How to interpret your card sort results Part 1: open and hybrid card sorts
Cards have been created, sorted and sorted again. The participants are all finished and you’re left with a big pile of awesome data that will help you improve the user experience of your information architecture. Now what?Whether you’ve run an open, hybrid or closed card sort online using an information architecture tool or you’ve run an in person (moderated) card sort, it can be a bit daunting trying to figure out where to start the card sort analysis process.
About this guide
This two-part guide will help you on your way! For Part 1, we’re going to look at how to interpret and analyze the results from open and hybrid card sorts.
In open card sorts, participants sort cards into categories that make sense to them and they give each category a name of their own making.
In hybrid card sorts, some of the categories have already been defined for participants to sort the cards into but they also have the ability to create their own.
Open and hybrid card sorts are great for generating ideas for category names and labels and understanding not only how your users expect your content to be grouped but also what they expect those groups to be called.In both parts of this series, I’m going to be talking a lot about interpreting your results using Optimal Workshop’s online card sorting tool, OptimalSort, but most of what I’m going to share is also applicable if you’re analyzing your data using a spreadsheet or using another tool.
Understanding the two types of analysis: exploratory and statistical
Similar to qualitative and quantitative methods, exploratory and statistical analysis in card sorting are two complementary approaches that work together to provide a detailed picture of your results.
Exploratory analysis is intuitive and creative. It’s all about going through the data and shaking it to see what ideas, patterns and insights fall out. This approach works best when you don’t have the numbers (smaller sample sizes) and when you need to dig into the details and understand the ‘why’ behind the statistics.
Statistical analysis is all about the numbers. Hard data that tells you exactly how many people expected X to be grouped with Y and more and is very useful when you’re dealing with large sample sizes and when identifying similarities and differences across different groups of people.
Depending on your objectives - whether you are starting from scratch or redesigning an existing IA - you’ll generally need to use some combination of both of these approaches when analyzing card sort results. Learn more about exploratory and statistical analysis in Donna Spencer’s book.
Start with the big picture
When analyzing card sort results, start by taking an overall look at the results as a whole. Quickly cast your eye over each individual card sort and just take it all in. Look for common patterns in how the cards have been sorted and the category names given by participants. Does anything jump out as surprising? Are there similarities or differences between participant sorts? If you’re redesigning an existing IA, how do your results compare to the current state?If you ran your card sort using OptimalSort, your first port of call will be the Overview and Participants Table presented in the results section of the tool.If you ran a moderated card sort using OptimalSort’s printed cards, now is a good time to double check you got them all. And if you didn’t know about this handy feature of OptimalSort, it’s something to keep in mind for next time!The Participants Table shows a breakdown of your card sorting data by individual participant. Start by reviewing each individual card sort one by one by clicking on the arrow in the far left column next to the Participants numbers.
Viewing individual participant card sorts in detail.
From here you can easily flick back and forth between participants without needing to close that modal window. Don’t spend too much time on this — you’re just trying to get a general impression of what happened.Keep an eye out for any card sorts that you might like to exclude from the results. For example participants who have lumped everything into one group and haven’t actually sorted the cards. Don’t worry - excluding or including participants isn’t permanent and can be toggled on or off at anytime.If you have a good number of responses, then the Participant Centric Analysis (PCA) tab (below) can be a good place to head next. It’s great for doing a quick comparison of the different high-level approaches participants took when grouping the cards.The PCA tab provides the most insight when you have lots of results data (30+ completed card sorts) and at least one of the suggested IAs has a high level of agreement among your participants (50% or more agree with at least one IA).
Participant Centric Analysis (PCA) tab for an open or hybrid card sort in OptimalSort.
