Duplicate content explained: what Google does with it and when it actually hurts
Why Google usually just picks one URL for duplicate pages, how redirects, rel-canonical and sitemaps differ as canonical signals, and where the spam policies draw the real line.
Duplicate content sounds like a rule you can break, but Google's own documentation treats it mostly as a practical problem: when several URLs serve the same or very similar content, Google picks one to show and consolidates the signals. The cases where duplication actually hurts are narrower, and they are spelled out in Google's spam policies. This article summarises Google's "consolidate duplicate URLs" documentation and spam policies, both read on 20 November 2026, so you can tell housekeeping from real risk.
What Google does with ordinary duplicates
Printer pages, tracking parameters, HTTP and HTTPS versions, www and non-www - these create duplicate or very similar URLs on almost every site. Google's documentation says that specifying a canonical preference is "generally not critical", but gives two concrete reasons to do it: to consolidate signals for similar or duplicate pages, and to avoid Googlebot spending crawling time on duplicates instead of new or updated pages. So the everyday cost of duplication is diluted signals and wasted crawling, not a penalty.
The consolidation tools, in Google's order of strength
Google lists several ways to indicate a canonical page, and is explicit about their relative weight. A redirect is described as a strong signal that the target should become canonical, and Google says to use it when deprecating a duplicate page. The rel="canonical" link element, added to the head of duplicate pages, can map an unlimited number of duplicates to one canonical page. Sitemaps work differently: every URL in a sitemap is suggested as canonical, and Google decides which pages are duplicates based on content similarity - Google's own "con" is that it must still determine the duplicates for any canonicals you declare this way. Housekeeping matters too: link internally to the canonical URL rather than to duplicates.
When duplication becomes a policy problem
The spam policies draw the line at manipulation. Scraping - taking content from other sites, often automatically, and hosting it to manipulate rankings - is abuse; Google's examples include scraping feeds or search results to generate many pages, including through automated transformations like synonymizing or translating, where little value is provided to users. The same policies discuss syndication: republishing syndicated content is not automatically abuse, but it needs to add value rather than exist to capture rankings. The distinction Google draws is intent and value, not the mere presence of the same words on two URLs.
A short self-check for your site
List the ways one piece of content can be reached: parameters, protocol and host variants, trailing slashes, print views. Pick the canonical URL for each set, add rel="canonical" where variants must stay live, redirect the ones you can retire, and keep internal links pointing at the canonical. If you syndicate your content to other sites, or publish others' content, check the arrangement against the scraping and syndication sections of the spam policies rather than assuming either direction is safe.
What to keep in mind
Canonical signals are hints Google weighs, not commands: the documentation repeatedly frames them as preferences and says Google decides duplicates based on similarity. Crawl-budget concerns matter mostly for large sites; a small site's main reason to consolidate is signal clarity. And this article covers Google's documentation only - other search engines treat these signals differently.