“Dear Optimal, I want to test the structure of a university website (well certain sections anyway). My gut instinct is that it's pretty 'broken'. Lots of sections feel like they're in the wrong place. I want to test my hypotheses before proposing a new structure. I'm definitely going to do some card sorting, and was planning a mixture of online and offline. My question is about when to bring in tree testing. Should I do this first to test the existing IA? Or is card sorting sufficient? I do intend to tree test my new proposed IA in order to validate it, but is it worth doing it upfront too?" — Matt
Dear Matt,
Ah, the classic chicken or the egg scenario: Which should come first, tree testing or card sorting? It’s a question that many researchers often ask themselves, but I’m here to help clear the air! You should always use both methods when changing up your information architecture (IA) in order to capture the most information.
Tree testing and card sorting, when used together, can give you fantastic insight into the way your users interact with your site. First of all, I’ll run through some of the benefits of each testing method.
What is card sorting and why should I use it?
Card sorting is a great method to gauge the way in which your users organize the content on your site. It helps you figure out which things go together and which things don’t. There are two main types of card sorting: open and closed.
Closed card sorting involves providing participants with pre-defined categories into which they sort their cards. For example, you might be reorganizing the categories for your online clothing store for women. Your cards would have all the names of your products (e.g., “socks”, “skirts” and “singlets”) and you also provide the categories (e.g.,“outerwear”, “tops” and “bottoms”).
Open card sorting involves providing participants with cards and leaving them to organize the content in a way that makes sense to them. It’s the opposite to closed card sorting, in that participants dictate the categories themselves and also label them. This means you’d provide them with the cards only, and no categories.
Card sorting, whether open or closed, is very user focused. It involves a lot of thought, input, and evaluation from each participant, helping you to form the structure of your new IA.
What is tree testing and why should I use it?
Tree testing is a fantastic way to determine how your users are navigating your site and how they’re finding information. Your site is organized into a tree structure, sorted into topics and subtopics, and participants are provided with some tasks that they need to perform. The results will show you how your participants performed those tasks, if they were successful or unsuccessful, and which route they took to complete the tasks. This data is extremely useful for creating a new and improved IA.
Tree testing is an activity that requires participants to seek information, which is quite the contrast to card sorting. Card sorting is an activity that requires participants to sort and organize information. Each activity requires users to behave in different ways, so each method will give its own valuable results.
Comparing tree testing and card sorting: Key differences
Tree testing and card sorting are complementary methods within your UX toolkit, each unlocking unique insights about how users interact with your site structure. The difference is all about direction.
Card sorting is generative. It helps you understand how users naturally group and label your content; revealing mental models, surfacing intuitive categories, and informing your site’s information architecture (IA) from the ground up. Whether using open or closed methods, card sorting gives users the power to organize content in ways that make sense to them.
Tree testing is evaluative. Once you’ve designed or restructured your IA, tree testing puts it to the test. Participants are asked to complete find-it tasks using only your site structure – no visuals, no design – just your content hierarchy. This highlights whether users can successfully locate information and how efficiently they navigate your content tree.
In short:
Card sorting = "How would you organize this?"
Tree testing = "Can you find this?"
Using both methods together gives you clarity and confidence. One builds the structure. The other proves it works.
Which method should you choose?
The right method depends on where you are in your IA journey. If you're beginning from scratch or rethinking your structure, starting with card sorting is ideal. It will give you deep insight into how users group and label content.
If you already have an existing IA and want to validate its effectiveness, tree testing is typically the better fit. Tree testing shows you where users get lost and what’s working well. Think of card sorting as how users think your site should work, and tree testing as how they experience it in action.
Should you run a card or tree test first?
In this scenario, I’d recommend running a tree test first in order to find out how your existing IA currently performs. You said your gut instinct is telling you that your existing IA is pretty “broken”, but it’s good to have the data that proves this and shows you where your users get lost.
An initial tree test will give you a benchmark to work with – after all, how will you know your shiny, new IA is performing better if you don’t have any stats to compare it with? Your results from your first tree test will also show you which parts of your current IA are the biggest pain points and from there you can work on fixing them. Make sure you keep these tasks on hand – you’ll need them later!
Once your initial tree test is done, you can start your card sort, based on the results from your tree test. Here, I recommend conducting an open card sort so you can understand how your users organize the content in a way that makes sense to them. This will also show you the language your participants use to name categories, which will help you when you’re creating your new IA.
Finally, once your card sort is done you can conduct another tree test on your new, proposed IA. By using the same (or very similar) tasks from your initial tree test, you will be able to see that any changes in the results can be directly attributed to your new and improved IA.