The PCA tab compares data from individual participants and surfaces the top three ways the cards were sorted. It also gives you some suggestions based on participant responses around what these categories could be called but try not to get too bogged down in those - you’re still just trying to gain an overall feel for the results at this stage.Now is also a good time to take a super quick peek at the Categories tab as it will also help you spot patterns and identify data that you’d like to dive deeper into a bit later on!Another really useful visualization tool offered by OptimalSort that will help you build that early, high-level picture of your results is the Similarity Matrix. This diagram helps you spot data clusters, or groups of cards that have been more frequently paired together by your participants, by surfacing them along the edge and shading them in dark blue. It also shows the proportion of times specific card pairings occurred during your study and displays the exact number on hover (below).
OptimalSort’s Similarity Matrix showing that ‘Flat sandals’ and ‘Court shoes’ were paired by 91% of participants (31 times) in this example study.
In the above screenshot example we can see three very clear clusters along the edge: ‘Ankle Boots’ to ‘Slippers’ is one cluster, ‘Socks’ to ‘Stockings & Hold Ups’ is the next and then we have ‘Scarves’ to ‘Sunglasses’. These clusters make it easy to spot the that cards that participants felt belonged together and also provides hard data around how many times that happened.Next up are the dendrograms. Dendrograms are also great for gaining an overall sense of how similar (or different) your participants’ card sorts were to each other. Found under the Dendrogram tab in the results section of the tool, the two dendrograms are generated by different algorithms and which one you use depends largely on how many participants you have.
If your study resulted in 30 or more completed card sorts, use the Actual Agreement Method (AAM) dendrogram and if your study had fewer than 30 completed card sorts, use the Best Merge Method (BMM) dendrogram.The AAM dendrogram (see below) shows only factual relationships between the cards and displays scores that precisely tell you that ‘X% of participants in this study agree with this exact grouping’.In the below example, the study shown had 34 completed card sorts and the AAM dendrogram shows that 77% of participants agreed that the cards highlighted in green belong together and a suggested name for that group is ‘Bling’. The tooltip surfaces one of the possible category names for this group and as demonstrated here it isn’t always the best or ‘recommended’ one. Take it with a grain of salt and be sure to thoroughly check the rest of your results before committing!
AAM Dendrogram in OptimalSort.
The BMM dendrogram (see below) is different to the AAM because it shows the percentage of participants that agree with parts of the grouping - it squeezes the data from smaller sample sizes and makes assumptions about larger clusters based on patterns in relationships between individual pairs.The AAM works best with larger sample sizes because it has more data to work with and doesn’t make assumptions while the BMM is more forgiving and seeks to fill in the gaps.The below screenshot was taken from an example study that had 7 completed card sorts and its BMM dendrogram shows that 50% of participants agreed that the cards highlighted in green down the left hand side belong to ‘Accessories, Bottoms, Tops’.
BMM Dendrogram in OptimalSort.
Drill down and cross-reference
Once you’ve gained a high level impression of the results, it’s time to dig deeper and unearth some solid insights that you can share with your stakeholders and back up your design decisions.Explore your open and hybrid card sort data in more detail by taking a closer look at the Categories tab. Open up each category and cross-reference to see if people were thinking along the same lines.Multiple participants may have created the same category label, but what lies beneath could be a very different story. It’s important to be thorough here because the next step is to start standardizing or chunking individual participant categories together to help you make sense of your results.In open and hybrid sorts, participants will be able to label their categories themselves. This means that you may identify a few categories with very similar labels or perhaps spelling errors or different formats. You can standardize your categories by merging similar categories together to turn them into one.OptimalSort makes this really easy to do - you pretty much just tick the boxes alongside each category name and then hit the ‘Standardize’ button up the top (see below). Don’t worry if you make a mistake or want to include or exclude groupings; you can unstandardize any of your categories anytime.
Standardizing categories in OptimalSort.
Once you’ve standardized a few categories, you’ll notice that the Agreement number may change. It tells you how many participants agreed with that grouping. An agreement number of 1.0 is equal to 100% meaning everyone agrees with everything in your newly standardized category while 0.6 means that 60% of your participants agree.Another number to watch for here is the number of participants who sorted a particular card into a category which will appear in the frequency column in dark blue in the right-hand column of the middle section of the below image.
