UX glossary
Card Sorting
Card sorting is a research method in which people sort terms written on cards into groups. It shows you how they mentally organize content, so you can plan navigation that makes sense to them.
Imagine a local sports club rebuilding its website. The board has spent weeks debating whether “Fees” belongs under “Membership” or “Services”, and whether parents would look for training times under “Teams” or “Calendar”. Everyone is sure they’re right. Instead of arguing any longer, the team asks the people who will actually use the site, with a stack of cards.
Explained simply
Card sorting is a research method in which participants sort terms written on cards into groups that make sense to them. Each card stands for a piece of content, such as a page or a feature. You know the principle from organizing a kitchen: some people sort by type of utensil, others by what they need for baking or cooking. Card sorting shows you the logic people carry in their heads, in other words their mental model.
The results help you build an information architecture: the structure and labeling of content, shaped by what users expect rather than by the company’s org chart.
Open, closed or hybrid
In open card sorting, participants create their own groups and name them. This is the most common variant and works well when you don’t have a structure yet. In closed card sorting, you provide the categories and participants only assign the cards. To validate a finished structure, though, the Nielsen Norman Group generally recommends tree testing instead. The hybrid form mixes both: some categories are given, and participants can add their own.
You can also run a card sort moderated or unmoderated. In a moderated session, someone sits in and asks why a card ends up in a particular group. Unmoderated sessions run through an online tool like Optimal Workshop, which gives you more data for quantitative analysis.
An example
Back to the sports club. Before, the site mirrors the club’s committees: “Board”, “Office”, “Youth Committee”. Anyone looking for under-10 training times has to click through three menus. An open card sort with 15 members and parents reveals a clear pattern: almost everyone puts training times, the gym schedule and the team contact on one pile and calls it something like “My team”. After the redesign, each team has its own page with exactly that content, and the committees move to an “About the club” section.
Common misconceptions
Card sorting gives you ideas for a structure, not a finished navigation. The group names participants come up with are suggestions, not final menu labels. And because different people often group the same content differently, you still need your own judgment and a second test that checks the structure. You can read more in Card Sorting Basics: Efficient Information Architecture for Beginners.
In practice
- Keep the deck manageable. The Nielsen Norman Group recommends 30 to 50 cards. With more, participants get tired and dump lots of cards into a big “Miscellaneous” pile.
- Recruit enough participants. For qualitative insights, at least 15 people is a sensible minimum. For solid quantitative results, plan for 30 to 50.
- Avoid giveaway words. When several cards share the same word, people tend to group by that word instead of by meaning. Word your cards differently and try them out with one or two people first.
Related terms
Information Architecture, Tree Testing, Mental Model, User Research