Canonical Tag
A canonical tag is a small snippet of HTML in the head of a web page with which you tell search engines and AI crawlers the preferred, authoritative version of a page. It helps when identical or very similar content is reachable under several addresses, and it consolidates their signals onto a single, official URL.
Why the canonical tag matters
The same page is often reachable via several addresses: with and without "www", with tracking parameters in the URL, or as a print version. To you it looks like one page, to crawlers like several. Without a clear signal, ranking signals such as links spread across competing copies, and no version becomes properly strong. The canonical tag solves this by consolidating all variants onto one main URL. That way you know which page appears in search results and in AI answers, instead of a weaker copy winning by chance. This matters especially for AI visibility: language models should cite exactly the one version that you maintain and keep up to date, not an outdated duplicate URL.
How the canonical tag works
Technically, you place one line in the head (the <head>) of each page: <link rel="canonical" href="https://your-domain.com/main-page">. With it you say: "This is the official address of this content." If the main URL points to itself, this is called a self-referencing canonical – considered clean, basic hygiene. For duplicates, all copies point to the same target URL. Important: the tag is a hint, not a strict instruction. Search engines may follow it or ignore it if other signals contradict it. That is why your internal links, your XML sitemap and any redirects should consistently point to the same address. The more free of contradictions all signals are, the more reliably the correct version is chosen.
Common mistakes
The classic mistake is the contradictory canonical: the page points to a URL that itself redirects elsewhere or no longer exists at all. Equally risky is a canonical pointing to a page that is blocked by robots.txt or set to "noindex" – then you are sending opposing signals. Many content management systems set the tag automatically, but incorrectly, for example pointing every page to the homepage. Mixing upper and lower case, or http and https, in the target URL also causes confusion. So check regularly that every important page points to itself and that filter or parameter URLs cleanly reference their main version. A single typo in the URL can make an entire product category invisible.
Relevance to AI recommendations
AI assistants like ChatGPT, Perplexity or Google AI Overviews access web content and cite sources. If your content exists under several URLs, a model may cite an outdated or incomplete copy – bad for your citability. A cleanly set canonical tag increases the chance that the same maintained address always appears as the source and that its authority is consolidated. This strengthens your brand mention in AI answers and prevents thin duplicates from weakening one another. In combination with structured data and clear internal linking, the canonical tag forms a foundation: it ensures that crawlers and language models can assign your content unambiguously, instead of having to choose between competing versions.
Example
Imagine an online shop for hiking boots. One model is reachable via three addresses: the normal product page, the same page with a colour filter in the URL, and a version with an appended promo code from an email campaign. In terms of content, everything is identical. Without a canonical tag, the three URLs compete and ranking signals fragment. If the shop sets a canonical tag on all three variants pointing to the clean product page, links and authority consolidate there. Search engines and AI assistants then reliably refer to this one official address.
Common questions
Is a canonical tag the same as a 301 redirect?
No. A 301 redirect forcibly sends visitors and crawlers to a different URL, and the old address disappears. A canonical tag keeps all variants reachable and merely recommends which one counts as the main version. Use redirects when a page moves for good, and canonicals for duplicates that exist in parallel.
Does every page need a canonical tag?
Ideally yes, in the form of a self-referencing canonical that points to the page itself. This creates unambiguous relationships and prevents duplicate content, even when no copies currently exist. All that matters is that the tag is set correctly and consistently – a wrong canonical does more harm than none.