Duplicate URLs are produced constantly and mostly by accident: tracking parameters, faceted filters, session identifiers, pagination, printer views, and the same product reachable through several category paths. Without a canonical, search engines pick a preferred version themselves, and the one they pick is often not the one you would have chosen.
The critical thing to understand is that a canonical is a hint, not an instruction. Search engines weigh it alongside internal linking, sitemap inclusion, redirects and content similarity. If your signals contradict each other (a canonical pointing one way while every internal link points another), the tag will be ignored. The diagnosis is almost always conflicting signals, not a broken tag.
Every page should carry a self-referencing canonical as a baseline. It costs nothing, it resolves parameter-based duplication automatically, and its absence is the most common reason a site's tracking-parameter URLs end up indexed alongside the clean ones.
Formula
<link rel="canonical" href="https://example.com/preferred-url" />
Use an absolute URL, place it in the head, and point it at a page that returns a 200 status. One canonical per page. Multiple tags cause search engines to disregard all of them.
Why Canonical Tag matters
Duplicate URLs split the ranking signals that should be concentrated on one page. Ten versions of a product page each accumulating a fraction of the links and engagement will collectively rank worse than one consolidated version. And on large catalogs they also consume crawl budget that should be spent discovering pages you actually want indexed.
Common mistakes
Canonicalising every page to the homepage
A surprisingly common misconfiguration, usually from a misapplied template. It tells search engines none of your inner pages deserve indexing, and it removes them from results as thoroughly as a mistaken noindex would.
Pointing at a redirected or broken URL
A canonical target that returns a 301 or a 404 is a contradictory signal. The chain gets resolved unpredictably or the tag gets ignored outright, and either way you have lost control of which version ranks.
Contradicting the canonical with internal links
If the tag names one URL while your navigation, sitemap and internal links all point at another, search engines will follow the weight of evidence. The tag isn't doing what you think it's doing.
Using a canonical where a redirect belongs
If the duplicate should not exist at all, redirect it. Canonicals leave the page reachable and crawlable, which is right for parameter variants and wrong for truly retired URLs.
Canonicalising paginated pages to page one
Pages two onward contain different content and different products. Pointing them at page one usually means the items only listed on later pages are never indexed at all.
Where we work on this