Card sorting is a UX research method where you ask real users to organize a set of topics, features, or content items, literally written on cards, physical or digital, into categories that make sense to them. It's one of the most direct ways to understand how people naturally group and label information, rather than relying on how a product team assumes they will. There are two main flavors: open card sorting, where participants create and name their own categories from scratch, and closed card sorting, where they sort items into categories you've already defined. Each answers a slightly different question, open sorting reveals how people think; closed sorting tells you whether your existing structure actually makes sense to them.
The reason this method has stuck around for decades, long before anyone called it "UX," is that it solves a problem teams consistently get wrong on their own: internal logic rarely matches user logic. A product team might organize a website by department or by how the backend is structured, while users are looking for things based on their own goals or vocabulary. Card sorting exposes that mismatch early, before it gets baked into a navigation menu or an information architecture that then has to be lived with for years.
For a business, the value shows up in a fairly direct way through reduced friction. When a site's navigation or menu structure mirrors how users actually categorize things in their heads, people find what they're looking for faster, bounce less, and generally trust that the product understands them. That has downstream effects on conversion, support tickets, and overall satisfaction that are easy to underestimate until you see what happens without it: users getting lost, calling support for things that should've been self-explanatory, or abandoning a task because the right option was three unexpected clicks away.
It also saves a company from a much costlier mistake restructuring a live product after launch because the information architecture didn't hold up. Doing a card sort during discovery or research phases is inexpensive and fast compared to redesigning navigation once thousands of users have already formed habits around a confusing structure. For enterprise clients with large, content-heavy platforms,banking portals, retail catalogs, internal tools, this upfront research often prevents exactly the kind of expensive rework that shows up much later as a "why is nobody using this feature" conversation.
