Card Sort

What can you learn from running a card sort?

7 minute read time

If you’re new to UX or just new to card sorting you may be wondering what information you could get out of a good card sort.

If you're a UX professional you may be wondering if you're using card sorts to their full advantage or if there is anything else you could learn from a card sort.

This is a great question becuase you'll likely be surprised at the answer.

How you answer this question will impact the type of card sort you run, how many participants you recruit and how you analyze the data.

In other words, figuring out what you want to learn will impact what you do learn.

Kind of meta.

Let’s walk through some of the reasons for running a card sort and what you could learn in each scenario.

Create an effective information architecture and website navigation

This is probably the first thing you think about as the reason for running a card sort.

It’s the outcome most talked about when suggesting a card sort. And for good reason.

Knowing how your customers think about your information, how they label it and how they would group it will be enormously helpful in creating an effective information architecture (IA).

Using that IA to build a website navigation based on how your customers think about your content is a solid strategy.

For this type of study you’ll want to run an open card sort with at least 30 participants. Let your audience tell you how they would group and label your content.

When looking at the data, in this type of situation, you’ll want to look and see how many different classification schemes people came up with, how many different groups they created, and what they put into those groups.

You’ll want to look at both the qualitative and the quantitative results.

If you’re in the initial phases of creating a product or a website, a card sort like this can inform the designs and allow you to create a logical and user-friendly organization right from the beginning. This will save you a lot of time down the road.

If you already have an existing website, consider running a tree test right before this type of study. With the tree test results in hand this type of card sort can give you insights into how to organize and label your content in a way that will make sense to your audience.

Determine menu labels

If you have a website and you’ve already figured out how to organize your information in a way that makes sense to your audience you can use a card sort for quick insights into what labels to use to describe each group.

To run a study like this, you’ll want your card sort to be an open sort. You don’t need a lot of accuracy here or statistical significance, as your main focus is on the words they use for the different groups.

You also don’t need that many participants. 5-10 is probably going to be enough. Just keep an eye on the results and once you see repeating patterings in the labeling of categories you can end the study.

You can use the results from this type of study to confidently label your website navigation and menus.

Learn about a topic

If you are a freelance UX professional or new to a company, especially if that company has a content-rich website, it can be a lot to take in at first. You can’t possibly know the domain in depth at first.

Running a card sort early on is a great way to learn about the domain, particularly from your customers point of view. You’ll see what types of groups they form and how they describe the content.

You’ll want to set up the cards so that each card represents a broad topic (not a specific page or blog post). You don’t need a lot of accuracy here or statistical significance so you don’t need to be super concerned about getting a ton of participants. 10-20 participants is probably going to be plenty.

A card sort done in this way will give you a lot of information about the domain and especially about how your audience views the information and thinks about it. What words they use and what makes sense to them.

You’ll better understand the relationships and nuances in your content.

You’re not going to use this to create an information architecture, but it’s going to give you a solid base to work from and will help you make progress much quicker.

Test your ideas and assumptions

You are not your users. Any good UX professional knows this. However, it’s also true that sometimes you do have good ideas.

Before you go very far down the path of acting on those good ideas, test them with your users.

Run a quick test to make sure you’re on track. You don’t need statistical significance here so 5 to 10 participants is plenty for this type of test.

You don’t even have to make sure to include all the information on your site, just the area in question or enough of it to test your assumptions.

Remember, this is just a quick test that you can put together and get some answers all in an afternoon. A card sort ran this way will give you the information you need to move forward.

Collaborate with and get team buy-in

Research has shown that organizations that involve more of the company are more profitable. Collaboration is an important part of ensuring your user research is successful.

Typically you don’t run a card sort just to get team buy-in. Instead you’ll probably have another reason for running a card sort, but this might be a secondary reason.

At Empathetic we want you to involve your whole team. That’s why we haven’t set any limits on the number of team members you can invite into your account. We want you to collaborate and make your product more successful.

Involve team members in your card sorts. They will begin to better understand how your audience thinks about your content, realize that other people think differently (if they don't already), and generally see the value of user research.

Dive into one area

You may run into one small section of your website that is not working.

Or you may simply be having a hard time organizing, structuring, prioritizing or labeling the content.

Or maybe you’re not sure which solution that you’ve come up with would be best.

In each of these scenarios a card sort can help.

For a situation like this you’ll want your card sort to be a little more robust. Most likely you’ll use this data to form a small part of your information architecture.

You’ll want to create between 30 to 50 cards, each one either for a specific page in that area or a specific topic (depending on how many pages there are).

You’ll also need at least 30 participants.

Keep this type of test as an open card sort so that participants can create their own categories and groupings without any bias from you.

When looking at the data, in this type of situation, you’ll want to look and see how many different classification schemes people came up with, how many different groups they created, and what they put into those groups.

You’ll want to look at both the qualitative and the quantitative results.

Compare your audiences, guide marketing and product direction

If you have an audience that seems to have some big differences between them, you may need to run a card test to compare how much and where those differences show up.

The differences may be due to location, life status, or simply how they’re using your product. The bottom line is that they think about and use your information differently.

For example, you may have an audience that are physicians and another that are patients. The two might think about your content and labels in very different ways.

It’s worthwhile to run a card sort testing these different groups. When doing this though, it’s important to note, that you should recruit 30-50 participants per user segment.

So, if there seem to be three different groups, you’ll want to recruit upwards of 90 participants.

When you look at the results you’re going to be interested in both the qualitative and the quantitative data, looking specifically at similarities and differences between your audience groups.

You can use the results in your marketing efforts allowing you to target specific audiences with the way you label and organize your information. This will also give you a good idea of which products and offerings to group together.

Back up your recommendation with data

There are times when you are dialed into your audience. You understand them. And all you really need is a quick test of your ideas or assumptions.

You could validate you are on the right track and be moving forward this afternoon.

Except that you know that a small study isn’t going to convince those who need convincing.

In her book “Card Sorting”, Donna Spencer shares her experience of what type of card sort to create when facing these challenges. She says, “In this situation, I think about how many participants I would normally involve to get the answers I want, and then I double it. I also prepare up front to do both qualitative and statistical analyses, even though I may feel like I need to do only qualitative analysis.”

Using data and results to back up what you know to be true is sometimes the smartest thing you can do.

My suggestion for you

There’s a lot you can learn from card sorts. They don’t always need to be statistically significant or take a lot of time.

It all depends on what situation you’re in and what questions you have.

Use card sorts to:

  • Create an effective navigation
  • Learn general information about your audience, your content, or the way your audience thinks about your content.
  • Validate your ideas and assumptions to make sure you’re on the right track.
  • Get team buy-in
  • Guide marketing
  • Determine menu labels
  • Dive into one specific area. Get a lot of information and insights into how your audience thinks that information should be organized, grouped, and labeled.
  • Get really clear about the differences and similarities between your audience segments in terms of how they think about your content.
  • Use data to back up your recommendations.

You can learn a lot from a card sort. They can provide a lot of insight and help to you in your projects.

Now that you know what you can learn, my recommendation is that before you start your study determine what exactly you hope to learn.

It will inform the type of card sort you run, how many participants you recruit, how you set up the cards, and even how you evaluate the results.