What Is Card Sorting?
Card sorting is a research method in which participants organize a set of labeled cards, representing topics, features, products, or concepts, into groups that make sense to them. The method reveals how people naturally categorize information, exposing the mental models that underlie their expectations about where things belong and how they relate. Card sorting is a cornerstone of information architecture (IA) and UX research, used to design navigation structures, menus, and content hierarchies that align with how users actually think rather than how organizations internally structure their offerings. The technique also has broader applications in market research, knowledge management, and organizational studies wherever understanding how people classify and relate concepts matters.
Why Card Sorting Matters in Research
The way researchers or product teams organize information often doesn't match how their audience thinks about it. Card sorting closes that gap by grounding structural decisions in user data. Studies consistently show that navigation and content structures based on card sort data outperform those designed by internal teams alone, users find what they're looking for faster and with fewer errors.
How Card Sorting Works
There are three main variants, each serving a different purpose.
Open Card Sort
Participants receive cards with no predefined categories and create their own groups, naming each one. This exploratory approach reveals how many natural categories exist, what they're called, and which items people associate together. Open sorts are best for early-stage research when you don't yet have a category structure.
Closed Card Sort
Participants receive both the cards and a set of predefined category labels, and sort the cards into those existing groups. This evaluative approach tests whether a proposed structure matches user expectations. Closed sorts are best for validating a draft navigation, taxonomy, or organizational scheme.
Hybrid Card Sort
Participants sort cards into predefined categories but can also create new categories if they feel an item doesn't fit anywhere. This captures the best of both approaches, you test your existing structure while leaving room for categories you may have missed.
Running a Card Sort
A typical study involves 30 to 60 cards and 20 to 50 participants. In-person sorts use physical index cards on a table; remote sorts use digital platforms where participants drag and drop cards into groups on screen. Sessions take 15 to 30 minutes per participant. After sorting, many researchers ask participants to think aloud about their reasoning or complete a brief follow-up questionnaire explaining their choices.
Analyzing Card Sort Data
Analysis focuses on a similarity matrix, a table showing how frequently each pair of cards was sorted into the same group. Pairs with high co-occurrence clearly belong together; pairs that are never grouped together probably belong in different categories. Cluster analysis (typically hierarchical clustering with a dendrogram visualization) identifies natural groupings in the data. For open sorts, researchers also analyze the category labels participants created, looking for consensus around naming conventions.
When to Use Card Sorting
- Designing or redesigning website navigation to ensure the menu structure reflects how users categorize content
- Organizing product catalogs or feature sets in a way that matches customer mental models
- Testing taxonomy proposals before committing development resources to building them
- Understanding how different audience segments categorize the same information: experts vs. Novices, for example, often have very different mental models
- Structuring knowledge bases, help centers, or documentation so users can self-serve effectively
Common Mistakes to Avoid
- Using too many cards (over 60-70), which produces fatigue and inconsistent sorting, participants start making arbitrary decisions toward the end rather than thoughtful ones
- Writing ambiguous or jargon-heavy card labels that participants interpret differently, producing sorting variation that reflects confusion about the labels rather than genuine differences in mental models
- Treating card sort results as definitive architecture without triangulating with other data, card sorts show how people group information in an abstract task, but navigation behavior on a live site involves additional factors like visual hierarchy, context, and task motivation
How Quali-Fi Supports Card Sorting
Quali-Fi's survey platform includes a dedicated card sort question type that supports open, closed, and hybrid variants on desktop and mobile. Participants drag and drop cards into groups within the survey interface, and the platform captures the full sorting matrix for export. Combined with follow-up open-ended questions and demographic data from the same survey, research teams can segment card sort results by audience type without needing a separate tool.
Frequently Asked Questions
How many participants do you need for a card sort?
For open card sorts, 20 to 30 participants typically produce stable category structures. Closed sorts can work with fewer, 15 to 20, since you're testing an existing structure rather than generating categories from scratch. If you're comparing segments (e.g., new vs. Returning customers), you'll need those numbers per segment.
What's the difference between card sorting and tree testing?
Card sorting asks participants to group items into categories, it's generative and reveals mental models. Tree testing asks participants to find specific items within a proposed hierarchy, it's evaluative and measures findability. They're complementary: use card sorting to design the structure, then tree testing to validate that users can navigate it successfully.
Can card sorting be combined with other methods?
Frequently. Card sorting pairs well with think-aloud protocols (to understand reasoning behind sorting decisions), concept mapping (to add dimensionality beyond simple grouping), and usability testing (to validate that the resulting structure works in practice). It's also common to follow a card sort with a survey that tests the proposed categories at larger scale.
Related Topics
- Affinity Diagramming
- Concept Mapping
- Q Methodology
- Think-Aloud Protocol
- Construct Validity
- Research Design
Need to run card sorts at scale? See how Quali-Fi's built-in card sort question type handles open, closed, and hybrid sorts within your survey workflow.