Canonicalization

Duplicate or near-duplicate content is far more common than most site owners realize — the same product page reachable with and without a trailing slash, with different URL parameters for sorting or session tracking, via both HTTP and HTTPS, or via both a www and non-www domain. Left unresolved, a search engine has to guess which version to treat as canonical, and it can split ranking signals like backlinks across multiple URL variants instead of consolidating them onto one.

The primary tool for controlling this is the rel="canonical" link tag, placed in a page's HTML to point to the preferred version, along with consistent internal linking and correctly configured redirects. Canonical tags are a strong hint, not a directive Google is contractually bound to obey — if other signals (internal links, sitemaps, actual user-facing URLs) consistently point somewhere else, Google can choose a different page as canonical regardless of what the tag says, which is why canonicalization has to be consistent across a site's whole technical setup, not just declared in one tag.

Getting this wrong is a common cause of unexplained ranking volatility: a page that should rank well split across three URL variants often ranks worse than the same content consolidated onto one clean URL, even though nothing about the content changed.