Categories table after groupings called ‘Accessories’ and ‘Bags’ have been standardized.
A closer look at the standardized category for ‘Accessories’.
From the above screenshot we can see that in this study, 18 of the 26 participant categories selected agree that ‘Cat Eye Sunglasses’ belongs under ‘Accessories’.Once you’ve standardized a few more categories you can head over to the Standardization Grid tab to review your data in more detail. In the below image we can see that 18 participants in this study felt that ‘Backpacks’ belong in a category named ‘Bags’ while 5 grouped them under ‘Accessories’. Probably safe to say the backpacks should join the other bags in this case.
Standardization Grid in OptimalSort.
So that’s a quick overview of how to interpret the results from your open or hybrid card sorts.Here's a link to Part 2 of this series where we talk about interpreting results from closed card sorts as well as next steps for applying these juicy insights to your IA design process.
Further reading
Card Sorting 101 – Learn about the differences between open, closed and hybrid card sorts, and how to run your own using OptimalSort.
I have two huge content structures I would like to A/B test. Do you think Treejack would be appropriate?"
— Mike
Hi Mike (and excellent question)!
Firstly, yes, Treejack is great for testing more than one content structure. It’s easy to run two separate Treejack studies — even more than two. It’ll help you decide which structure you and your team should run with, and it won’t take you long to set them up.
When you’re creating the two tree tests with your two different content structures, include the same tasks in both tests. Using the same tasks will give an accurate measure of which structure performs best. I’ve done it before and I found that the visual presentation of the results — especially the detailed path analysis pietrees — made it really easy to compare Test A with Test B.
Plus (and this is a big plus), if you need to convince stakeholders or teammates of which structure is the most effective, you can’t go past quantitative data, especially when its presented clearly — it’s hard to argue with hard evidence!
Here’s two example of the kinds of results visualizations you could compare in your A/B test: the pietree, which shows correct and incorrect paths, and where people ended up:
And the overall Task result, which breaks down success and directness scores, and has plenty of information worth comparing between two tests:
Keep in mind that running an A/B tree test will affect how you recruit participants — it may not be the best idea to have the same participants complete both tests in one go. But it’s an easy fix — you could either recruit two different groups from the same demographic, or test one group and have a gap (of at least a day) between the two tests.
I’ve one more quick question: why are your two content structures ‘huge’?
I understand that sometimes these things are unavoidable — you potentially work for a government organization, or a university, and you have to include all of the things. But if not, and if you haven’t already, you could run an open card sort to come up with another structure to test (think of it as an A/B/C test!), and to confirm that the categories you’re proposing work for people.
You could even run a closed card sort to establish which content is more important to people than others (your categories could go from ‘Very important’ to ‘Unimportant’, or ‘Use everyday’ to ‘Never use’, for example). You might be able to make your content structure a bit smaller, and still keep its usefulness. Just a thought... and of course, you could try to get this information from your analytics (if available) but just be cautious of this because of course analytics can only tell you what people did and not what they wanted to do.
The key purpose of running a card sort is to learn something new about how people conceptualize and organize the information that’s found on your website. The insights you gain from running a card sort can then help you develop a site structure with content labels or headings that best represent the way your users think about this information. Card sorts are in essence a simple technique, however it’s the details of the sort that can determine the quality of your results.
Adding context to cards in OptimalSort – descriptions, links and images
In most cases, each item in a card sort has only a short label, but there are instances where you may wish to add additional context to the items in your sort. Currently, the cards tab in OptimalSort allows you to include a tooltip description, a link within the tooltip description or to format the card as an image (with or without a label).
