Duplicate Content.

Duplicate content exists whenever the same or a nearly identical piece of content can be retrieved at more than one URL. A distinction is drawn between internal duplicate content, where several URLs of the same website carry it, and external duplicate content, where the material appears across several domains through copying or syndication. In both cases Google has to decide which version to show — and the version it picks is frequently not the one you would have chosen.

The common internal causes are technical rather than editorial: a site reachable with and without 'www', over HTTP and HTTPS, with and without a trailing slash, through URL parameters for filters or tracking, through print views or session IDs. This kind of unintentional duplication is extremely widespread and almost always fixable with configuration rather than rewriting. It rarely reflects anything wrong with the content itself. External duplication is harder to control, because the other domain is not yours — syndication agreements, scraped copies and partner sites republishing your text all fall into that category.

Contrary to a persistent myth, there is no general 'duplicate content penalty'. The problem is subtler than that. Google groups the variants, picks one as authoritative and effectively discards the rest, which spreads ranking signals across addresses and may leave the wrong URL in the index. The remedies are the canonical tag to name the preferred version, 301 redirects to merge variants permanently, consistent internal linking that always points at the same address, and a single URL convention applied site-wide.

An example: an online shop shows the same product under several category paths and additionally with filter parameters appended. A self-referencing canonical on the clean product URL means Google consolidates every signal on that one address instead of splitting it across dozens of variants. The shop keeps all its navigation paths, the visitor notices nothing, and the product page stops competing with itself. Doing the same for paginated category listings and for sort-order parameters usually removes the bulk of what a crawl report flags on a shop of any size.