A canonical tag (<link rel="canonical" href="URL">) is an HTML element that tells search engines which version of a page is the preferred, authoritative version. It helps consolidate duplicate content signals, preventing SEO dilution when multiple URLs point to the same or very similar content. Google, Bing, and other major search engines respect canonical tags as strong hints for indexing.

A self-referencing canonical is when a page's canonical tag points to its own URL. This is considered a best practice by Google — it clearly signals that this exact URL is the preferred version. Even if a page has no duplicates, including a self-referencing canonical protects against unexpected duplicate content issues (like URL parameters, session IDs, or scraped copies). Every page should ideally have a self-referencing canonical tag unless you intentionally want a different URL indexed.

Having multiple conflicting canonical tags on the same page is a serious SEO issue. When search engines encounter more than one rel="canonical" link, they typically ignore all of them, treating the page as if no canonical was specified. This defeats the purpose of using a canonical tag and can lead to indexing the wrong URL. Always ensure each page has exactly one canonical tag pointing to a single, consistent URL.

Google strongly recommends using absolute URLs in canonical tags (e.g., https://example.com/page rather than /page). Absolute URLs eliminate ambiguity and reduce the risk of misinterpretation by search engines, especially when pages are crawled in different contexts. While relative URLs may work in many cases, absolute URLs are the safer, more explicit choice and are considered an SEO best practice.

Search engines also recognize canonical signals sent via the HTTP Link header (e.g., Link: <https://example.com/page>; rel="canonical"). This is particularly useful for non-HTML resources like PDFs, images, or API responses. If both an HTML canonical tag and an HTTP Link header are present, they should agree — conflicts can confuse search engines and may lead to unpredictable indexing behavior.