The act of card sorting is a hands-on and quick way of creating a category tree (information architecture) based on users’ understanding of the topics of a product. The simple idea of the workshop is to ask users to organise cards that already contain written content into groups.
In this tutorial you will learn how to plan, execute, record and analyse the card sorting as if you were doing it yourself. We’ll also discuss why card sorting is a good UX tool, and how it can benefit you and your organisation.
1. Plan Your Card Sort
What is Card Sorting?
The first step is to decide which type of card sort is going to be most beneficial, since there are several different kinds. The most common are open, closed, and a hybrid of the two. They can then be handled as remote, in-person, or by group.
Open Card Sort
An open card sort allows the users to create their own group names based on the kind of cards they have grouped together. This is beneficial when wanting to understand how the users would structure the content if they didn’t know the main topics of the website or product. It allows you to think of new ways of approaching the hierarchy of the product and can give creative insights to the structure that makes more sense to users than what the organisation would have previously thought of.
In this card sort, the participants sort the cards with written content, topics, labels, or pages in the order that makes sense to them. They can choose to remove certain cards that don’t fit in and also add new cards that would match the context better. For example, if there’s a card with the wording “contact form” written on it and another one with the wording “address”, it would make sense for the participant to add a new card on which they write “Google map”.
After they’re happy with the card sort, they write a name for each group that’s been created on a card and place it above the relevant group.
Closed Card Sort
A closed card sort works in the same way as an open one except that the group names are defined beforehand and the participant can’t change them. Neither are they allowed to add their own cards to the sort.
This works best if you have a website or a product that already has main topics that work well. For example, a product that has already conducted user research on their main topics and received beneficial results, or a website where the organisation doesn’t have a budget to change their core topics but are able to change the sorting of the content.
Hybrid Card Sort
If you have the time, I would recommend mixing the two and starting with an open workshop to uncover all the potential groups the users connect the content to. Then run a closed workshop with the predefined categories the organisation has chosen for this workshop to see if the ideas could be understood by the audience.
A hybrid card sort has a few defined group names but still lets the participant add their own.
In-person Card Sort
Once you’ve decided which type of card sort might work best, it’s time to look at the location and the participants of the workshop. In-person card sorts work best if you have easy access to a user group in the same area. I prefer to do card sort workshops in person since they allow the participants to explain their decisions as they’re making them, giving a deeper insight into how the users will use the product.
Remote Card Sort
A remote card sort works well if the product is used by a broad audience with different geographic locations. In this case, an online card sorting tool is needed. Via tools like OptimalSort or UserZoom the participant can sort cards while sitting at their own computer. It gives great qualitative data but doesn’t explain why a user might have chosen to sort things in a particular way.
Group Card Sort
This is when a group receives the instructions and works together to group the topics. The participants can influence each other’s answers but this can be a good way of receiving a general view.
Conducting a group workshop can be a bit tricky with a lot of different personalities involved and a lot of notes to keep track of. Therefore, it’s best to run a single person card sort as your first card sorting experience and then move to groups, once you feel more comfortable.
Participants
Once you’ve decided which type of card sort is best for the situation, it’s time to contact the participants.
If the organisation has an already established user base on which they run tests, that’s great. If not, it could be worth checking if they have a list of email subscribers or somethings similar with people that match their target group. You can then send out an email to them and ask if they are willing to be a part of a study that benefits how the product will evolve. It’s advisable to offer some form of compensation for this.
If the organisation doesn’t hold such a list, it’s best to go with the remote card sorting option since online card sorting tools often provide a database of users to test on where you can choose the preferred demographic.
The NN group recommends running card sorting on 15 participants to receive results that give you enough data to create an information architecture based on users’ understanding of the product’s topics.
Prepare the Space
- If the card sort is in-person, the space needs to be prepared. Start by buying pens, cards, post-it notes or paper in different colours. The different paper colours will be used for each different group and group name so that it’s easier to differentiate them. Make sure the space has a large empty table or wall so that the cards can be spread out.
- Write down the topics (product categories, article categories, pages in a site map, tags or menu items) with a correlating number on your computer and then do the same on each card. Make sure to put the number at the back of the card so it won’t distract the participant. This will help with the analysis of the results. Keep the card to a maximum of 40 since researchers have found that any more than that makes the participant less engaged. Leave a few cards empty without a topic if it’s an open sort.
- Place the cards on the table and shuffle them around so that they’re not in the numbered order.
- Something that really helps me in these sessions is to record them so that I can watch them later to refresh my memory and help with the analysis. It also helps to write down exactly what I’m going to say in case I miss something. You’d be surprised how easy it is to forget to explain the process properly if you’ve already done it ten times, which can make the participants confused or nervous.
- And of course, don’t forget to prepare snacks and drinks!
2. Execute
Now you’ve prepared everything and the day has come for the actual workshop.
- Start by giving the participant(s) an introduction to why you’re doing the card sort and ask them if it’s okay to record the session. Let them know that it’s important for them to speak loudly and for them to explain why they are putting things in certain categories as much as possible. It’s also important to let them know that there are no right or wrong answers, and if there are a few cards that are unclear, that’s a mistake in the system and not theirs.
- After the introduction you can ask them to start sorting the cards into groups. When they are finished, they can add the group name above the group topics.
- Take notes during the session and don’t try to interfere unless they say something like “umm, hmm” and it isn’t clear as to why they’ve sorted something in a certain way. You can then ask what they are thinking. Don’t try to interrupt the natural process because you might create an unrealistic sort.
3. Record Results
After the sorting has been completed you can bring up questions you might have had before. Questions like:
- Was it difficult?
- Which parts were you unsure about?
- Could we rename some of them to make it easier?
Thank the participant, take a picture of the sort and note which numbers were grouped together. Make sure you label the photo with the participant’s name.
Re-shuffle the cards for the next person!
4. Analyse
When all the sessions have been completed, you’ll probably already have a general understanding of how the users group the topics together and their thought processes behind doing so. Analysis still needs to be run to be able to present the results.
If an online sorting tool was used, there will most likely be an analysis of the data within the tool that you can use. But if not, follow the steps below.
- Since you’ve taken a photo of the sort and recorded which numbers were in which order for each participant, start by sorting them digitally in accordance with the workshop outcomes.
- Add all the groups to a spreadsheet program and start counting how many common denominators there are. For example, count how many participants chose to put “banana” in the group “fruit”, how many participants added the group name “contact” or how many participants added the topic “search” as the first topic in the group “home page”. In this way you’ll find common patterns for the quantitative data.
- Read through your comments to see if there’s something worth adding to the analysis. Alternatively, you can add all the comments to an online text analysis tool like Voyant. Through a tool like this, you can look at the most common sentences, words (it’s possible to exclude short words such as “and, have, no, yes, one” etc.) and specify common thoughts amongst the participants.
- Once the quantitative and qualitative data have been prepared, combine everything in a report or presentation for the purchaser of the card sort.
Conclusion
In a world where it’s very easy for the HIPPO (highest paid person’s opinion) to decide how a website structure should be built without having any knowledge of user preferences, card sorting can be a quick, cost-effective, and easy tool to show real data about what users want and therefore what is most beneficial to the organisation.
It gives a good starting point for the website or product structure, but also gives an insight into whether the users understand the organisation’s path. Based on this, you can embark on the next step in the UX process; the information architecture.