What Is Card Sorting and Why Does It Matter for Website Navigation?
If your website visitors struggle to find what they need, the problem usually isn’t the content itself. It’s how that content is organized. Card sorting is a proven UX research method that helps you understand how real users mentally group and label information, so you can build navigation menus that actually make sense to them.
Instead of guessing which pages belong under which menu, card sorting lets your target audience show you. The result? An information architecture (IA) grounded in real user behavior rather than internal assumptions.
In this guide, we walk you through everything you need to run a card sorting session for website navigation: the types, the tools, the ideal sample size, and exactly how to turn your results into an intuitive site structure.
How Card Sorting Works: The Basics
The concept is refreshingly simple:
- You create a set of cards, each representing a page, topic, or piece of content on your website.
- You ask participants to sort those cards into groups that feel logical to them.
- You analyze the groupings to discover patterns and build (or rebuild) your navigation.
Card sorting uncovers what UX researchers call mental models: the internal frameworks people use to organize information in their heads. When your website navigation aligns with these mental models, visitors find what they need faster, bounce rates drop, and conversions go up.
Open Card Sorting vs. Closed Card Sorting vs. Hybrid
Not all card sorts are the same. Choosing the right type depends on where you are in your design or redesign process.
| Type | How It Works | Best For |
|---|---|---|
| Open Card Sort | Participants sort cards into groups they create themselves and label each group. | Discovering new navigation structures, early-stage design, or complete site redesigns. |
| Closed Card Sort | Participants sort cards into predefined categories that you provide. | Validating an existing navigation structure or testing proposed menu labels. |
| Hybrid Card Sort | Participants sort into predefined categories but can also create new ones if needed. | Refining a draft navigation while leaving room for surprises. |
When to Use Open Card Sorting
Use an open card sort when you are starting from scratch or have no confidence in your current navigation. It is exploratory by nature. You will get rich, sometimes messy data, but it reveals how users naturally think about your content.
When to Use Closed Card Sorting
Use a closed card sort when you already have a proposed navigation structure and want to validate it. Since categories are fixed, the data is easier to analyze. It answers a focused question: “Do users place these items where we expect them to?”
When to Use Hybrid Card Sorting
A hybrid card sort gives you the best of both worlds. It tests your proposed categories while still capturing feedback you might not have anticipated. This is a strong choice for iterative redesign projects.
Step-by-Step: How to Run a Card Sorting Session
Follow these steps to set up and run a card sorting study that produces actionable results for your website navigation.
Step 1: Define Your Goals
Before touching a single card, be clear about what you want to learn. Common goals include:
- Understanding how users group your existing pages
- Identifying which menu labels feel natural to your audience
- Validating a proposed site map before development begins
- Prioritizing what content belongs on the homepage vs. sub-pages
Step 2: Create Your Cards
Each card should represent one piece of content, one page, or one feature. Keep the wording short and clear. Avoid jargon unless your users use that jargon themselves.
Tips for creating effective cards:
- Start with your sitemap or content inventory.
- Aim for 30 to 60 cards. Fewer than 20 won’t reveal enough patterns. More than 80 causes participant fatigue.
- Use plain language. “Pricing” is better than “Commercial Offerings Overview.”
- Avoid cards that are too similar, which confuses participants.
Step 3: Recruit Participants
Your participants should represent your actual target audience. If you run an e-commerce store, recruit online shoppers. If you run a B2B SaaS site, recruit decision-makers or end users in your niche.
How Many Participants Do You Need?
| Card Sort Type | Recommended Sample Size | Notes |
|---|---|---|
| Open Card Sort | 15 to 30 participants | More variability in responses requires a larger sample to find stable patterns. |
| Closed Card Sort | 15 to 20 participants | Fixed categories produce more consistent data, so fewer participants are needed. |
| Hybrid Card Sort | 15 to 25 participants | Falls between open and closed in terms of data variability. |
Research from the Nielsen Norman Group suggests that 15 users will uncover approximately 80% or more of grouping patterns in a card sort. For critical projects like a full website redesign, aim for the higher end of these ranges.
Step 4: Choose a Format (In-Person or Remote)
In-person card sorting uses physical index cards or sticky notes on a table. It allows you to observe body language, hear participants think aloud, and ask follow-up questions in real time.
Remote card sorting uses online tools (covered below). It is faster to set up, reaches more participants, and works well for unmoderated studies where participants complete the task on their own time.
For most website navigation projects in 2026, remote card sorting is the practical choice because of its speed and scalability. But if you have access to local users, an in-person session with 5 to 8 people can give you incredibly rich qualitative data to complement your remote study.
Step 5: Run the Session
If running a moderated session (in person or via video call):
- Explain the task clearly: “Please group these cards in whatever way makes sense to you.”
- Encourage thinking aloud.
- Do not guide or correct. Let participants struggle.
- Ask follow-up questions: “Why did you put these two together?”
If running an unmoderated remote session:
- Write clear, concise instructions in the tool.
