Card Sorting: The Ultimate Guide (in 2025)

Whenever you’re working in user experience (UX) design, you’ve got to understand user expectations if you’re going to be able to group information effectively, and card sorting is a handy—and relatively simple—way to engage participants so they categorize topics in a way they find logical. This approach is an effective one to help you uncover user preferences for content arrangement to help create a coherent and user-friendly structure for websites or apps.

You’ve likely come across the concept of card sorting—whether or not you’re in UX—and researchers use the popular, low-tech research technique to organize data sets. What it’s especially useful for are information architecture, menu structures, workflows, and website navigation. You may find it easy to run a card sort, but—be that as it may—there’s an art to it and things that you’ll need to get right; let’s get into the do’s and don’ts of it now.

Table of contents

What Is Card Sorting in UX Research?

When you do a card sort, you get users involved in organizing information and they’ll sort labeled cards into groups from what they understand about the relevant things. The purpose behind the process is it shows how users expect to find information on a website, and the insights you get from card sorting are—potentially—very powerful tools you can use to create intuitive and user-friendly website structures as a designer, structures that run in line with the users' mental models for seamless and efficient navigation.

Here’s a little example of how a card sorting session might go. Picture a group of user participants and you’re the researcher for a grocery store brand that’s creating a new website. You want to see how customers are going to approach the store—namely, what they’d put various items under in terms of categories. So, you deal out cards to the participants, cards with things like “Dairy,” “Bread,” “Fruits,” “Toothpaste,” “Deodorant,” “Eggs,” “Insect Repellent” and “Garbage Bags.” Now, the users put these into categories, like “Dairy” and “Bread” under “Daily Essentials,” “Toothpaste” and “Deodorant” under “Toiletries,” and “Garbage Bags” and “Insect Repellent” under “Sundries.” But some participants might consider, say, toothpaste a daily essential, for example. Even though they wouldn’t need to buy a whole tube of it more than once every two weeks, they might still term it a daily essential.

Welcome to the “magic” of card sorting, where logic can be subjective! In the above example, you’d be able to harvest—no pun intended—these insights to organize the website's menu and categories in the way that works best, and make it easy for shoppers to find what they need.

Watch as Design Consultant Donna Spencer provides an excellent example of a card sort in her IxDF Master Class, “How To Use Card Sorting For Better Information Architecture.”

Show Hide video transcript

00:00:00 --> 00:00:34

The first thing you want to do is just *poke around*. If you've got it in a big spreadsheet or something, you might print it out on the wall, you sort of look at: What *happened* here? What did people *do*? What's kind of *interesting*? What *ways* did they do things? So, just an initial – 'What is this?' You might look through and say, 'Okay, Participant 1 – how did they do that? Oh, look – they sorted the groceries into the way it's sorted in a physical store.'
00:00:34 --> 00:01:03

Or, 'They sorted the groceries according to taste – interesting. It's just interesting.' Then you might go through and go, 'What is interesting about those groupings? Ah, wow – okay, this is different to that. I didn't expect to see people put those things together.' And then look at like different participants and say, 'What did people do that is *different*? Did everybody do exactly the same thing? Did they do all different things?
00:01:03 --> 00:01:26

Did they do some similar things and some differents? What is interesting about that?' – What kind of variations did you notice? And then, also, *what words* did people put on the cards? What language and labeling came out of this activity is super interesting. So... you're kind of just burying yourself in data initially; you're not trying to get to a result really quickly.

Why Do We Use Card Sorting?

Card sorting helps you understand how people think—and that’s not just about items but about ideas and concepts, too—and when you’ve got a good idea of how users think about encountering a digital design, for instance, you’ll be able to arrange content in a way that makes sense to them when you are organizing a website's structure or navigation. A card sort can show how someone’s background and personal experience affect their thoughts and the way they envision things, and give you a handy vantage point to accommodate how they see their world and fine-tune your interface design to their way of seeing and grouping things.

When you’ve done a successful card sort, you’ll learn the following, too:

What goes together and why
Different ideas and methods of organizing content
If people think similarly or differently about something or things

While people use card sorting more often to create IAs, website navigation, and menu structures, the strength of card sorting is how it translates across the board, and you can also use it for internal communication and politics.

In this clip, Donna Spencer explains one of the less obvious uses of card sorting and what it can tell us.

