Card Sorting
Definition
Card sorting is a research method that helps you understand how people naturally organize information. You give participants a set of cards with different topics or content written on them, and they group the cards into categories that make sense to them.
This technique helps you understand users' mental models and create navigation systems that feel intuitive to them. Instead of guessing how to organize your content, you're learning from real users how they think about and categorize information.
Why Card Sorting Matters
Card sorting helps you understand user mental models by seeing how people naturally group information. It helps create intuitive navigation that matches how users think about your content, and avoids assumptions about how information should be organized.
It improves findability by structuring content in ways that make sense to users, reduces user confusion by organizing information logically from the user's perspective, and validates existing structures to see if your current organization works for users.
Types of Card Sorting
There are several ways to conduct card sorting, each with different benefits:
Open card sort lets participants create their own categories and name them. This is great for discovering how users naturally think about your content and generating new ideas for organization.
Closed card sort asks participants to sort cards into categories you've already defined. This is useful for testing or validating an existing structure you want to improve.
Hybrid card sort combines both approaches - participants can use your predefined categories but also create their own. This gives you the structure of closed sorting with the flexibility of open sorting.
Tree testing (also called reverse card sort) tests an existing hierarchy by asking participants to find specific items in a text-based site structure. This is useful for validating navigation after you've done card sorting.
The type you choose depends on what you're trying to learn. If you're starting from scratch, open sorting is usually best. If you're improving an existing structure, closed or hybrid sorting might be more appropriate.
How to Conduct Card Sorting
Here's a practical approach to running a card sorting study:
Preparation:
- Define your goals and what specific questions you want to answer
- Choose representative content items (usually 30-70 cards work well)
- Recruit 15-20 participants who represent your target users
- Decide whether to do it in-person or online
Execution:
- Explain the process clearly to participants
- Have them organize cards into logical groupings
- Ask them to name the categories they've created (for open sorts)
- Conduct follow-up discussions about their reasoning
Analysis:
- Look for patterns in how participants grouped items
- Identify consistent category names and groupings
- Create visualizations to help understand the data
- Use your findings to inform your information architecture
The key is to keep it simple and focus on understanding how users naturally think about your content, not on getting perfect results.
In-Person vs. Online Card Sorting
You can conduct card sorting in different ways:
In-person card sorting uses physical cards on a table. This gives you rich qualitative data and lets you observe participant behavior directly, but it's more time-consuming and geographically limited.
Online card sorting uses digital tools and software. This allows for larger sample sizes and easier analysis, but you get less insight into participant thinking and behavior.
Choose the method that works best for your situation. If you need deep insights and can gather participants in one place, in-person is great. If you need larger sample sizes or participants are spread out, online might be better.
When to Use Card Sorting
Card sorting is particularly valuable when you're starting a new product and need to create initial information architecture, redesigning an existing product and want to restructure content organization, conducting a content audit to understand how to categorize your inventory, designing navigation and want to create intuitive menu structures, testing terminology to find user-friendly labels for categories, or working with different audiences to understand how various groups categorize information.
Analyzing Your Results
Look for patterns in how participants grouped items together.
Pay attention to category names participants used - these often reveal how they think about your content.
Consider participant comments from follow-up discussions for additional insights.
Don't ignore outliers - unusual groupings might contain valuable insights.
Focus on consistency - if most participants grouped items the same way, that's probably a good organizational approach.
Use visualizations like dendrograms or similarity matrices to help understand the data.
Best Practices
Here are some tips for successful card sorting:
Use clear language on your cards - avoid jargon or technical terms that participants might not understand.
Keep the number of cards manageable - usually 30-70 items work well. Too many cards can overwhelm participants.
Include diverse users from different segments of your audience to get a well-rounded perspective.
Combine with other methods - use card sorting alongside other research techniques for a complete picture.
Iterate and refine - conduct multiple sorts, refining your content and categories based on what you learn.
Validate your results by testing the resulting structure with tree testing or usability testing.
Consider context - evaluate your findings in light of business goals and technical constraints.
Common Mistakes to Avoid
Using too many cards can overwhelm participants and make the exercise less effective.
Creating leading categories in closed sorts that bias participants toward certain groupings.
Working with too few participants - you need enough people to identify meaningful patterns.
Ignoring outliers - unusual groupings might contain valuable insights you shouldn't dismiss.
Over-relying on card sorting - use it as one input among many, not the only basis for decisions.
Using technical content that participants don't understand or relate to.
Applying results too rigidly without considering context and other factors.
Getting Started
If you want to try card sorting, begin with these fundamentals:
Start with a clear goal. Know what specific questions you want to answer.
Choose your content carefully. Select representative items that cover your main topics.
Recruit the right participants. People who represent your target users.
Keep it simple. Don't overcomplicate the process or analysis.
Focus on patterns. Look for consistent ways participants group items.
Use the insights to inform your information architecture and navigation design.
Remember, card sorting is a tool for understanding how users think about your content. The goal isn't to get perfect results, but to gain insights that help you create more intuitive and user-friendly information structures.