Duplicate content
Duplicate content is substantial blocks of content that appear at more than one URL, on the same site or across different sites.
Most duplication is accidental: product pages reachable through several categories, tracking parameters, printer-friendly versions, HTTP and HTTPS or www and non-www versions, and syndicated articles.
In more detail
Google does not usually penalise duplication; it picks one version to show and filters the rest. The problem is that it may pick the wrong one, and links and signals get split across copies. Fixes include canonical tags, 301 redirects to one version, consistent internal links and parameter handling. Copying other sites’ content wholesale is a different issue, and adds no value.
Why it matters
Consolidating duplicates focuses ranking signals on the right URL and saves crawl effort. On large stores, it can be the difference between pages ranking and pages being ignored.
A shop's product pages exist at four URLs because of colour and sort parameters. Adding canonicals to the clean product URL and linking only to that version stops search engines splitting signals across copies.