Show Hide video transcript

00:00:00 --> 00:00:30

One of the not-obvious uses of card sorting is to show people who have a really good, deep technical understanding of a domain – to show them that people who don't know this yet think really, really different about it. And that can actually be one of the big findings – it's: You think like this because you *know* stuff. The people who are using it think like this because *they don't*.
00:00:30 --> 00:00:54

If we arrange it your way, they won't be able to comprehend it. Or even different teams in an organization – you think that way; they think that way. There isn't a right way to do this. So, it's not only about putting content together to make a hierarchy for an IA. There are deeper and more useful aspects for this process.

You’ll find card sorting effective for enhancing your IA, sure, but don’t expect it to actually create the IA for you—and that’s because card sorting is qualitative and exploratory, and therefore often yields results that are inconsistent and renders it unreliable to create an information architecture from scratch. On a similar note, it can’t tell you whether your IA works, and even though the card sort will inform your IA, you won’t know if it’s any good. Donna Spencer suggests you use a tree test instead.

What’s more, card sorting isn’t suitable when you’re after the following:

Definitive, black and white answers
Quantitative, statistically valid data

Advantages and Disadvantages of Card Sorting

Advantages of Card Sorting

Simplicity: You’ll find that very few techniques are as easy as just handing someone a deck of cards and then asking them to sort them out—it’s so direct and basic.
Cost-effectiveness: You’re talking about some plain card and either printer ink or pen ink. You may also use some sticky notes or some Sellotape.
Speed: You can run this exercise quickly—and as many times as you have to until you get the data you need out of it.
User-centric: Card sorting focuses on user input—and that’s crucial for product success—and since it values user perspectives over assumptions, it can save many a design headache later.
Familiarity: It’s a technique that’s been around for hundreds of years and people—including your users—can “get” it easily and not need much explaining on how to do it.
Insightful: It isn’t the most in-depth approach, to be sure; still, card sorting does manage to provide valuable insights as regards how users categorize information—and they’re insights you can leverage well.

Disadvantages of Card Sorting

Task ignorance: Card sorting doesn't always consider exactly what you’re going to do with the insights that come from it—the practical application of data, in other words—so it pays to be cautious about how you’re going to translate those results into your product.

Inconsistency: Card sorting results can vary thanks to all those little individual differences in perception and categorization; you know, the little quirks and idiosyncrasies people have about how they see things in their world—or, maybe more apt here, their worlds, plural.

Time-consuming analysis: It’s quick to perform a card sort, sure, but to analyze card sorting results—that’s another matter, and it can translate to long treatment times, especially when you’ve got complex data sets at hand.

Surface-level insights: Some fundamental issues or context may be beyond the technique’s scope, since it can often only go so deep, and the top-level view a card sort might give you can have deeper, more intricate matters under it—matters which you might well need to look more closely into.

Types of Card Sorting

Card sorting is a versatile UX research technique and there are different types of it that can suit various needs, with each one offering unique insights—so, without further ado, let's explore these card sorting methods in detail, plus some examples.

1. Open Card Sorting

In this one, participants sort cards into categories which they make themselves, and that’s great if you want to know how users naturally group information and without preconceived categories of yours (or someone else’s) that may guide them away from how they’d see things otherwise.

A diagram showing an open card sort process. — Say you get a stack of cards and you sort cards into groups and label these groups; that’s an open card sort.
© Interaction Design Foundation, CC BY-SA 4.0

2. Closed Card Sorting

Closed card sorting finds participants sorting cards into predefined categories, and it’s a method you’ll find useful whenever you want to test specific groupings or if you’ve got a basic structure.

A diagram showing a closed card sort's process. — You get a stack of cards, and you’ve got to sort them based on groups the researchers have created—that’s a closed card sort.
© Interaction Design Foundation, CC BY-SA 4.0

3. Hybrid Card Sorting

Hybrid card sorting is a blend of open and closed card sorting where participants sort cards into provided categories but they can create new ones of their own, too, which means a good deal of flexibility for them and deeper insights into user preferences for you.

Example: Picture a website you might be building for a travel operator or brand that handles tourist resorts or the like. You’re doing a card sort where users can sort destinations and activities into different categories (e.g., Beaches, Mountains, and Cities). The beauty of this card sort, though, is that it’s a hybrid one and users aren’t restricted to the categories you’ve given them—in fact, they can make up totally new ones like “Adventure Travel,” “Historic Trails,” “Family-friendly” (the list can go on) and, from this, you’ll be able to get more nuanced insights.

