Card Sorting
Card sorting shows how people *actually* think about information. Not how you *assume* they do. Give participants cards with content or features. Ask them to group related items. You'll see mental models you never expected. The real value isn't the final categories. It's watching someone struggle with a card, hearing them explain why 'Settings' could be in multiple places. You'll realize your navigation only makes sense to you.
- Learn how users naturally group information.
- Find terminology that matches user expectations.
- Spot confusing parts of your information architecture.
- Build navigation based on real mental models.
- Uncover disagreements about categories before launch.
- Information architecture reflecting user mental models.
- Validated category labels and navigation.
- Insights into how different user groups think.
Staying quiet is the hardest part. You'll want to explain. Don't. If they ask, shrug and say, 'Whatever you think it is.' Confusion is valuable.
Analyze agreement and disagreement. 12/15 putting 'Billing' with 'Account' is a clear signal. A three-way split is a problem. Create a similarity matrix showing grouping frequency. Tools like OptimalSort do this.
Open sorts (participants create categories) are best for early discovery. Closed sorts (you provide categories) are best for validating a structure. Open first, then closed.
Watch for 'Miscellaneous' piles. These cards need attention. Labels may be confusing, or items don't fit. Solve these problems.
For qualitative insights, 5-8 participants reveal patterns. For statistical confidence, aim for 15-30. More than 15 rarely adds much.
Start the conversation
Be the first to share your thoughts, experiences, or questions!