We generally don’t recommend using tooltip descriptions and links, unless you have a specific reason to do so. It’s likely that they’ll provide your participants with more information than they would normally have when navigating your website, which may in turn influence your results by leading participants to a particular solution.
Legitimate reasons that you may want to use descriptions and links include situations where it’s not possible or practical to translate complex or technical labels (for example, medical, financial, legal or scientific terms) into plain language, or if you’re using a card sort to understand your participants’ preferences or priorities.
If you do decide to include descriptions in your sort, it’s important that you follow the same guidelines that you would otherwise follow for writing card labels. They should be easy for your participants to understand and you should avoid obvious patterns, for example repeating words and phrases, or including details that refer to the current structure of the website.
A quick survey of how card descriptions are used in OptimalSort
I was curious to find out how often people were including descriptions in their card sorts, so I asked our development team to look into this data. It turns out that around 15% of cards created in OptimalSort have at least some text entered in the description field. In order to dig into the data a bit further, both Ania and I reviewed a random sample of recent sorts and noted how descriptions were being used in each case.
We found that out of the descriptions that we reviewed, 40% (6% of the total cards) had text that should not have impacted the sort results. Most often, these cards simply had the card label repeated in the description (to be honest, we’re not entirely sure why so many descriptions are being used this way! But it’s now in our roadmap to stop this from happening — stay tuned!). Approximately 20% (3% of the total cards) used descriptions to add context without obviously leading participants, however another 40% of cards have descriptions that may well lead to biased results. On occasion, this included linking to the current content or using what we assumed to be the current top level heading within the description.
Testing the effect of card descriptions on sort results
So, how much influence could potentially leading card descriptions have on the results of a card sort? I decided to put it to the test by running a series of card sorts to compare the effect of different descriptions. As I also wanted to test the effect of linking card descriptions to existing content, I had to base the sort on a live website. In addition, I wanted to make sure that the card labels and descriptions were easily comprehensible by a general audience, but not so familiar that participants were highly likely to sort the cards in a similar manner.
I selected the government immigration website New Zealand Now as my test case. This site, which provides information for prospective and new immigrants to New Zealand, fit the above criteria and was likely unfamiliar to potential participants.
Navigating the New Zealand Now website
When I reviewed the New Zealand Now site, I found that the top level navigation labels were clear and easy to understand for me personally. Of course, this is especially important when much of your target audience is likely to be non-native English speaking! On the whole, the second level headings were also well-labeled, which meant that they should translate to cards that participants were able to group relatively easily.
There were, however, a few headings such as “High quality” and “Life experiences”, both found under “Study in New Zealand”, which become less clear when removed from the context of their current location in the site structure. These headings would be particularly useful to include in the test sorts, as I predicted that participants would be more likely to rely on card descriptions in the cases where the card label was ambiguous.
I selected 30 headings to use as card labels from under the sections “Choose New Zealand”, “Move to New Zealand”, “Live in New Zealand”, “Work in New Zealand” and “Study in New Zealand” and tweaked the language slightly, so that the labels were more generic.
I then created four separate sorts in OptimalSort:Round 1: No description: Each card showed a heading only — this functioned as the control sort
Round 2: Site section in description: Each card showed a heading with the site section in the description
Round 3: Short description: Each card showed a heading with a short description — these were taken from the New Zealand Now topic landing pages
Round 4:Link in description: Each card showed a heading with a link to the current content page on the New Zealand Now website
For each sort, I recruited 30 participants. Each participant could only take part in one of the sorts.
What the results showed
An interesting initial finding was that when we queried the participants following the sort, only around 40% said they noticed the tooltip descriptions and even fewer participants stated that they had used them as an aid to help complete the sort.
