product/blue-widget/</loc></url>
Self-referencing canonicals
Every indexable page should point its canonical to itself. This looks redundant but it prevents parameterised versions of the same URL from being indexed as duplicates. If someone shares /page?utm=newsletter, the canonical inside that page still says /page/, and Google indexes only /page/.
Cross-domain canonicals
When your content is republished on Medium, Substack, or a partner site, the syndicated copy's canonical points back to your original URL. Google indexes your version, not the copy.
Canonical URL vs other duplicate-content signals
| Signal | Strength | User impact | When to use |
|---|---|---|---|
| rel=canonical | Strong hint | Both URLs remain reachable | Duplicate variants that must stay live |
| 301 redirect | Directive | User forwarded to new URL | Permanent moves, retired URLs |
| noindex meta tag | Directive | URL reachable but hidden from index | Thin variants, filter pages you want out |
| hreflang | Regional hint | Both URLs reachable | Same content, different language/region |
| Sitemap inclusion | Weak hint | None | Reinforces canonical choice |
| Internal links | Weak hint | None | Consistency reinforces canonical |
Real canonical URL examples
Four situations where canonicals earn their keep.
1. E-commerce filter parameters
/products/blue-widget/?color=blue&sort=price
/products/blue-widget/?ref=email
canonical→ /products/blue-widget/
2. HTTPS + www consolidation
https://example.com/ canonical→ https://www.example.com/
3. Syndicated content on Medium
<link rel="canonical" href="https://yoursite.com/blog/post-title/">
4. Paginated series (rel=canonical to self, not page 1)
/blog/page/2/ canonical→ /blog/page/2/
/blog/page/3/ canonical→ /blog/page/3/
Canonical URL vs 301 redirect — which to use
Both consolidate ranking strength. The difference is whether both URLs need to remain reachable.
Use a canonical when
- Both URLs must stay accessible to users
- Filter, sort, or tracking-parameter variants
- Cross-domain syndication (Medium, Substack)
- You're giving Google a preference hint
- The URLs might legitimately be different (e.g. AMP variants)
Use a 301 redirect when
- The old URL is truly retired or merged
- Visitors should never see the old URL
- URL slug changes, domain moves, http to https
- You want a hard, unambiguous signal
- You're consolidating equity permanently
7 canonical URL best practices
- Self-reference every indexable page. Even without duplicates. It prevents accidental URL variants from getting indexed.
- Use absolute URLs, not relative paths.
href="https://example.com/page/"— neverhref="/page/". Absolute is unambiguous. - Point only to indexable, live pages. Canonicals pointing to noindexed, 404, or redirected URLs are ignored and can cause Google to pick a different canonical.
- Keep signals aligned. Canonical, sitemap, internal links, and hreflang should all point to the same URL. Contradictions weaken the signal.
- Match protocol and trailing slash exactly.
https://example.com/page/andhttps://example.com/pageare different URLs. Pick one and stick with it site-wide. - Audit after migrations. CMS moves, template changes, and platform switches routinely break canonicals. Run a full crawl after every significant change.
- Use rel=canonical, not noindex, for duplicate variants. Noindex removes the URL entirely; canonical keeps signals flowing.
Pointing a canonical to a URL that has <meta name="robots" content="noindex"> tells Google two contradictory things: "index this version" and "don't index it." Google ignores the canonical and picks its own. Fix it by choosing one signal per URL.
Common canonical URL mistakes to avoid
- Relative URLs —
href="/page/"without the domain leaves protocol and host ambiguous. - Canonical to a redirect — the canonical target 301s to another URL. Google follows the chain but this weakens the signal. Point directly to the final URL.
- Canonical to a 404 — target no longer exists. Google ignores the tag entirely.
- Multiple canonical tags on one page — Google treats it as a bug and may ignore both.
- Every page canonicalised to the homepage — kills indexing of every inner page. Common bug in bad CMS templates.
- Contradictory hreflang — a Spanish page canonicalising to English cancels the hreflang cluster.
Canonical URLs on a small business site
Most small sites never hand-write a canonical tag, and they do not need to. The platform sets one automatically: WordPress via Yoast or Rank Math, Shopify, Squarespace, and Wix all output a self-referencing canonical on every page by default. The job is to check that the output is correct, not to author it.
Four situations create real duplication on a small site. Tracking parameters on links handed out in ads or email, so the same page exists at a dozen URLs. Print or feed versions of a page. Product or service pages reachable through more than one category path. And a staging site that got indexed because nobody removed the crawl block before launch.
To check yours, open any page, view source, and search for rel="canonical". It should point at the clean, https, trailing-slash-correct version of that same page. If it points at the homepage, or at a URL carrying a session parameter, the plugin is misconfigured and worth fixing before anything else on the technical list. Search Console's Pages report names the affected URLs under "Duplicate, Google chose a different canonical".
Frequently asked questions
Yes. Google consolidates ranking signals from duplicate pages to the canonical URL. Unlike a 301 redirect, both URLs stay reachable but the canonical version accumulates the equity.
A 301 physically redirects visitors to a new URL and is the strongest signal for permanent moves. A canonical is a hint — both URLs remain accessible, and search engines pick the preferred one for the index.
Google treats rel=canonical as a strong hint, not a command. If content, internal links, sitemap, or hreflang contradict the tag, Google may choose a different canonical.
Yes. Cross-domain canonicals are supported and are the standard way to handle syndicated content — republished articles on Medium or Substack should point their canonical back to the original domain.
Yes. Every indexable page should include a self-referencing canonical. It removes ambiguity for crawlers and prevents parameter or tracking-string variants from splitting equity.
Related glossary terms
Sources
