Card-sorting is a form of user testing that asks participants to group notecards with site elements/sections written on them to help us gain insight on user expectations for navigation and information architecture (IA). This information can be used to create a user-centered taxonomy for the project.
Why use this exercise?
- Allows us to see how users naturally group content to determine how to intuitively design the navigation structure
- It helps our team determine key content for the project
- Can be used to evaluate effectiveness of current IA when redesigning a site or application
- Can also help with designing new IA if you have already built a list of content to be built
What you’ll need
- While we explain the process for in-person testing here, most card sorting is now done using tools such as OptimalSort. This makes it much easier to test a large number of participants and provides analytics for results.
- Notecards. One for every section or topic found in your website.
- A Sharpee. A pen can be difficult to read from a distance. As the facilitator, you will want to be able to easily read cards without having to read over the participant’s shoulder.
- Table(s). The number of notecards will determine how many tables you need. You will want enough space for the user to be able to play around with groups without feeling confined or having cards fall on the floor.
- A Wall (Optional). If others are observing, it can be useful to tape finished groupings on the wall so they can see from a distance.
- Tape (Optional). See last item.
There are two types of card-sorting techniques: open and closed.
This method is used to see how users naturally group content and label those groups. The key is that they do it without guidance.
- Write all items to be sorted on notecards. Items can contain sections, topics, features, or anything else you want categorized. Keep in mind, though, that more items make tests take longer to complete. It is definitely possible to create user fatigue. Optimal amount is approximately 40-60 items.
- Give the testing participants the notecards and some post-its.
- Ask them to sort or seperate the cards into catagories they make sense to them. Explain that there is no wrong answer, but if they are confused or have trouble deciding we ask that they think aloud and explain why they are having trouble.
- Allow them to work with minimal interruptions. However, remind them to think aloud if you think they are hung up on something.
- Once everything is grouped, have them label each group with some sort of title (and it is ok if the last group is labeled, “I don’t know”)
- Ask them to explain their thoughts behind the grouping of the content.
This method is used when you have pre-determined groups of content or topics. Helps see how users apply the content into those specific groups.
- All of the steps are the same as above, but they will not need post-its since the groups are already labeled. Instead, facilitator should write pre-determined groups on post-its and create columns on the table for participant to sort items into.
How long does this take?
It will depend on number of participants and number of items to sort. Typical tests sort between 40-60 items and run 30-45 minutes. Number of participants can vary greatly. If using a large number of users, is much better to use online tools.
Usability.gov Shows best practices and some of the pros and cons of certain card-sorting techniques.