4. Reverse Card Sorting

Reverse card sorting—or, tree testing—is a card sort that (you guessed it) works in reverse and users have got to work backwards. So, instead of actually sorting cards, they deduce where a card should go inside of an existing structure, and that’s a great way to test how intuitive a site’s navigation is. Then, at the end of a session, you can evaluate the results—compare how many users placed the item in the “right” category.

Example: Let’s imagine in a website that’s all about cookery (recipes and maybe some products and services, too) that you hand out a specific recipe card to users (say, “Vegetarian Lasagna”) and they’ve got to find where it fits inside of a pre-made set of categories. That way, you can test if the organization of the site really does make sense to each user.

5. Modified-Delphi Card Sorting

Modified-Delphi card sorting is a method that evolves with each participant, and it begins with the first participant’s conducting a complete card sort where they organize and arrange the items in a way that makes sense to them. That forms the baseline model, and then subsequent participants refine this model—each one builds on the previous person's work, and so the process continues.

Example: We’ve got a team working on designing a health app, and they categorize features like “Exercise Tracking,” “Diet Plans,” and “Mental Wellness.” Each one of the participants first sorts these features on their own; then, when they’re done, they discuss their choices collectively and that way can reach a consensus.

6. Remote Card Sorting

Remote card sorting—just as it sounds—lets participants sort cards through online tools, and it’s a handy and convenient way to reach a broader and more diverse audience.

Example: An e-commerce brand’s site might use an online tool to have users sort product categories like Electronics, Home Goods, and Fashion, and participants do the task on their computers instead of having to attend a formal session.

Tree Testing vs Card Sorting: Which One Should You Prefer?

First off, here’s a quick overview of both techniques:

In tree testing, users navigate through a simplified text version of the site's structure to find items—it’s something that reveals how easily they can locate information.
Card sorting helps you understand how users categorize information, and, there, participants organize topics or items into groups that make sense to them.

William Hudson, UX Designer and Author, discusses tree testing in more detail here, so watch this video to understand more about getting started with tree testing.

Show Hide video transcript

00:00:00 --> 00:00:33

We're going to be talking about getting started with tree testing. Tree testing *simulates* a hierarchical navigation structure. So, those are just the layers of menu that you expect your users to wade through in a typical application or website or mobile app, for example. And it's purely text-based; so, this is simply how Optimal Workshop, and their solution is called "Treejacks".
00:00:33 --> 00:01:00

Here, for example, it's Home phone, My Account, Help and Support, Internet and so on. And this is how – you can change the colors. I believe you might even be able to change the fonts, but that's all you can do. There is nothing else that you can do in terms of the visual presentation. And, again, that's one of the strengths – that you're not distracting people with unnecessarily complex visual designs – you're looking purely at the navigation.
00:01:00 --> 00:01:31

So, starting from the home page, each level of navigation is shown to participants in turn. And so, that's the top level. So, being given a task, I now have to click on one of those *six* items. Most tree testing services ask you to upload a spreadsheet with your hierarchy. And I'll give you a full example of a project where we've done exactly that, and you can see the format; it's very straightforward. It's just done hierarchically in a spreadsheet. Separately, you provide a list of tasks. And, for a lot of these testing services,
00:01:31 --> 00:02:02

you provide a correct response or a set of correct responses. I did work on one site where we ended up giving people all kinds of correct responses because they could get to the products that were being sold a number of different ways and they were all valid; but, of course, that ended up contributing to users' confusion, as well, unfortunately. But, for basic tree testing, no visual design is needed. So, we just upload the spreadsheet of the menu hierarchy and off we go.
00:02:02 --> 00:02:33

As the researcher, you can specify how many of the tasks each participant must do and whether tasks can be skipped. And if you're really keen-eyed, you might see in the top right-hand corner of this screenshot on the right that it says "Skip this task" – which is user-definable, researcher-definable. You can decide whether your participants are able to skip tasks as they're going along. Optimal Workshop has some recommendations for their Treejack tool. They suggest a maximum of 10 tasks, and of course that can vary.
00:02:33 --> 00:03:02

If they're really complicated tasks, you might want to reduce that number. If they're really simple tasks and you've got a really short tree navigation hierarchy, then you might increase that number. Do write tasks that test the part of the website you want to improve. So, if you're working in a particular area, then focus on that – obviously. Write tasks as hypothetical scenarios based on your typical visitors. So, you want to make this sound like something realistic from their perspective.
00:03:02 --> 00:03:33

