| name | run-card-sorting |
| description | Use when designing or evaluating information architecture — navigation structure, menu labels, content categorization — to align the organization with how users mentally group and label concepts. |
| source | NNG "Card Sorting" guideline; Spencer "Card Sorting: Designing Usable Categories" (Rosenfeld Media, 2009); IBM Design Thinking IA practices |
| tags | ["user-research","information-architecture","card-sorting","navigation","ia","ux-research","mental-models"] |
Run Card Sorting
Have users group and label concepts on cards to reveal how they mentally organize the subject matter — then use those patterns to structure navigation, menus, and content hierarchies.
Why This Is Best Practice
Adopted by: NNG positions card sorting as the primary research method for information architecture decisions; IBM Design Thinking uses open card sorting as the standard IA input method; OptimalSort (Optimal Workshop), Maze, and UserZoom include card sorting as a core research tool, reflecting widespread adoption across product and UX teams
Impact: NNG: card sorting with 15 users identifies 80%+ of IA problems before development; Spencer (2009) documents that navigation redesigns informed by card sorting reduce findability failures by 40–60% compared to designer-led IA; fixing IA post-launch — after content, nav, and URLs are established — costs 10–100× more than validating IA during design
Why best: Designer-led IA reflects the organization's internal structure (product features, team ownership) rather than users' mental models; analytics show where users go but not why they struggle; only card sorting reveals the categories users expect and the labels they use — inputs that cannot be inferred from usage data or expert review alone