Of course, what people say they do does not always reflect what they do in practice! To measure the effect that different descriptions had on the results of this sort, I compared how frequently cards were sorted with other cards from their respective site sections across the different rounds.Let’s take a look at the “Study in New Zealand” section that was mentioned above. Out of the five cards in this section,”Where & what to study”, “Everyday student life” and “After you graduate” were sorted pretty consistently, regardless of whether a description was provided or not. The following charts show the average frequency with which each card was sorted with other cards from this section. For example in the control round, “Where & what to study” was sorted with “After you graduate” 76% of the time and with “Everyday day student life” 70% of the time, but was sorted with “Life experiences” or “High quality” each only 10% of the time. This meant that the average sort frequency for this card was 42%.
On the other hand, the cards “High quality” and “Life experiences” were sorted much less frequently with other cards in this section, with the exception of the second sort, which included the site section in the description.These results suggest that including the existing site section in the card description did influence how participants sorted these cards — confirming our prediction! Interestingly, this round had the fewest number of participants who stated that they used the descriptions to help them complete the sort (only 10%, compared to 40% in round 3 and 20% in round 4).Also of note is that adding a link to the existing content did not seem to increase the likelihood that cards were sorted more frequently with other cards from the same section. Reasons for this could include that participants did not want to navigate to another website (due to time-consciousness in completing the task, or concern that they’d lose their place in the sort) or simply that it can be difficult to open a link from the tooltip pop-up.
What we can take away from these results
This quick investigation into the impact of descriptions illustrates some of the intricacies around using additional context in your card sorts, and why this should always be done with careful consideration. It’s interesting that we correctly predicted some of these results, but that in this case, other uses of the description had little effect at all. And the results serve as a good reminder that participants can often be influenced by factors that they don’t even recognise themselves!If you do decide to use card descriptions in your cards sorts, here are some guidelines that we recommend you follow:
Avoid repeating words and phrases, participants may sort cards by pattern-matching rather than based on the actual content
Avoid alluding to a predetermined structure, such as including references to the current site structure
If it’s important that participants use the descriptions to complete the sort, you should mention this in your task instructions. It may also be worth asking them a post-survey question to validate if they used them or not
We’d love to hear your thoughts on how we tested the effects of card descriptions and the results that we got. Would you have done anything differently?Have you ever completed a card sort only to realize later that you’d inadvertently biased your results? Or have you used descriptions in your card sorts to meet a genuine need? Do you think there’s a case to make descriptions more obvious than just a tooltip, so that when they are used legitimately, most participants don’t miss this information?
Card sorting is an invaluable tool for understanding how people organize information in their minds, making websites more intuitive and content easier to navigate. It’s a useful method outside of information architecture and UX research, too. It can be a useful prioritization technique, or used in a more traditional sense. For example, it’s handy in psychology, sociology or anthropology to inform research and deepen our understanding of how people conceptualize information.
The introduction of remote card sorting has provided many advantages, making it easier than ever to conduct your own research. Tools such as our very own OptimalSort allow you to quickly and easily gather findings from a large number of participants from all around the world. Not having to organize moderated, face-to-face sessions gives researchers more time to focus on their work, and easier access to larger data sets.
One of the main disadvantages of remote card sorting is that it eliminates the opportunity to dive deeper into the choices made by your participants. Human conversation is a great thing, and when conducting a remote card sort with users who could potentially be on the other side of the world, opportunities for our participants to provide direct feedback and voice their opinions are severely limited.Your survey design may not be perfect.
The labels you provide your participants may be incorrect, confusing or redundant. Your users may have their own ideas of how you could improve your products or services beyond what you are trying to capture in your card sort. People may be more willing to provide their feedback than you realize, and limiting their insights to a simple card sort may not capture all that they have to offer.So, how can you run an unmoderated, remote card sort, but do your best to mitigate this potential loss of insight?
A quick look into the data
In an effort to evaluate the usefulness of the existing “Leave a comment” feature in OptimalSort, I recently asked our development team to pull out some data.You might be asking “There’s a comment box in OptimalSort?”If you’ve never noticed this feature, I can’t exactly blame you. It’s relatively hidden away as an unassuming hyperlink in the top right corner of your card sort.