And don't mimic the language that's used in your navigation. Use different language than the labels on your tree. Otherwise, it just becomes a word-matching exercise and you will have achieved nothing; you will end up going to all this effort and expense and you will get results that are of no value to you because all people did was to look at what you wrote and match it to the text on the screen. So, don't do that, as you must use common words that are *natural* for your tasks without mirroring the navigation terms.
00:03:33 --> 00:04:00

This is a simple summary result for the tree testing. This one again is from Optimal Workshop, but most of the tools have a very similar kind of summary. In this case, we've got a pie chart and the successes are shown in green. And there is the idea of a direct success and an indirect success. The direct success is when people have gone straight to the correct answer without any variation.
00:04:00 --> 00:04:32

Indirect is where they've gone down one path in your navigation hierarchy and had to backtrack. So, if they still end up in the right place, that's still referred to as a success – but it's an *indirect* success. And you can analyze these figures separately if you wish. If you're getting a lot of indirect success, it means people are getting there in the end but they're confused. And, of course, in *real life*, if you're talking about an actual website or mobile app, you may see a lot of people simply giving up rather than persevering with the study, which is what they're doing when you're doing the research.
00:04:32 --> 00:05:03

The red is fail, and that too is direct and indirect. And in both, in the same situation, the same concepts apply that if they've wandered around a bit to fail, it's indirect, but if they've actually gone straight to an incorrect solution without backtracking, then it's called a direct failure. And then, finally, at the bottom of this chart, you have the skips, which hopefully will be zero, and certainly will be zero if you said that users are not allowed to skip tasks.

If you’re going to pick between these methods, it’ll depend on your project's stage and goals for which one’s best. If your current IA seems to be causing problems and you need to assess how effective it is, then do start with tree testing, as it’ll show you where users struggle to find information; what’s more, it’ll give you a clear picture of the areas that need improvement.

After you’ve grasped what’s wrong with your current IA through tree testing, then it’s time to go ahead with card sorting so you can get a helping hand at understanding how users think you should organize your content. An open card sort is what lets you capture your users' natural categorizations and terminologies—that’s a crucial ingredient for making an intuitive and user-friendly IA.

Last—but not least—do another tree test after you’ve reworked your IA based on card sorting insights. It’ll help you compare the new IA against the old one using similar tasks, and it’s a comparison that will highlight just how effective your changes actually are—plus, it’ll help you get your updated IA more in line with users’ expectations and behaviors.

What Tools Do You Need to Conduct a Card Sort?

Card sorting doesn’t call for any fancy tools, and you can just use basic office supplies to sort with, say, index cards and markers. With that said, though, there are several digital tools you’ll find that can enhance the research process, and below we show six online tools that can help you streamline your card-sorting process.

Tools to use for card sorting. — © Interaction Design Foundation, CC BY-SA 4.0

Considerations for Choosing a Card Sorting Tool

You’ve got to consider several factors while you’re picking the right card-sorting tool for your project, so let’s see four important ones:

Study size: Tools vary in their capacity to handle different study sizes, so, for larger studies, go for tools that have got robust data-handling capabilities.

Study format: Decide if your study is to be a remote or an in-person one, and that’s because tools for remote sorting have got features like online collaboration, while in-person tools might focus more on physical interaction.

Analysis needs: Some tools come equipped with advanced analytics, which are ideal for detailed data interpretation—so, work out how deeply you’ll need the analysis to go.

Budget constraints: Be sure to balance the need for features with what you’re prepared to pay for.

How to Run a Card Sort?

Seven steps to run a card sort. — © Interaction Design Foundation, CC BY-SA 4.0

1. Set Clear Objectives

Just like any other research methodology, you’ve got to work out exactly what it is that you want to learn and then—once you’ve done that—consider whether card sorting is actually the best way to learn it.

2. Select the Right Tools

You’ll need some cards, and plain index cards or Post-it notes work well for an in-person card sort. You can use anywhere from a handful to about 100—sometimes more, but it’ll depend on the project—but do try to limit the number of cards to 50 if you’re going to do a virtual card sort.

A person arranging a set of different colored post-it notes in the manner of a card sort. — Index cards or Post-it notes work well for card sorting.
© UX Indonesia, Unsplash License

You can find several card sorting tools online—such as OptimalSort, but applications like Trello, Miro, or GitHub can work well, too. If you use a card sorting tool, it’ll offer you a great benefit, and that’s immediate availability for analysis on the computer and—what’s more and a happy little extra—you’ll be spared the tedious task of transferring results from physical cards to a spreadsheet.

