Canonical URL (rel=canonical)
A canonical URL is the one page a site designates, via a rel="canonical" link tag in its <head>, as the authoritative version when several URLs serve the same or substantially similar content. Search engines treat the tag as a strong hint, not a strict directive, for which URL to index and rank, and they consolidate the ranking signals of the duplicate URLs onto whichever one it points to.
How the tag works
The tag itself is a single line in the page's <head>:
<link rel="canonical" href="https://example.com/products/widget-123" />
A few practical rules govern how it's used:
- Self-referencing by default. Every page should carry a canonical tag pointing at itself, even with no known duplicate, so the page stays unambiguous if it's ever reached through a tracking parameter or a different URL casing.
- One per page, absolute URL. A page should declare exactly one canonical, using the full HTTPS URL, not a relative path.
- A hint, not a command. Google's own guidance is explicit that it weighs the canonical tag alongside other signals (redirects, internal linking, sitemap entries) and can choose a different URL as canonical if those signals disagree.
Where this shows up in a product catalog
Canonicalization is a same-domain (or cooperating-domain) fix for duplicate URLs, which makes it useful for a specific slice of catalog duplication:
| Situation | Canonical tag helps? |
|---|---|
| Color/size variant URLs that render the same body copy | Yes |
| Faceted-search parameter URLs (sort, filter, pagination) | Yes |
| www vs. non-www, or http vs. https, duplicates | Yes |
| A competitor publishing the same manufacturer description | No — can't point at a domain you don't control |
| The same listing syndicated to a marketplace | Sometimes, if the marketplace honors cross-domain canonicals |
The last row matters for anyone running product content syndication: a marketplace or retailer that republishes your feed may or may not respect a canonical pointing back at your own PDP, and most large marketplaces don't. Duplicate content across companies is a content problem the tag can't solve; duplicate URLs within your own domain are what it's built for.
Canonical tag vs. 301 redirect vs. noindex
Three tools get reached for interchangeably and shouldn't be:
| Tool | Use when | Effect |
|---|---|---|
| 301 redirect | The duplicate URL should stop existing entirely | Users and crawlers are sent to the new URL; the old one stops resolving |
| Canonical tag | Both URLs need to stay live and reachable | Both resolve; ranking signals consolidate on the one named canonical |
| Noindex | A page shouldn't appear in search results at all, but should still be crawlable and pass links | Excluded from the index; not a duplicate-content tool |
A parameter URL that users land on from an active ad campaign is the clearest case for a canonical over a redirect — it has to keep working for the person who clicked it, even though it shouldn't be the version search engines rank.
Frequently asked questions
Is rel=canonical a command or a suggestion to Google?
A hint. Google's documentation describes it as one signal among several (redirects, internal linking, sitemaps) that gets weighed together, and the stated canonical can be overridden if those other signals point elsewhere. It's reliable in the common case but not guaranteed.
Should every page have a canonical tag, even with no known duplicate?
Yes, as a self-referencing canonical that points the page at itself. It isn't fixing a current problem; it removes ambiguity if the page later becomes reachable through a tracking parameter, a different casing, or a trailing slash variant.
Can a canonical tag point to a page on a different domain?
Yes — cross-domain canonicals are valid when the same content is intentionally published in two places, such as a press release syndicated to a partner site. Both sites need to agree on which is canonical; it isn't something you can impose on a domain you don't control.
What's the difference between a canonical tag and a 301 redirect for a duplicate product page?
A redirect sends everyone to the new URL and the old one stops resolving, which is right when the old URL should disappear for good. A canonical tag leaves both URLs live and reachable while pointing ranking signals at one, which is right when the duplicate still needs to work for someone who lands on it directly.