Comments left by your participants can be viewed in the “Participants” tab in your results section, and are indicated by a grey speech bubble.
The history of the button is unknown even to long-time Optimal Workshop team members. The purpose of the button is also unspecified. “Why would anyone leave a comment while participating in a card sort?”, I found myself wondering.As it turns out, 133,303 comments have been left by participants. This means 133,303 insights, opinions, critiques or frustrations. Additionally, these numbers only represent the participants who noticed the feature in the first place. Considering the current button can easily be missed when focusing on the task at hand, I can’t help but wonder how this number might change if we drew more attention to the feature.
Breaking down the comments
To avoid having to manually analyze and code 133,303 open text fields, I decided to only spend enough time to decipher any obvious patterns. Luckily for me, this didn’t take very long. After looking at only a hundred or so random entries, four distinct types of comments started to emerge.
This card/group doesn’t make sense.Comments related to cards and groups dominate. This is a great thing, as it means that the majority of comments made by participants relate specifically to the task they are completing. For closed and hybrid sorts, comments frequently relate to the predefined categories available, and since the participants most likely to leave a comment are those experiencing issues, the majority of the feedback relates to issues with category names themselves. Many comments are related to card labels and offer suggestions for improving naming conventions, while many others draw attention to some terms being confusing, unclear or jargony. Comments on task length can also be found, along with reasons for why certain cards may be left ungrouped, e.g., “I’ve left behind items I think the site could do without”.
Your organization is awesome for doing this/you’re doing it all wrong. A substantial number of participants used the comment box as an opportunity to voice their general feedback on the organization or company running the study. Some of the more positive comments include an appreciation for seeing private companies or public sector organizations conducting research with real users in an effort to improve their services. It’s also nice to see many comments related to general enjoyment in completing the task.On the other hand, some participants used the comment box as an opportunity to comment on what other areas of their services should be improved, or what features they would like to see implemented that may otherwise be missed in a card sort, e.g., “Increased, accurate search functionality is imperative in a new system”.
This isn’t working for me. Taking a closer look at some of the comments reveals some useful feedback for us at Optimal Workshop, too. Some of the comments relate specifically to UI and usability issues. The majority of these issues are things we are already working to improve or have dealt with. However, for researchers, comments that relate to challenges in using the tool or completing the survey itself may help explain some instances of data variability.
#YOLO, hello, ;) And of course, the unrelated. As you may expect, when you provide people with the opportunity to leave a comment online, you can expect just about anything in return.
How to make the most of your user insights in OptimalSort
If you’re running a card sort, chances are you already place a lot of value in the voice of your users. To ensure you capture any additional insights, it’s best to ensure your participants are aware of the opportunity to do so. Here are two ways you may like to ensure your participants have a space to voice their feedback:
Adding more context to the “Leave a comment” feature
One way to encourage your participants to leave comments is to promote the use of the this feature in your card sort instructions. OptimalSort gives you flexibility to customize your instructions every time you run a survey. By making your participants aware of the feature, or offering ideas around what kinds of comments you may be looking for, you not only make them more likely to use the feature, but also open yourself up to a whole range of additional feedback. An advantage of using this feature is that comments can be added in real time during a card sort, so any remarks can be made as soon as they arise.
Making use of post-survey questions
Adding targeted post-survey questions is the best way to ensure your participants are able to voice any thoughts or concerns that emerged during the activity. Here, you can ask specific questions that touch upon different aspects of your card sort, such as length, labels, categories or any other comments your participants may have. This can not only help you generate useful insights but also inform the design of your surveys in the future.
Make your remote card sorts more human
Card sorts are exploratory by nature. Avoid forcing your participants into choices that may not accurately reflect their thinking by giving them the space to voice their opinions. Providing opportunities to capture feedback opens up the conversation between you and your users, and can lead to surprising insights from unexpected places.
