Survey Experience
January 12, 2025
The Voice of the Partner: How Independent Facilitation Transformed Global Strategy for a Tech Leader
March 17, 2025Survey Experience
January 12, 2025The Voice of the Partner: How Independent Facilitation Transformed Global Strategy for a Tech Leader
March 17, 2025Insights
Card Sorting Isn’t Just a UX Exercise
Using Mental Models to Build Better Structures for Taxonomy, Content, and Features
Card sorting gets a lot of airtime in the UX world, typically as a method to validate navigation or group labels on a website. But its value extends much further than that. At its core, card sorting is a way to uncover how people expect information to be structured. It surfaces their mental models, their instincts about what belongs together and why.
We’ve used card sorting for things like organizing learning content, clarifying product feature sets, and, in one case, exploring how people naturally categorize professional skills. While none of these are traditional UX challenges, they all involve the same problem: we were building something for people to use, and our internal logic wasn’t enough.
Card sorting can interrupt that bubble.
Whether you’re using an open sort (where people group freely) or a closed one (where you give the categories), the act of seeing how others chunk and label your content can reveal patterns you never expected, and gaps you didn’t realize you’d created.
But card sorting isn’t magic. It creates messy data. People label things inconsistently. They use different logic. They see connections you hadn’t considered. That mess is where the value is, but it’s also where synthesis gets hard.
To make use of it, you need to look for clustering patterns across respondents, not just what one person did. You need to pay attention to language overlap and category shapes. Some groupings will emerge clearly. Others may feel fractured or ambiguous. Not everything has to fit neatly, and trying to force edge cases into place often weakens the overall structure.
One small but important note:
When you're testing the structure that emerges from your synthesis, don’t just rely on researchers or real respondents. Real respondents rarely tell you what feels repetitive or cognitively taxing, because you’re only seeing their choices, not the thought process behind them. On the other hand, research teammates are often too close to the material. The sweet spot? Test with colleagues who aren’t part of the project but know enough to ask: “Wait, didn’t I already see this?” or “What’s the difference between these two?”
At its best, card sorting gives you the raw material for a structure that reflects how people actually think, not just how we want them to think. If you’re building a system that people need to move through, whether it’s a content hub, a feature framework, or something else entirely, it’s worth starting there.