Play to Your User’s Deck: Card Sorting for User Research

Card sorting was developed by psychologists to study how people organize and categorize their knowledge. They used it to learn how other people organize topics and relate them to one another. Today it’s used in the same way, but for a vast array of audiences and contexts.

In fact, card sorting is one of those sneaky research tools that seems straightforward, but takes some practice to get the hang of.

In short, card sorting has you write words or phrases on cards and ask participants to organize them into categories. This lets you witness how your users prioritize, organize, and see what they find alike or different — building your understanding of their mental model of the topic at hand.

In general, we use card sorting to:

• Organize information in a way that is intuitive for users
• Segment tasks into larger goals
• Build a common language (more on this later)

What’s so great about card sorting:

• It can be low tech. You can do this with flash cards, post-its, pieces of paper, whatever.
• It can be high tech. There are many resources to conduct card-sorting online, and while these options are always helpful for conducting research at scale, they are more important than ever while we’re social distancing. A few I like are Optimal Sort, User Zoom, and Usability Tools (UXSuite).
• It’s flexible. Once you understand the principles, you can easily modify the method to meet your needs.
• It’s evidence-based. You can point to real, user generated metrics to support your design decisions.
• It’s user-centered. You can empower and witness your audience in action.

We could all benefit from further insight into diverse mental models. For this reason, I always include a lesson on card sorting in Human-Centered Design training workshops to teach how and when to effectively use the technique.

How To Do It

I’ll start with a flexible method, outlining the basics and speaking to variations afterwards. To run an open card sorting session — that is, a card sorting session where the participant comes up with the categories as opposed to you — with one person, in-person (or through a socially distant Skype or Zoom session, where each person can see the other, or by using a tool like Mural) obtain:

1. Flashcards
2. Sharpie
3. Sticky tabs (optional)

Choose a Set of Topics

For example, if you’re conducting a sort for an online store, a “topic” would be a product. I suggest keeping the count under 40.

Be careful with the terminology you use on your topic cards. You might be leading users and spoiling your results. Jakob Nielsen especially cautions against repeating words across cards. For example, don’t list the following as topics: wool boots, wool socks, and wool coat. Users will group them simply because they share the word “wool.” Try using synonyms, such as the following: felt boots, wool socks, and pea coat.

Prepare a Deck of Notecards

Write one topic on the face of each card. If you want to conduct multiple sessions at once, you will need multiple decks of cards. I suggest having them printed to save time. You can also consider adding photos if they are relevant to what you’re testing.

One limitation of card sorting is that it assumes that one topic can only go into one category. My work-around for this problem is to have blank note cards on hand and allow the user to add duplicates if they want.

Test Run the Exercise

Try out your setup with a friend and time it. The duration will vary based on volume of topics and subject matter, and you’ll want to know how long it will take before you start recruiting participants. I recommend keeping these sessions under an hour.

There are a couple ways to cut time from the exercise. You could reduce the number of topics, or eliminate steps 3, 4, and/or 5 in “Steps for the Participant” (below).

Recruit Participants

To make this easier, pay people for their time. They’re giving you valuable information — that should be worth some money!

Introduce the Activity

Explain the purpose of your research. This can be as simple as “I want your help in organizing my online store.” However, be sure to avoid suggesting solutions! By which I mean, don’t say “I want to know if you categorize shirts by sleeve length.” You would be leading the participant.

Then, let them know how long the exercise will take and shuffle the deck.

Steps for the Participant

Ask the participant to do the following while thinking out loud as much as possible.

1. Organize the topics on the cards [B] into groups
2. Add a blank post it to the top of each group [A] and write a category name

The next steps are optional but helpful when you’re thinking about “featured” topics, such as what sections of your website to feature on the homepage, what clothing items to put in the window, what news stories should be sent out via push notification.

3. Order the categories from most to least important.

4. Order the topics on the cards within each category from most to least important.

5. Add a sticky tab to the topics that are most important to the participant.

Document the Results

Take a picture of the layout the participant created. Ask them to explain their reasoning and take notes.

Types of Card Sorting

There are a few common variations for card sorting. We just outlined how to conduct an open card sort. Your card sort can also be closed, combination, or reverse. You can even conduct a card sort as a group activity.

Open Card Sorting

In “open card sorting,” participants create their own categories and give them names. This open ended modality is great for learning. Your users are basically problem solving for you, and sharing their own language with you. It can help you come up with great navigation labels. When I say that you can use card sorting to build a common language, this is the version that does that. Users provide you with their own words for things, which you can use to better communicate with them.

Closed Card Sorting

Alternatively, you can conduct closed card sorting, where you supply the user with predefined categories and ask the user to sort within those groups. Use this when you have fixed categories but want to know how the user would organize set items within them. For example, say you have a kitchen with 3 drawers and 2 cabinets, and you want to know how each of the five roommates who share that kitchen would stock it. You can use closed card sorting to see how the roommates as participants would each organize food and kitchen supplies within those fixed options.

Combination

This pairs elements of open and closed card sorting. You could conduct an open card sort first to identify categories and then use a closed card sort to see how well the category labels work.

Reverse Card Sorting

Build your ideal result and create a diagram with a blank space for each category and topic as you see fit. Present the user with the blank diagram and a deck of topics, and see how accurately they place cards compared with your ideal result.

Conclusion

Card sorting is one of many useful tools to study how individuals organize information so that your systems and solutions will feel more natural to them. If you’re inspired to try one of the above techniques we’d love to hear about it on Twitter, LinkedIn, or in an email.