Card Sorting – How companies develop better structures

What is the card sorting method?

Card sorting is a UX research method where participants group content into categories that make sense to them. Originating from cognitive psychology, it reveals users' mental models. Participants are given cards featuring page titles, features, or content types, which they then organize into groups and label. The method can be conducted analogly with physical cards or digitally using specialized tools. Card sorting does not provide ready-made solutions, but rather insights into potential groupings that must be validated through further UX research methods.

‍

Variations of the card sorting method

The card sorting method exists in various forms, each addressing different research goals. The choice of the right variant depends on whether an existing structure is already in place or if an entirely new information architecture needs to be developed.

Open card sorting

In open card sorting, participants receive cards with content but no predefined categories. They group the cards freely and create their own labels for the resulting groups. This variant is particularly suitable when no existing structure is available or when fundamentally new organizational approaches are to be explored. Open card sorting provides insights into the natural thought processes of users and reveals the terminology they use to describe content. It is the most flexible form and allows for the most accurate representation of users' mental models.

Closed card sorting

In closed card sorting, the categories are predefined. Participants sort the cards into these existing groups. This variant is used when an existing navigation structure needs to be reviewed or optimized. It provides data on whether users can assign content to the intended categories or if misunderstandings arise. Closed card sorting is less exploratory but faster to evaluate and is well-suited for validating structural hypotheses.

Hybrid card sorting

Hybrid card sorting combines elements of both approaches: there are predefined categories, but participants can add their own groups if needed. This variant offers a middle ground between structure and flexibility and is suitable for projects where parts of the information architecture are already fixed, while other areas remain open.

The choice of variant should be based on the project status and specific research questions. In practice, it is common to start with open card sorting to gain a broad understanding and subsequently validate with closed card sorting.

‍

Process and methodology

Conducting a card sort requires careful planning. First, cards are created that represent elements from inventory analyses, interviews, or existing websites. The wording of the card labels is crucial: they should be neutral and contain no hints regarding categorization. The number of cards is typically between 30 and 60. The selection of card sorting participants should represent the actual target audience. For quantitative evaluations, 30 to 50 participants are recommended; for qualitative studies, 15 may be sufficient.

Card sorting can be conducted in a moderated or unmoderated format. Moderated sessions provide deeper qualitative data but are more time-consuming. Unmoderated remote studies allow for larger sample sizes and faster data collection. Many companies combine both approaches.

‍

From data to structure

Card sorting analysis uses similarity matrices and dendrograms to reveal patterns. A dendrogram visualizes which cards were sorted together most frequently. It is important to note that dendrograms do not provide ready-made answers, but rather show trends that must be interpreted. Qualitative data from moderated sessions complements the quantitative results and explains why certain groupings were made.

The results are incorporated into the development of navigation structures, menus, and sitemaps. Card sorting is generative and exploratory. The results must subsequently be validated through evaluative methods such as tree testing.

‍

Online Card Sorting Tools: Digital Implementation

Online card sorting tools enable remote studies with large numbers of participants. Platforms such as Optimal Workshop, UXtweak, or Maze offer features for creation, participant management, and automated analysis. The advantage lies in scalability: participants can join from anywhere, data is captured automatically, and results are available immediately. Physical card sorts still have their place for specific target groups or during early exploratory phases.

‍

Pros and Cons of Card Sorting

The card sorting method offers significant advantages, but it also has clear limitations that must be considered during the planning phase.

Pros

• Directly involves users in the design process
• Makes mental models visible that would otherwise remain implicit
• Reduces the risk of developing structures based solely on an internal company perspective
• Relatively inexpensive and quick to conduct, especially with digital tools
• Provides both quantitative data for statistical analysis and qualitative insights into thought processes
• Flexible enough to be used from early concept phases to the optimization of existing structures

Cons

• Results require interpretation: users rarely sort in a uniform way
• Deciding on prioritized patterns requires expertise and further research
• Shows only how users group content, not whether they can actually find it
• Supplementary methods like tree testing are necessary
• Often provides only superficial insights; deeper contextual connections may remain undiscovered
• Quality depends heavily on the wording of the card labels and the selection of participants
• Limited effectiveness for complex hierarchies
• Iterative approaches are required for deeply nested structures

‍

Card sorting in the design phase: Strategic application

The design phase is the ideal time to use card sorting. At this stage, structural decisions are still open, and the cost of making changes is low. The process typically follows an iterative approach: start with open card sorting for a broad understanding, then validate with closed or hybrid card sorting. A common mistake is to view card sorting in isolation. The method is most effective when embedded in a comprehensive research process and combined with tree testing.

‍

Developing user-friendly navigation

The goal of every card sorting study is to develop user-friendly navigation. The results form the basis for decisions regarding menu structures, category names, and information hierarchies. Translating these into concrete design decisions requires balancing user needs, business goals, and technical feasibility. User-centered information architecture is created by intelligently interpreting data within the context of the entire project. Optimizing navigation also means continuously questioning existing structures. Reverse card sorting can reveal where structures no longer align with current mental models.

‍

Card sorting cost-benefit: An economic perspective

The costs for card sorting are manageable. Digital tools are often available as part of UX research platforms. Participant compensation is typically in the low to mid double-digit range per person. Overall, costs are often in the low four-figure range, which is significantly less than the cost of a redesign due to misguided structural decisions.

The benefit lies in risk reduction. An information architecture based on actual user needs reduces support requests, increases conversion rates, and improves user satisfaction. The ROI of card sorting is difficult to quantify precisely, as the benefits are often indirect. However, studies show that investments in early UX research pay off many times over.

‍

Conclusion

Card sorting is a proven tool for developing user-centered information architecture and creating user-friendly navigation. It makes users' mental models visible and provides a sound foundation for structural decisions. The various types—open, closed, and hybrid card sorting—allow for flexible adaptation to different project contexts. Execution is efficiently scalable with modern online card sorting tools, and the analysis provides both quantitative patterns and qualitative insights. Card sorting is particularly valuable in the design phase when fundamental decisions about improving website structure need to be made. The method has clear limitations; it does not replace other UX research methods and requires careful interpretation, but in combination with tree testing and usability testing, it creates a comprehensive picture. The card sorting cost-benefit ratio is favorable, as the investment reduces risks and contributes to long-term improvements in navigation usability. Companies that use card sorting strategically create information architectures that not only seem logical but actually work.

‍