Substantially similar text across multiple URLs. Causes search engines to pick one canonical and ignore the rest, sometimes the wrong way.
Duplicate Content in plain English
Duplicate content rarely triggers a penalty in the manual-action sense — Google has been clear that the default response is to choose one canonical version and exclude the others from search results. The damage is in losing control of which version Google picks. A product page that exists at /products/widget, /collection/blue/widget, and /collection/sale/widget can have Google index the wrong one, losing ranking signal accumulated on the canonical URL.
Standard ecommerce sources of duplicate content: filter and sort parameters that produce indexable URL variants (?sort=price-asc, ?color=red), tag pages that duplicate category content, www-vs-non-www and http-vs-https serving the same content on two URLs, syndicated product descriptions copied from manufacturer feeds (identical across thousands of competing stores), and CMS-generated archive pages (date archives, author archives) that surface the same posts.
The fix toolbox: rel="canonical" to tell Google which URL is the primary (works for parameter variants and intentional duplicates), 301 redirects to consolidate truly-equivalent URLs onto one (best for www/non-www and slug changes), noindex to remove duplicates from index entirely (best for tag archives and filter URLs you don't want competing), and rel="alternate" for paired desktop/mobile URLs.
For syndicated content (same product description across multiple retailers), the canonical solution is rewrite — replace the manufacturer copy with your own. This is the highest-effort but only durable fix; canonical tags can't help when 50 other stores have the same content. Many sites combine partial rewrite with structured-data additions (your own FAQ, your own use-cases) to differentiate.
Why duplicate content matters for ecommerce
Duplicate content silently divides ranking signals across multiple URLs. Imagine a product page that has 200 inbound links, but split across three URL variants (/products/x, /collection-a/x, /collection-b/x) — Google may index the variant with 50 links and ignore the one with 150. Consolidating via canonical or 301 concentrates the signal back onto one URL. For mature stores with years of accumulated parameter variations and tag pages, this consolidation often produces visible ranking improvements within weeks without writing any new content.