Once your test has concluded, you can use this data to compare the performance from the tree test for your original information architecture.
Why using both methods together is most effective
Card sorting and tree testing aren’t rivals, view them as allies. Used together, they give you end-to-end clarity. Card sorting informs your IA design based on user mental models. Tree testing evaluates that structure, confirming whether users can find what they need. This combination creates a feedback loop that removes guesswork and builds confidence. You'll move from assumptions to validation, and from confusion to clarity – all backed by real user behavior.
A great information architecture (IA) is essential for a great user experience (UX). And testing your website or app’s information architecture is necessary to get it right.
Card sorting and tree testing are the very best UX research methods for exactly this. But the big question is always: which one should you use, and when? Very possibly you need both. Let’s find out with this quick summary.
What is card sorting and tree testing? 🧐
Card sorting is used to test the information architecture of a website or app. Participants group individual labels (cards) into different categories according to criteria that makes best sense to them. Each label represents an item that needs to be categorized. The results provide deep insights to guide decisions needed to create an intuitive navigation, comprehensive labeling and content that is organized in a user-friendly way.
Tree testing is also used to test the information architecture of a website or app. When using tree testing participants are presented with a site structure and a set of tasks they need to complete. The goal for participants is to find their way through the site and complete their task. The test shows whether the structure of your website corresponds to what users expect and how easily (or not) they can navigate and complete their tasks.
What are the differences? 🂱 👉🌴
Card sorting is a UX research method which helps to gather insights about your content categorization. It focuses on creating an information architecture that responds intuitively to the users’ expectations. Things like which items go best together, the best options for labeling, what categories users expect to find on each menu.
Doing a simple card sort can give you all those pieces of information and so much more. You start understanding your user’s thoughts and expectations. Gathering enough insights and information to enable you to develop several information architecture options.
Tree testing is a UX research method that is almost a card sort in reverse. Tree testing is used to evaluate an information architecture structure and simply allows you to see what works and what doesn’t.
Using tree testing will provide insights around whether your information architecture is intuitive to navigate, the labels easy to follow and ultimately if your items are categorized in a place that makes sense. Conversely it will also show where your users get lost and how.
What method should you use? 🤷
You’ve got this far and fine-tuning your information architecture should be a priority. An intuitive IA is an integral component of a user-friendly product. Creating a product that is usable and an experience users will come back for.
If you are still wondering which method you should use - tree testing or card sorting. The answer is pretty simple - use both.
Just like many great things, these methods work best together. They complement each other, allowing you to get much deeper insights and a rounded view of how your IA performs and where to make improvements than when used separately. We cover more reasons why card sorting loves tree testing in our article which dives deeper into why to use both.
Ok, I'm using both, but which comes first? 🐓🥚
Wanting full, rounded insights into your information architecture is great. And we know that tree testing and card sorting work well together. But is there an order you should do the testing in? It really depends on the particular context of your research - what you’re trying to achieve and your situation.
Tree testing is a great tool to use when you have a product that is already up and running. By running a tree test first you can quickly establish where there may be issues, or snags. Places where users get caught and need help. From there you can try and solve potential issues by moving on to a card sort.
Card sorting is a super useful method that can be instigated at any stage of the design process, from planning to development and beyond. As long as there is an IA structure that can be tested again. Testing against an already existing website navigation can be informative. Or testing a reorganization of items (new or existing) can ensure the organization can align with what users expect.
However, when you decide to implement both of the methods in your research, where possible, tree testing should come before card sorting. If you want a little more on the issue have a read of our article here.
Check out our OptimalSort and Treejack tools - we can help you with your research and the best way forward. Wherever you might be in the process.
We love getting stuck into scary, hairy problems to make things better here at Trade Me. One challenge for us in particular is how best to navigate customer reaction to any change we make to the site, the app, the terms and conditions, and so on. Our customers are passionate both about the service we provide — an online auction and marketplace — and its place in their lives, and are rightly forthcoming when they're displeased or frustrated. We therefore rely on our Customer Service (CS) team to give customers a voice, and to respond with patience and skill to customer problems ranging from incorrectly listed items to reports of abusive behavior.
The CS team uses a Customer Relationship Management (CRM) system, Trade Me Admin, to monitor support requests and manage customer accounts. As the spectrum of Trade Me's services and the complexity of the public website have grown rapidly, the CRM system has, to be blunt, been updated in ways which have not always been the prettiest. Links for new tools and reports have simply been added to existing pages, and old tools for services we no longer operate have not always been removed. Thus, our latest focus has been to improve the user experience of the CRM system for our CS team.
