Card Sort questions and usability testing: Best practices
Follow these tips to make the most of Card Sort questions in your usability testing workflows.
Conduct Card Sort activities early
Using Card Sort early in the design and development process increases efficiency and decreases the likelihood of expensive rework later. You ensure that user expectations are met before further investment.
For example, let's say you are redesigning the navigation experience on your company's website. You want to understand which pages users expect to find under specific headings. Conducting a Card Sort activity as the first step in the design process establishes the heading/page associations that resonate most with users. Using this information, you can begin work on a prototype with confidence.
Conduct Card Sort activities before Tree Test activities
- Card Sort tests for idea association (which ideas users associate with each category) and how users conceptualize information.
- Tree Test evaluates users' first clicks (which heading or node a user clicks first when looking for something specific), the findability of information, and the user friendliness of a hierarchical navigation structure.
For this reason, we recommend doing a card sorting exercise before a tree test activity in usability testing. Use the category and item groupings from the card sorting exercise to design the navigation structures that you'll evaluate in the tree test.
Consult existing resources for card and category content
- A brainstorm or mind map
- Internal documentation such as a sitemap, org chart, or product inventory list
- A content audit of a website or knowledge base
- A list of ideas from stakeholders
- A list of feedback or suggestions from customers
- A list of ideas compiled from competitor research
Create groupable cards
The cards need to be on the same level, hierarchically speaking. For example, if you have cards that name different types of cookware, servingware, kitchen appliances, and kitchen utensils and you are asking participants to sort them into categories, you should not also have a card that says "Cookware."
Likewise, cards should overlap enough in conceptual similarity so that participants can instinctively cluster several of them together. Going back to the example above, there should be cards that show a frying pan, stock pot, and saucepan that users can easily conceptualize and group together in a "Cookware" category.
Optimize the Card Sort user experience
- We recommend using 3-6 categories, ideally closer to 4. This way, participants don't have to scroll too much to view all the categories that they're sorting cards into.
- Be explicit in instructions that there are no right or wrong answers, it's about how participants perceive things.
- In large Card Sort questions, we recommend not including a "I'm not sure" or "I don't know" category where appropriate. Having these types of categories can result in too many "unsorted" cards. Instead, advise participants that if something fits in more than one bucket, they should put it in the "best fit" category.
- Large Card Sort questions with many cards and categories can be difficult to complete on mobile devices. If you expect the majority of responses to come from mobile participants, keep the Card Sort questions on the smaller side. Show participants a subset of cards in random or rotated order. Alternately, consider ways you can chunk a large Card Sort question into a series of smaller ones.
- For larger Card Sort questions, we recommend limiting the number of cards to 30-50 and the estimated completion time to 10-15 minutes.
- Be mindful of the overall cognitive load you're asking of participants. If the Card Sort exercise is on the simpler side (for example, you are asking them to sort cards into Yes/No/Maybe categories), you can have as many as 60+ cards. However, if there are more categories and the cards require more thoughtful reflection to sort properly, follow the sizing recommendation provided above.