An Introduction to Card Sorting

Out of all the types of research available to improve your web site, card sorting has to be one of the most enjoyable.

What is a Card Sort?

Card sorting is a unique method of observing the ways in which people group information together. It uses actual index cards with content printed on each one. You observe as participants create meaning by organizing the cards.

There are a few basic steps to take in the most common type of card sort:

  1. Write a word or phrase on every card that represents example web site content. Each card should be approximately the same level in the content hierarchy (so you would not have some cards that are blog post titles and others that are main navigation items).
  2. Explain in general terms why you’re doing this activity and ask the participant to group similar cards together in a way that makes the most sense. You can mention that this will help inform the structure of your web project, but be careful not to place heavy emphasis the web aspect. This will help them focus on the content itself without getting caught up in web site layout.
  3. Make sure you get permission to record the audio and then start a recording. If you can record the video in a way that doesn’t feel intrusive, then go for it. In either case, have a laptop or notepad handy to take notes.
  4. As they begin sorting, listen and watch. Answer any questions about content as they come up because you want to make sure each participant clearly understands the content on a card. After you’ve run a few sorts, you’ll realize which cards are worded poorly and need to be reworked.
  5. Ask the participant to create a label and description for each group. Have some sticky notes and a pen available for this.
  6. Thank them and then record your results in a spreadsheet.

That’s it! This style of card sort is called “open” because you’re not determining the group titles for them; they can choose whatever labels they deem appropriate. In a “closed” card sort, you’d be giving them the main categories and they would place cards under the appropriate heading.

Not a silver bullet

A common misconception about the value of card sorts is that the data should be collected, compiled, and then the new site structure created to mimic the results. The data you collect should be part of your overall research methods and considering everything together will help shape your final structure.

The value of card sorting lies in observing the process as much as it does in collecting the data. The process informs the way you view the data. In order to capitalize on this, it’s important to explain the activity to participants and then ask them to verbalize their thought processes as they go along.

The inherent problems with this, though, are that

a) some people find it extremely difficult to verbalize their thoughts and

b) results can be skewed once participants become aware that someone is recording and analyzing every word they speak.

Any observation of behavior is capable of distorting the behavior itself, but one way to potentially lessen this burden is the group card sort.

How group sorts can help

A group card sort is one in which you have two or three people working together to complete the task.

I was initially hesitant to try this because it sounded similar to focus groups, which tend to suffer from groupthink and social influence. In the end, though, what I found is that it’s really quite different.

When two or three people are tasked with the job of sorting information and labeling it, they tend to get pretty into it. When they’re interacting, discussing, arguing, and joking, you and your recording device become less of an imposing presence.

I’ve conducted sorts where the results from the group sort ended up being messier than individuals’ results, but the conversation I recorded gave me insight I wouldn’t have otherwise had. With difficult-to-place content, an individual might decide internally to make and stick to an initial decision but a pair of participants will end up verbalizing the pros and cons in order to reach consensus.

This technique can give you insight, challenge assumptions, provide valuable discussion points in meetings, and inform your overall strategy. If you’re unfamiliar, I hope that this gives you a reason to consider trying it out!

Free book: Jump Start HTML5 Basics

Grab a free copy of one our latest ebooks! Packed with hints and tips on HTML5's most powerful new features.

  • awasson

    Nice to see some new discussion about Card Sorting. It’s generally the first step in our usability testing suite and can provide valuable insight.

    I’m of mixed opinions about how useful the resulting data is and I think it really depends on the demographics of the site’s audience. We did card sorting for a bank website and I think that was useful because of the wide array of age groups, backgrounds, etc… who visit the site. Card sorting allowed us to hone in on consistencies in the results and choose a flow of information that resonated (somewhat) with most of the subjects. In the case of narrower audiences, I wonder how useful it is. For instance for a website that caters to engineers or scientists, you may very well get scattered results because there may be a common information flow but the subjects can over think their role and provide artificial results which echoes your inherent problem ‘b’.

    I still find it useful and ever card sort in my experience has been different. We’ve even done remote card sorting for international websites.

    I thought Jakob Nielson had some really great information about card sorting. Especially his noted about how many subjects to include and his thoughts on diminishing results: http://www.useit.com/alertbox/20040719.html

  • http://twitter.com/usabiliTEST Sergey S, CUA

    Here’s a budget-friendly online Card Sorting tool with built-in data analytics: http://www.usabilitest.com/CardSorting