And though on the surface it looks like we're working on a product with only 90 internal users, our changes will have flow on effects to tens of thousands of our members at any given time (from a total number of around 3.6 million members).
The challenges of designing customer service systems
We face unique challenges designing customer service systems. Robert Schumacher from GfK summarizes these problems well. I’ve paraphrased him here and added an issue of my own:
1. Customer service centres are high volume environments — Our CS team has thousands of customer interactions every day, and and each team member travels similar paths in the CRM system.
2. Wrong turns are amplified — With so many similar interactions, a system change that adds a minute more to processing customer queries could slow down the whole team and result in delays for customers.
3. Two people relying on the same system — When the CS team takes a phone call from a customer, the CRM system is serving both people: the CS person who is interacting with it, and the caller who directs the interaction. Trouble is, the caller can't see the paths the system is forcing the CS person to take. For example, in a previous job a client’s CS team would always ask callers two or three extra security questions — not to confirm identites, but to cover up the delay between answering the call and the right page loading in the system.
4. Desktop clutter — As a result of the plethora of tools and reports and systems, the desktop of the average CS team member is crowded with open windows and tabs. They have to remember where things are and also how to interact with the different tools and reports, all of which may have been created independently (ie. work differently). This presents quite the cognitive load.
5. CS team members are expert users — They use the system every day, and will all have their own techniques for interacting with it quickly and accurately. They've also probably come up with their own solutions to system problems, which they might be very comfortable with. As Schumacher says, 'A critical mistake is to discount the expert and design for the novice. In contact centers, novices become experts very quickly.'
6. Co-design is risky — Co-design workshops, where the users become the designers, are all the rage, and are usually pretty effective at getting great ideas quickly into systems. But expert users almost always end up regurgitating the system they're familiar with, as they've been trained by repeated use of systems to think in fixed ways.
7. Training is expensive — Complex systems require more training so if your call centre has high churn (ours doesn’t – most staff stick around for years) then you’ll be spending a lot of money. …and the one I’ve added:
8. Powerful does not mean easy to learn — The ‘it must be easy to use and intuitive’ design rationale is often the cause of badly designed CRM systems. Designers mistakenly design something simple when they should be designing something powerful. Powerful is complicated, dense, and often less easy to learn, but once mastered lets staff really motor.
Our project focus
Our improvement of Trade Me Admin is focused on fixing the shattered IA and restructuring the key pages to make them perform even better, bringing them into a new code framework. We're not redesigning the reports, tools, code or even the interaction for most of the reports, as this will be many years of effort. Watching our own staff use Trade Me Admin is like watching someone juggling six or seven things.
The system requires them to visit multiple pages, hold multiple facts in their head, pattern and problem-match across those pages, and follow their professional intuition to get to the heart of a problem. Where the system works well is on some key, densely detailed hub pages. Where it works badly, staff have to navigate click farms with arbitrary link names, have to type across the URL to get to hidden reports, and generally expend more effort on finding the answer than on comprehending the answer.
Groundwork
The first thing that we did was to sit with CS and watch them work and get to know the common actions they perform. The random nature of the IA and the plethora of dead links and superseded reports became apparent. We surveyed teams, providing them with screen printouts and three highlighter pens to colour things as green (use heaps), orange (use sometimes) and red (never use). From this, we were able to immediately remove a lot of noise from the new IA. We also saw that specific teams used certain links but that everyone used a core set. Initially focussing on the core set, we set about understanding the tasks under those links.
The complexity of the job soon became apparent – with a complex system like Trade Me Admin, it is possible to do the same thing in many different ways. Most CRM systems are complex and detailed enough for there to be more than one way to achieve the same end and often, it’s not possible to get a definitive answer, only possible to ‘build a picture’. There’s no one-to-one mapping of task to link. Links were also often arbitrarily named: ‘SQL Lookup’ being an example. The highly-trained user base are dependent on muscle memory in finding these links. This meant that when asked something like: “What and where is the policing enquiry function?”, many couldn’t tell us what or where it was, but when they needed the report it contained they found it straight away.
Sort of difficult
Therefore, it came as little surprise that staff found the subsequent card sort task quite hard. We renamed the links to better describe their associated actions, and of course, they weren't in the same location as in Trade Me Admin. So instead of taking the predicted 20 minutes, the sort was taking upwards of 40 minutes. Not great when staff are supposed to be answering customer enquiries!
We noticed some strong trends in the results, with links clustering around some of the key pages and tasks (like 'member', 'listing', 'review member financials', and so on). The results also confirmed something that we had observed — that there is a strong split between two types of information: emails/tickets/notes and member info/listing info/reports.