3. Write Clear Instructions for Effective Participation

You’ve got to prepare your instructions before your card sort begins—and that’s no matter if it’s to be done physically or digitally—and they’ve got to be clear so your participants know what to do.

Think about the questions people may ask, and, if you’re conducting your card sort online, be sure to include the answers to those questions in your instructions so you’ll pre-empt them. If you’re going to be doing an in-person card sort, prepare your answers in advance.

Participants may ask questions like:

“This card fits with more than one group; can I put it in multiple places?”
“I don’t understand this card; can I exclude it?”
“There seems to be some content missing; can I add a card?”

Another important factor for you to consider before you start is your analysis, and it’s here where you’ve got to think about the optimal results you want; as in, you may wish to have only five people in your sample or as many as 25. If you do an in-person card sort, then it’s important that you figure out how to get your data from the physical world into the digital.

4. Organize the Content

Your content is one of the most critical parts of card sorting—if not the most—and your results will come from the content you give your participants; and that’s why you’ve got to evaluate them well.

Donna Spencer describes some of the elements and their significance of your card sorting content:

Show Hide video transcript

00:00:00 --> 00:00:30

So, what we want is content that is all at the *same level*. So, what I mean by that is we need it to be the same kind of granularity, the same kind of part of a hierarchy. An example that I saw a very long time ago that's a good example for this is I saw somebody running a card sort about tourism content. And in the tourism content, they had things like a hotel name, a zoo to visit,
00:00:30 --> 00:01:00

a winery to visit, they had some stuff about the geography, and then they had a card called 'accommodation'. So, you can think like if you've got a hotel – actually, I think they had a card of 'things to do' as well. So, you have like a 'hotel' card and an 'accommodation' card. They're different levels of hierarchy. A hotel is more detailed than the broad concept of accommodation. So, we don't want that 'accommodation' card in the mix because people will put those together
00:01:00 --> 00:01:34

– you cannot stop them from putting a hotel and accommodation together. And then, you haven't really learned about the content and the sort. You've just learned the thing that was going to happen. So, you want to make sure that they're – and the level is a hard concept, but you can see it in your content; like, you want either quite detailed things or fairly broad things, but they need to be the same. You want the cards to be a good selection that represents the content you'll be working with in the end.
00:01:34 --> 00:02:04

If you were card sorting for a supermarket, a grocery store, but you didn't include anything cold, so no frozen foods, no refrigerated foods, and then you ran the card sort, you're going to learn about some things but you're not going to learn about everything. And then you won't have information to cover the things that you haven't covered in the sort. So, you want to have really good coverage across everything that you're arranging.
00:02:04 --> 00:02:35

And this can be hard because if you're working on a site that's large, you need to be able to get good coverage. You also need people to be able to actually group the content. So, if you get a good broad selection that covers your content but then people look at it and they go, 'Well, I can see some chocolate and some apples and some cheese and some tinned foods and things, but I can't make a *group* out of those,'
00:02:35 --> 00:02:49

then you haven't learned anything from the activity. So, this is a juggle – you need to get a breadth that covers your content but still people can make groups out of.

If you’ve got an extensive content set, it may be wise to split them into several card sorts so you get the desired results.

Steer clear of jargon. Avoiding technical terms like that is a vital thing, and that’s because if people can’t understand your content, they won’t be able to group the cards that well and it may affect your results and distort things. People tend to force associations and make connections that aren’t there when they don’t fully understand the information, and that’s why it’s so important to be sure you use clear and direct content; it’s one of the factors that your getting the best possible outcome depends on.

5. Choose the Participants

The people you select for your card sort are really important; Donna Spencer explains what you need to consider.

Show Hide video transcript

00:00:00 --> 00:00:32

The people who you invite to be involved are also super important – not quite as important as the content, but still very important. Small groups of people working in-person around a table is literally the best thing because you can listen to the people and the conversation that they have and that conversation is so much more useful than the actual results that you get. We still could do this remotely, especially if people can get together and you can moderate via video.
00:00:32 --> 00:01:01

Or... I haven't ever done this, but we could kind of do it with a Zoom call and like a MURAL board and get people to poke things around, but it's still – any time that you have a chance to do it face-to-face, do that. And one of the things that I like to do if I need to do remote card sorts is just at least a couple face-to-face because then I can really hear that discussion and understand what's going on with just some of the group and then I can look at the results from the others.
00:01:01 --> 00:01:37

