Card Sorting Methodology
Generative user research technique to design or evaluate information architecture.
The Specific Question
How do users expect your information to be grouped?
Card sorting answers this by forcing users to categorize concepts into mental models rather than relying on internal business jargon.
Worked Example: E-commerce Taxonomy
Instead of guessing if "Belts" belongs under "Accessories" or "Menswear", you give 50 users cards. If 80% put them under "Accessories", that becomes your structural truth.
| Type | Use Case | Pros |
|---|---|---|
| Open Sort | Generative (New structures) | Discovers user vocabulary |
| Closed Sort | Evaluative (Testing existing structures) | Validates specific categories |
| Hybrid Sort | Refining | Balances rigidity with discovery |
Common Mistakes
- Using internal acronyms on cards.
- Testing more than 60 cards (causes fatigue).
FAQ
- Q: How many participants do I need?
- A: Statistical significance in card sorting typically plateaus at 15-20 users.