We built and tested two IAs
After card sorting, we created two new IAs, and then customized one of the IAs for each of the three CS teams, giving us IAs to test. Each team was then asked to complete two tree tests, with 50% doing one first and 50% doing the other first. At first glance, the results of the tree test were okay — around 61% — but 'Could try harder'. We saw very little overall difference between the success of the two structures, but definitely some differences in task success. And we also came across an interesting quirk in the results.
Closer analysis of the pie charts with an expert in Trade Me Admin showed that some ‘wrong’ answers would give part of the picture required. In some cases so much so that I reclassified answers as ‘correct’ as they were more right than wrong. Typically, in a real world situation, staff might check several reports in order to build a picture. This ambiguous nature is hard to replicate in a tree test which wants definitive yes or no answers. Keeping the tasks both simple to follow and comprehensive proved harder than we expected.
For example, we set a task that asked participants to investigate whether two customers had been bidding on each other's auctions. When we looked at the pietree (see screenshot below), we noticed some participants had clicked on 'Search Members', thinking they needed to locate the customer accounts, when the task had presumed that the customers had already been found. This is a useful insight into writing more comprehensive tasks that we can take with us into our next tests.
What’s clear from analysis is that although it’s possible to provide definitive answers for a typical site’s IAs, for a CRM like Trade Me Admin this is a lot harder. Devising and testing the structure of a CRM has proved a challenge for our highly trained audience, who are used to the current system and naturally find it difficult to see and do things differently. Once we had reclassified some of the answers as ‘correct’ one of the two trees was a clear winner — it had gone from 61% to 69%. The other tree had only improved slightly, from 61% to 63%.
There were still elements with it that were performing sub-optimally in our winning structure, though. Generally, the problems were to do with labelling, where, in some cases, we had attempted to disambiguate those ‘SQL lookup’-type labels but in the process, confused the team. We were left with the dilemma of whether to go with the new labels and make the system initially harder to use for staff but easier to learn for new staff, or stick with the old labels, which are harder to learn. My view is that any new system is going to see an initial performance dip, so we might as well change the labels now and make it better.
The importance of carefully structuring questions in a tree test has been highlighted, particularly in light of the ‘start anywhere/go anywhere’ nature of a CRM. The diffuse but powerful nature of a CRM means that careful consideration of tree test answer options needs to be made, in order to decide ‘how close to 100% correct answer’ you want to get.
Development work has begun so watch this space
It's great to see that our research is influencing the next stage of the CRM system, and we're looking forward to seeing it go live. Of course, our work isn't over— and nor would we want it to be! Alongside the redevelopment of the IA, I've been redesigning the key pages from Trade Me Admin, and continuing to conduct user research, including first click testing using Chalkmark.
This project has been governed by a steadily developing set of design principles, focused on complex CRM systems and the specific needs of their audience. Two of these principles are to reduce navigation and to design for experts, not novices, which means creating dense, detailed pages. It's intense, complex, and rewarding design work, and we'll be exploring this exciting space in more depth in upcoming posts.
In Part 1 of this series we looked at how to interpret results from open and hybrid card sorts and now in Part 2, we’re going to talk about closed card sorts. In closed card sorts, participants are asked to sort the cards into predetermined categories and are not allowed to create any of their own. You might use this approach when you are constrained by specific category names or as a quick checkup before launching a new or newly redesigned website.In Part 1, we also discussed the two different - but complementary - types of analysis that are generally used together for interpreting card sort results: exploratory and statistical. Exploratory analysis is intuitive and creative while statistical analysis is all about the numbers. Check out Part 1 for a refresher or learn more about exploratory and statistical analysis in Donna Spencer’s book.
Getting started
Closed card sort analysis is generally much quicker and easier than open and hybrid card sorts because there are no participant created category names to analyze - it’s really just about where the cards were placed. There are some similarities about how you might start to approach your analysis process but overall there’s a lot less information to take in and there isn’t much in the way of drilling down into the details like we did in Part 1.Just like with an open card sort, kick off your analysis process 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. Does anything jump out as surprising? Are there similarities or differences between participant sorts?
If you’re redesigning an existing information architecture (IA), how do your results compare to the current state? If this is a final check up before launching a live website, how do these results compare to what you learned during your previous research studies?If you ran your card sort using information architecture tool OptimalSort, head straight to 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, you’ve probably been scanning them in after each completed session, but 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. 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 how the cards were sorted into your predetermined categories. 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.Once you’re happy with the individual card sorts that will and won’t be included in your results visualizations, it’s time to take a look at the Results Matrix in OptimalSort. The Results Matrix shows the number of times each card was sorted into each of your predetermined categories- the higher the number, the darker the shade of blue (see below).
