Card sorting is a design research tool that helps you design your product (e.g., a website, an app) so users can more easily find what they want. Card sorting is usually used near the beginning of the design process, and can be used to guide the design of the navigation menu and the overall structure (information architecture) of the product.

Why Card Sorting?

Card sorting is effective and simple. There are many reasons to use it:

  • Easy to conduct (can be set up by anyone!)
  • Inexpensive (can be done with just paper and pens!)
  • Quick (can see patterns from as little as 5 card sorts!)
  • Powerful insights and results (can reveal a lot about how users think!)

Card sorting is all about bridging the gap between how a user understands the subject and how a website or app can match that understanding and present information or functionality to users in a way that’s easy to find.

How does Card Sorting work?

The process of a card sort is exactly what it sounds like - participants (ideally, people who would normally use the website or app) categorize things written on cards into different groups that just simply make sense to them. The cards contain either content or functionality that the website or app would have.

Card sorts can be open or closed:

  Open Card Sort Closed Card Sort
What is it? Participants generate their own categories Participants sort cards into pre-defined categories
Useful for Generating ideas as inputs for a potential structure Refining ideas; when you want to add content to an existing information architecture; or when categories aren’t changeable
Challenges Analysis will likely take more effort, given the various participant-generated categories May not represent an overall structure that best matches participants’ expectations (since categories were pre-defined)

Card sorts can also be team or individual:

  Team Card Sort Individual Card Sort
What is it? Participants sort cards in small groups of 2-3 people Participants sort cards on their own
Advantages Allows for more discussion and ideas between participants that reveal valuable insight Faster to complete; more data points can be taken; ensures individual opinion is recorded; logistically easier to organize, especially if participants complete the sort online
Challenges One person could dominate the discussion, and bias the sort towards their point of view. To counteract this, the person running the card sort should moderate discussion so that everyone has a turn to speak. May miss valuable insight and discussion, especially if using online tools. To counteract this, include follow-up questions asking participants to explain how they got to their final sort, and any cards that they found tricky to put in a particular category.

Card sorting can return valuable insights using any combination of open vs. closed and team vs. individual, so don’t let scheduling or other logistical challenges stop you from giving it a try!

Learning from Card Sorting

Last week, Yang Chen led a Boltmade Session to explore these concepts. We had a lot of fun learning and discussing how card sorting can be used as an effective UX research tool, and also performing a couple of card sorts ourselves.

Cardsorting activity1

Activity #1: Museum Sort (an open, team sort)

With the first activity, we were able to get a sense of what a real participant would experience in a card sort. In teams of 2 or 3, everyone sorted a set of 54 cards that had been pre-printed with the names of museum exhibits such as Chinese Architecture, Terry Fox, and Bronze Age. Because this was an open card sort, teams also had to generate their own categories to put cards under.

Many good discussions and questions came up during the sort:

  • How should cards be sorted? By chronology? By subject? By type of exhibit?
  • Where should Kid’s Zone go, since it doesn’t quite fit with anything else?
  • Should World War I be categorized under World History or under Canada? Who is the museum for, and what does the exhibit actually cover?

All of these struggles and debates about which cards to group together are perfectly normal during a card sort, as there is no one right answer. These discussions that happen during the activity of sorting are very useful when working on a proposed design; much of it might have been missed if we had only looked at the final outcome of the sort.

Key takeaway: The process of card sorting and the reasonings that come out of it are as valuable as the product of the groupings that are generated as a result.

Cardsorting activity1 result
An example of a card sort result. Note that the discussion and rationale for why certain cards are in certain categories (not captured in this photo) are also very valuable!

Activity #2: Website Sort (a closed, individual sort)

In our second activity, we saw how a card sort done individually and online is different from a team, in-person sort. We also saw how a closed card sort can be used to improve an existing site.

Using a tool called OptimalSort, we sorted content from the website for Banff National Park into predefined categories that were taken from the existing website. This card sort was very easy to analyze because the online tool did all the work automatically. At a glance, it was easy to see that some content on the existing site didn’t match our expectations. For example, most of us grouped Top 10 Things to Do into the Activities category, but on the actual website it is under Visitor Information, where few of us placed it.

Cardsorting activity2 result
Partial results from OptimalSort’s analysis tool. Note that only 9% of participants put “Top 10 Things” under Visitor Information, where it’s currently located.

We speculated that Top 10 had been placed under Visitor Information because website visitors are likely to explore that section and thus be more likely to discover it. Interestingly, on the current website, there are no sub-categories under Activities. Card sorting results suggest that either the website navigation structure should be reconsidered, or alternatively, the Activities section could be removed so that users don’t try to find activity information under it when it’s not actually there.

Key takeaways: Card sorting results should not be taken at face value as the final answer. It is one data source to consider, and should be triangulated with other information, criteria, and informed intuition when creating a proposed design. Follow-up evaluation such as usability testing is also important.

Card Sorting Key Points

  • Card sorting is simple and effective
  • The discussion that happens during a card sort is just as or sometimes even more valuable than the actual outcome of the card sort
  • The results of a card sort are a great data point, but should be considered along with other research and informed intuition

Further Reading and Resources

Thanks to all who came out to the workshop! Our next Boltmade Session, Amazingly Awesome Ansible, is on April 21st and will introduce automating apps and IT infrastructure with Ansible. Hope to see you then!