Uncover user mental models for organizing information to create intuitive navigation and content hierarchies.
Card Sorting reveals how users mentally organize information by asking them to sort labeled cards into groups, informing navigation design.
Card Sorting is a foundational user research technique that reveals how people naturally organize, categorize, and label information. During a session, participants are presented with a set of labeled cards -- each representing a piece of content, feature, or concept -- and asked to sort them into groups that make sense to them. UX designers, information architects, and content strategists use card sorting to create navigation structures, taxonomies, and content hierarchies that align with user mental models rather than internal organizational logic. The method comes in three formats: open sorting where participants create and name their own categories, closed sorting where predefined categories are provided, and hybrid sorting that combines both approaches. Card sorting can be conducted in person or remotely using specialized software, making it scalable from small qualitative studies to large quantitative analyses. It is especially valuable when designing new websites, restructuring existing information architecture, or validating proposed navigation before development begins. The insights help teams make evidence-based decisions about how to organize content so that users can find what they need efficiently.
Determine the goals of your card sorting session, such as identifying user expectations, understanding how users categorize information, or improving the navigation structure of a website or application.
Recruit a diverse group of participants who represent your target user demographic to ensure a wide range of perspectives in the card sorting exercise.
Develop a set of cards that include the key concepts, categories, or tasks that are relevant to your design project. These cards should be brief, clear, and written in the user's language.
Decide between an open, closed, or hybrid card sorting format. In open sorting, participants create and label their own categories; in closed sorting, pre-defined categories are provided; and in hybrid sorting, a mix of both approaches is used.
Facilitate the card sorting session by providing clear instructions to participants, ensuring they understand the goals of the exercise, and encouraging them to think aloud while organizing the cards. Monitor their progress and ask clarifying questions if needed.
Document the categories and card order created by each participant, either by taking photographs, recording the session, or transcribing the results. If using a digital card sorting tool, the data will be captured automatically.
Analyze the card sorting results by looking for patterns and trends in the way participants grouped and labeled the cards. Identify common themes and outliers and consider how these findings align with your design objectives.
Interpret the findings to gain insights into user mental models, expectations, and preferred content organization. Use these insights to inform your design decisions, such as the labeling and structure of navigation menus, or the organization of content within your website or application.
Apply the insights from the card sorting session to your design and conduct additional testing with users to ensure that the new organization aligns with their expectations and is easy to navigate.
After conducting card sorting sessions, your team will have clear data on how users naturally group and categorize your content. You will receive similarity matrices and dendrograms showing which items users consistently group together, category labels that reflect user language, and a proposed information architecture grounded in user mental models. The analysis will reveal areas of strong agreement among participants as well as contentious items that users categorize inconsistently. Teams typically use these results to create or refine sitemaps, navigation menus, and content taxonomies that are intuitive for end users, significantly reducing the risk of building an information architecture that only makes sense to internal stakeholders.
Card sorting is a flexible method that works as a complement to qualitative research or as a standalone quantitative method.
Use online card sorting tools like Optimal Sort or UserZoom for remote research at scale with larger sample sizes.
Cards can represent individual concepts, content types, products, services, activities, or even images.
Use 30 to 60 cards for optimal cognitive load -- too many overwhelms participants, too few limits insights.
Consider hybrid sorts when you have some predetermined categories but want user input on others.
Ask participants to think aloud during open card sorts to understand the rationale behind their categorization.
Analyze results using dendrogram or similarity matrix tools to identify clustering patterns across participants.
Run pilot sessions with 3 to 5 participants before scaling up to identify confusing or ambiguous card labels.
Internal terminology or technical language confuses participants and skews results. Write card labels in plain language that matches how users actually talk about the content.
Exceeding 60 to 80 cards causes cognitive fatigue and unreliable groupings. Keep the set to 30 to 60 items and prioritize the most important content for the exercise.
Running card sorts with fewer than 10 participants produces unreliable patterns. Aim for 15 to 20 participants for open sorts and 30 or more for quantitative closed sorts.
Dismissing unusual categorizations means missing valuable insights about edge cases. Analyze outliers carefully as they may reveal important alternative mental models.
Implementing card sorting results without tree testing creates risk. Always validate your resulting information architecture with a follow-up study before building.
Target participant profiles and strategies for recruiting representative users.
Facilitator guide with step-by-step instructions and participant prompts.
Physical or digital cards representing content, categories, or features.
Dataset capturing each participant's sorting decisions and group labels.
Report with findings, suggested groupings, and IA recommendations.
Summary of participant insights and suggestions about content structure.
Visual representations showing item relationships based on participant input.
Suggested sitemap or content hierarchy based on card sorting results.