Results Matrix in OptimalSort.
This table enables you to quickly and easily get across how the cards were sorted and gauge the highest and lowest levels of agreement among your participants. This will tell you if you’re on the right track or highlight opportunities for further refinement of your categories.If we take a closer look (see below) we can see that in this example closed card sort conducted on the Dewey Decimal Classification system commonly used in libraries, The Interpretation of Dreams by Sigmund Freud was sorted into ‘Philosophy and psychology’ 38 times in study a completed by 51 participants.
Results Matrix in OptimalSort zoomed in with hover.
In the real world, that is exactly where that content lives and this is useful to know because it shows that the current state is supporting user expectations around findability reasonably well. Note: this particular example study used image based cards instead of word label based cards so the description that appears in both the grey box and down the left hand side of the matrix is for reference purposes only and was hidden from the participants.Sometimes you may come across cards that are popular in multiple categories. In our example study, How to win friends and influence people by Dale Carnegie, is popular in two categories: ‘Philosophy & psychology’ and ‘Social sciences’ with 22 and 21 placements respectively. The remaining card placements are scattered across a further 5 categories although in much smaller numbers.
Results Matrix showing cards popular in multiple categories.
When this happens, it’s up to you to determine what your number thresholds are. If it’s a tie or really close like it is in this case, you might review the results against any previous research studies to see if anything has changed or if this is something that comes up often. It might be a new category that you’ve just introduced, it might be an issue that hasn’t been resolved yet or it might just be limited to this one study. If you’re really not sure, it’s a good idea to run some in-person card sorts as well so you can ask questions and gain clarification around why your participants felt a card belonged in a particular category. If you’ve already done that great! Time to review those notes and recordings!You may also find yourself in a situation where no category is any more popular than the others for a particular card. This means there’s not much agreement among your participants about where that card actually belongs. In our example closed card sort study, the World Book Encyclopedia was placed into 9 of 10 categories. While it was placed in ‘History & geography’ 18 times, that’s still only 35% of the total placements for that card- it’s hardly conclusive.
Results Matrix showing a card with a lack of agreement.
Sometimes this happens when the card label or image is quite general and could logically belong in many of the categories. In this case, an encyclopedia could easily fit into any of those categories and I suspect this happened because people may not be aware that encyclopedias make up a very large part of the category on the far left of the above matrix: ‘Computer science, information & general works’. You may also see this happening when a card is ambiguous and people have to guess where it might belong. Again - if you haven’t already - if in doubt, run some in-person card sorts so you can ask questions and get to the bottom of it!After reviewing the Results Matrix in OptimalSort, visit the Popular Placements Matrix to see which cards were most popular for each of your categories based on how your participants sorted them (see below 2 images).
Popular Placements Matrix in OptimalSort- top half of the diagram.
Popular Placements Matrix in OptimalSort- scrolled to show the bottom half of the diagram.
The diagram shades the most popular placements for each category in blue making it very easy to spot what belongs where in the eyes of your participants. It’s useful for quickly identifying clusters and also highlights the categories that didn’t get a lot of card sorting love. In our example study (2 images above) we can see that ‘Technology’ wasn’t a popular card category choice potentially indicating ambiguity around that particular category name. As someone familiar with the Dewey Decimal Classification system I know that ‘Technology’ is a bit of a tricky one because it contains a wide variety of content that includes topics on medicine and food science - sometimes it will appear as ‘Technology & applied sciences’. These results appear to support the case for exploring that alternative further!
Where to from here?
Now that we’ve looked at how to interpret your open, hybrid and closed card sorts, here are some next steps to help you turn those insights into action!Once you’ve analyzed your card sort results, it’s time to feed those insights into your design process and create your taxonomy which goes hand in hand with your information architecture. You can build your taxonomy out in Post-it notes before popping it into a spreadsheet for review. This is also a great time to identify any alternate labelling and placement options that came out of your card sorting process for further testing.From here, you might move into tree testing your new IA or you might run another card sort focussing on a specific area of your website. You can learn more about card sorting in general via our 101 guide.
When interpreting card sort results, don’t forget to have fun! It’s easy to get overwhelmed and bogged down in the results but don’t lose sight of the magic that is uncovering user insights.I’m going to leave you with this quote from Donna Spencer that summarizes the essence of card sort analysis quite nicely:Remember that you are the one who is doing the thinking, not the technique... you are the one who puts it all together into a great solution. Follow your instincts, take some risks, and try new approaches. - Donna Spencer
Further reading
Card Sorting 101 – Learn about the differences between open, closed and hybrid card sorts, and how to run your own using OptimalSort.
