A great, reliable, inexpensive method for discovering patterns in how users would expect to find content or functionality. Card sorting is used to test the taxonomy of data with a group of subjects, usually to help inform the creation of the information architecture, user flow, or menu structure on a project.
A card sorting test is usually conducted by design team members responsible for developing the user experience of the website or product being designed. Key concepts or labels are usually created in advance and subjects are asked to arrange them into what they feel are logical relationships or groupings. The exercise can be done in groups, but better results are often achieved by testing with multiple individuals and consolidating the results into an agreed median.
The output of the exercise can help inform the information architecture of the website or product being designed. The technique is employed when there might be some debate in a project team as to the obvious groupings of items, the number of items to group is large and difficult to predict, or there are too many similarities to draw obvious conclusions.
We use card sorting as an exercise when dealing with projects that have a large sitemap, or large groupings of content whose associations might not be completely obvious. It is also useful to validate any preexisting assumptions around how a target audience might group the content help by the client or ourselves.