We want to make sure that *the people are also of the same level*. And this goes for any kind of group research or group activities. If you have a boss and somebody much more junior, the junior people will always just defer to the boss. If you have technical people and non-technical people, the non-technical people might just go, 'I don't know!' and follow the technical people. So, you want to make sure that people are at the same kind of levels, and, if you can, the same kinds of work or life situation, or the same kinds of users
00:01:37 --> 00:02:06

because then you can listen to them discuss their experiences and they'll do that better if they're the same types. We do that for most kind of focus groups and things. If you are doing it in-person, just don't try to fill up a room with lots of people. Do it with one group or have one facilitator per group because, as I said, the conversation is the super-useful part of this and it's really worth listening.
00:02:06 --> 00:02:32

What I used to do when I did these in-person is I would kind of just sit up at the end of the table, kind of be typing in results from a previous sort; so, I don't look like I'm hovering over them. I look like I'm busy kind of just listening a bit. But I'm listening to every single word they're saying! So, again, if we're working remotely and we're doing a remote card sort, you still want to have a good representative set of people and see if you can figure out a way of getting
00:02:32 --> 00:02:44

people to work together remotely. Otherwise, they'll just do it by themselves one at a time, and you'll still get good results, but you just won't hear them discussing their reasoning.

You might wonder about what the ideal number of participants is for your card sort. Sure, due to space and time constraints, you’ll find in-person card sorts easier to manage. For an online card sort, send the exercise to a small group first, as it’ll let you revise and adapt your card sort and, if you have to, you can always reach out to more people later.

6. Facilitate the Process

It’s best to start out with an introduction, give your participants a bit of context, and then go through the instructions. In a physical card sort, you can speak to your participants, but, if you do it remotely via Zoom or something similar, then you’ll have to write them for most virtual card sorts.

Don’t keep explaining or overhelp. One or some of your participants may likely look to you for guidance in the session; you can repeat the instructions, sure, but don’t go into more detail. You’ll get better results if they carry on without further assistance from you.

It’s better not to tell the people in your card sort that they’ll label the groups they create. Unfortunately, that’s not something you can do in a virtual card sort with an online tool. Language and labels will influence the categories—your participants are going to think more about the categories than the content that goes into them, and you’ll lose the fluidity of the card-sorting process.

Take notes during the card sorting process, and be sure to listen to what your participants say. Try not to hover, though, as this can get in the way of the organic conversation between the participants.

Offer incentives—they’re remarkably effective—and you’ll find this especially valuable in virtual settings where people commit less. Think about what might motivate your participants—such as a small gift, a discount voucher, or even a chance to win something bigger.

7. Analyze the Results

You'll find analysis tends to be easier with an online card sort, and that’s because it directly inputs results into your computer. For an in-person card sort, though, you can use numbers and a spreadsheet to simplify things, and number each card and put it into your spreadsheet, along with the card’s content; then, all you’ve got to do is input the numbers.

First, you want to get an overview. Donna Spencer shares how you do this:

Show Hide video transcript

00:00:00 --> 00:00:34

The first thing you want to do is just *poke around*. If you've got it in a big spreadsheet or something, you might print it out on the wall, you sort of look at: What *happened* here? What did people *do*? What's kind of *interesting*? What *ways* did they do things? So, just an initial – 'What is this?' You might look through and say, 'Okay, Participant 1 – how did they do that? Oh, look – they sorted the groceries into the way it's sorted in a physical store.'
00:00:34 --> 00:01:03

Or, 'They sorted the groceries according to taste – interesting. It's just interesting.' Then you might go through and go, 'What is interesting about those groupings? Ah, wow – okay, this is different to that. I didn't expect to see people put those things together.' And then look at like different participants and say, 'What did people do that is *different*? Did everybody do exactly the same thing? Did they do all different things?
00:01:03 --> 00:01:26

Did they do some similar things and some differents? What is interesting about that?' – What kind of variations did you notice? And then, also, *what words* did people put on the cards? What language and labeling came out of this activity is super interesting. So... you're kind of just burying yourself in data initially; you're not trying to get to a result really quickly.

Then, go into more depth, compare participants in detail, and see what the similarities and differences in their results are. And then look at what insights you can pull from the variations you find. For an IA, did they all make an “About Us” section? If they did, then what cards did they put into that group? Sometimes, people will create the same groupings but put different cards into the groupings. You can look at specific cards, too, and see where each person puts them.

Remember, it’s a qualitative research exercise, so you’re trying to get insights—and you want to discover what’s interesting or what stands out in your results. Review your user research objectives and what you wanted to learn from the card sort.

Learn more about quantitative and qualitative user research in this video.

Show Hide video transcript

00:00:00 --> 00:00:32

Ah, well – it's a lovely day here in Tiree. I'm looking out the window again. But how do we know it's a lovely day? Well, I could – I won't turn the camera around to show you, because I'll probably never get it pointing back again. But I can tell you the Sun's shining. It's a blue sky. I could go and measure the temperature. It's probably not that warm, because it's not early in the year. But there's a number of metrics or measures I could use. Or perhaps I should go out and talk to people and see if there's people sitting out and saying how lovely it is
00:00:32 --> 00:01:01

or if they're all huddled inside. Now, for me, this sunny day seems like a good day. But last week, it was the Tiree Wave Classic. And there were people windsurfing. The best day for them was not a sunny day. It was actually quite a dull day, quite a cold day. But it was the day with the best wind. They didn't care about the Sun; they cared about the wind. So, if I'd asked them, I might have gotten a very different answer than if I'd asked a different visitor to the island
00:01:01 --> 00:01:31

or if you'd asked me about it. And it can be almost a conflict between people within HCI. It's between those who are more *quantitative*. So, when I was talking about the sunny day, I could go and measure the temperature. I could measure the wind speed if I was a surfer – a whole lot of *numbers* about it – as opposed to those who want to take a more *qualitative* approach. So, instead of measuring the temperature, those are the people who'd want to talk to people to find out more about what *it means* to be a good day.
00:01:31 --> 00:02:02

And we could do the same for an interface. I can look at a phone and say, "Okay, how long did it take me to make a phone call?" Or I could ask somebody whether they're happy with it: What does the phone make them feel about? – different kinds of questions to ask. Also, you might ask those questions – and you can ask this in both a qualitative and quantitative way – in a sealed setting. You might take somebody into a room, give them perhaps a new interface to play with. You might – so, take the computer, give them a set of tasks to do and see how long they take to do it. Or what you might do is go out and watch
00:02:02 --> 00:02:30

people in their real lives using some piece of – it might be existing software; it might be new software, or just actually observing how they do things. There's a bit of overlap here – I should have mentioned at the beginning – between *evaluation techniques* and *empirical studies*. And you might do empirical studies very, very early on. And they share a lot of features with evaluation. They're much more likely to be wild studies. And there are advantages to each. In a laboratory situation, when you've brought people in,
00:02:30 --> 00:03:00

you can control what they're doing, you can guide them in particular ways. However, that tends to make it both more – shall we say – *robust* that you know what's going on but less about the real situation. In the real world, it's what people often call "ecologically valid" – it's about what they *really* are up to. But it is much less controlled, harder to measure – all sorts of things. Very often – I mean, it's rare or it's rarer to find more quantitative in-the-wild studies, but you can find both.
00:03:00 --> 00:03:34

You can both go out and perhaps do a measure of people outside. You might – you know – well, go out on a sunny day and see how many people are smiling. Count the number of smiling people each day and use that as your measure – a very quantitative measure that's in the wild. More often, you might in the wild just go and ask people. It's a more qualitative thing. Similarly, in the lab, you might do a quantitative thing – some sort of measurement – or you might ask something more qualitative – more open-ended. Particularly quantitative and qualitative methods,
00:03:34 --> 00:04:01

which are often seen as very, very different, and people will tend to focus on one *or* the other. *Personally*, I find that they fit together. *Quantitative* methods tend to tell me whether something happens and how common it is to happen, whether it's something I actually expect to see in practice commonly. *Qualitative* methods – the ones which are more about asking people open-ended questions – either to both tell me *new* things that I didn't think about before,
00:04:01 --> 00:04:32

but also give me the *why* answers if I'm trying to understand *why* it is I'm seeing a phenomenon. So, the quantitative things – the measurements – say, "Yeah, there's something happening. People are finding this feature difficult." The qualitative thing helps me understand what it is about it that's difficult and helps me to solve it. So, I find they give you *complementary things* – they work together. The other thing you have to think about when choosing methods is about *what's appropriate for the particular situation*. And these things don't always work.
00:04:32 --> 00:05:00

Sometimes, you can't do an in-the-wild experiment. If it's about, for instance, systems for people in outer space, you're going to have to do it in a laboratory. You're not going to go up there and experiment while people are flying around the planet. So, sometimes you can't do one thing or the other. It doesn't make sense. Similarly, with users – if you're designing something for chief executives of Fortune 100 companies, you're not going to get 20 of them in a room and do a user study with them.
00:05:00 --> 00:05:07

That's not practical. So, you have to understand what's practical, what's reasonable and choose your methods accordingly.

Last—but not least—apply your findings to your project, and it’s best to combine them with other research methods.

Card Sorting Tips and Best Practices

You can leverage card sorting as a powerful tool in user experience research, but its success really does hinge on how you conduct it. You can consider this method a vital step in designing intuitive interfaces.

With that said, though, you’ve got to stick to certain tips and best practices if you’re to fully capitalize on its benefits. Follow these guidelines (which are numbered to match the diagram above) as they’ll help you conduct productive card-sorting sessions that yield actionable results.

Use relevant and straightforward cards. Avoid jargon and complex terms—remember, the clarity of your cards can impact the quality of your results.
Choose a diverse group of participants who represent your target audience; diversity is a powerful engine that brings varied perspectives.
Whether it’s online or in-person, make sure you provide a user-friendly sorting environment for participants. For online sorts, pick a tool that's easy to navigate. For in-person ones, make sure there’s a comfortable and distraction-free space.
Give your participants concise, understandable instructions. Be sure to clarify any ambiguities and be ready to answer questions, remembering that good instructions lead to more accurate sorting.
Too many cards can swamp and overwhelm participants, so it’s best to aim for a balanced number that's comprehensive yet manageable, typically around 40-50 cards.
Encourage participants to think freely in open sorts, as that openness is what can lead to innovative categorizations and unexpected insights.
Observe things quietly during the sort, and make notes about how participants interact with the cards and listen to their comments. The user observations you make can end up being as valuable as the sort itself.
“Post-sort” (i.e. after the session), spend time analyzing the results. Look for patterns, outliers, and surprising groupings, and this detailed analysis is where the real value of card sorting lies.
Don't hesitate to run multiple card sorting sessions, and that’s because iterations can refine your understanding and lead to more robust conclusions.
Use card sorting with other research methods like surveys or user interviews—it’s wise as this multi-method approach will give you a more comprehensive view.
Test your method before using card sorting, and it’ll ensure that it’s not just effective but means it’s nice and tailorable for meaningful insights, too.

The Take Away

When you want to know how your users see and think about concepts and items you need to design for, then card sorting is a handy way to reveal patterns so you can organize websites, apps, and the like in a way that matches their mental models best. Key factors like approach (i.e., in-person or remote), objectives, tools, content, participants (including the number of them who are participating in the card sort) are things to bear in mind, although you’ll likely find the process straightforward—but you’ll need to consider each element with care.

To be sure, card sorting is one technique that can offer invaluable insights, but it’ll still call for you to be mindful that the results you get from it may not always be clear. That said, it’s still a handy go-to thanks to its low cost, simplicity, and low investment of effort—so try it, use it carefully, and you may come away with some of the most helpful insights you can hope for to guide how your website or app ends up looking and functioning successfully.

References and Where to Learn More

Dive deeper into card sorting with Donna Spencer’s Master Class, How To Use Card Sorting For Better Information Architecture

Learn how to run card sorts and what tools to use:

William Hudson explains How to Screen Research Participants

Card Sorting: The Ultimate Guide (in 2025)

What Is Card Sorting in UX Research?

Why Do We Use Card Sorting?

Advantages and Disadvantages of Card Sorting

Advantages of Card Sorting

Disadvantages of Card Sorting

Types of Card Sorting

1. Open Card Sorting

2. Closed Card Sorting

3. Hybrid Card Sorting

4. Reverse Card Sorting

5. Modified-Delphi Card Sorting

6. Remote Card Sorting

Tree Testing vs Card Sorting: Which One Should You Prefer?

What Tools Do You Need to Conduct a Card Sort?

Considerations for Choosing a Card Sorting Tool

How to Run a Card Sort?

1. Set Clear Objectives

2. Select the Right Tools

3. Write Clear Instructions for Effective Participation

4. Organize the Content

5. Choose the Participants

6. Facilitate the Process

7. Analyze the Results

Card Sorting Tips and Best Practices

The Take Away

References and Where to Learn More

User Research – Methods and Best Practices

Get Weekly Design Insights

Topics in This Article

What You Should Read Next

Revolutionize UX Design with VR Experiences

Open Access—Link to us!

Popular related searches

Filters

Privacy Settings