- Set an estimated completion time (usually 10 to 20 minutes).
- Include one or two post-sort questions, like “Was anything difficult to place?”
Step 6: Analyze the Results
This is where the real value emerges. Here are the key analysis techniques:
Similarity Matrix
A similarity matrix shows how often any two cards were placed in the same group. The higher the percentage, the stronger the association. Cards that are grouped together by 70% or more of participants almost certainly belong in the same navigation category.
Dendrogram (Cluster Analysis)
A dendrogram is a tree diagram that visually clusters cards based on how frequently they were grouped together. Most card sorting tools generate this automatically. Look for clear clusters that form at high similarity levels; these become your top-level menu categories.
Category Labels (Open Sorts)
In open card sorts, review the names participants gave their groups. Look for the most frequently used terms. These are strong candidates for your navigation menu labels, because they are the words your users naturally use.
Placement Rates (Closed Sorts)
In closed card sorts, look at what percentage of participants placed each card in each category. A card placed in the “correct” category by 80% or more of participants validates your structure. A card with scattered placement signals a problem: either the card label is confusing, the category labels are unclear, or the card genuinely fits multiple categories.
Step 7: Build Your Navigation
Translate your findings into a draft sitemap:
- Use the strongest clusters from your dendrogram as top-level navigation categories.
- Use subclusters as dropdown or secondary navigation items.
- Use participant-generated labels as starting points for menu names.
- Place orphan cards (items no one grouped consistently) into a catch-all section or reconsider whether that content is needed.
- Test the draft with a closed card sort or tree test before finalizing.
Best Card Sorting Tools in 2026: Free and Paid Options
You don’t need a big budget to run a card sort. Here are the most widely used tools, ranging from free to enterprise-level:
| Tool | Free Plan? | Open / Closed / Hybrid | Notable Features |
|---|---|---|---|
| Optimal Workshop (OptimalSort) | Limited free plan | All three | Industry standard. Generates similarity matrices, dendrograms, and participant-level data. |
| UXtweak | Free plan available | All three | Built-in participant panel, solid analytics, and intuitive setup. |
| Maze | Free plan available | Open and Closed | Combines card sorting with usability testing and prototype testing in one platform. |
| UserZoom (now UserTesting) | No free plan | All three | Enterprise-grade. Strong analytics and integration with broader UX research workflows. |
| Miro / FigJam | Free plans available | Manual (any type) | Not purpose-built for card sorting, but works well for moderated in-person or video sessions using virtual sticky notes. |
| Google Sheets / Spreadsheets | Free | Manual (any type) | Works for very small studies. You will need to do analysis manually or export data to a dedicated tool. |
Our recommendation: If you are running your first card sort, start with UXtweak or Optimal Workshop’s free plan. They handle the analysis for you, which saves hours of manual work.
Common Mistakes to Avoid
Card sorting is simple in concept but easy to get wrong in practice. Watch out for these pitfalls:
- Using internal jargon on cards. If your users don’t say “knowledge base,” don’t put it on a card. Use the language your audience actually speaks.
- Too many cards. Anything above 60 to 80 cards leads to participant fatigue and sloppy sorting.
- Too few participants. Running a card sort with only 5 people rarely produces stable patterns. Aim for 15 or more.
- Skipping the follow-up. A card sort tells you what users grouped together, but not always why. Pair your card sort with brief interviews or post-sort questions.
- Treating results as absolute. Card sorting is one input into your information architecture, not the final answer. Combine it with tree testing, analytics data, and usability testing for the best outcomes.
Card Sorting vs. Tree Testing: What’s the Difference?
These two methods are closely related but serve different purposes:
| Method | Purpose | When to Use |
|---|---|---|
| Card Sorting | Discover how users group content | Before building or redesigning navigation |
| Tree Testing | Validate whether users can find content in a proposed structure | After building a draft navigation based on card sort results |
The ideal workflow is: Card Sort first, then Tree Test. Card sorting generates the structure. Tree testing validates it.
Real-World Example: How Card Sorting Transformed a Website
Imagine a mid-sized consulting firm with 120 pages on their website. Their navigation has seven top-level menu items, and analytics show that users rarely venture past the first two. Bounce rates on service pages are high.
Here is how they used card sorting to fix the problem:
- They listed all 120 pages and consolidated them into 55 cards (merging redundant content).
- They ran an open card sort with 25 participants recruited from their email list.
- The dendrogram revealed that users naturally created 4 top-level groups, not 7.
- Participants consistently used different labels than the company’s internal terms. For example, users said “How We Help” instead of “Service Offerings.”
- The team built a new sitemap based on the 4 clusters and tested it with a closed card sort (20 participants). Placement accuracy averaged 85%.
- After launch, average pages per session increased by 40% and bounce rate dropped by 22%.
This is a typical outcome. Card sorting doesn’t just reorganize your menu. It aligns your website with how your users actually think.
How Card Sorting Fits Into a Larger UX Process
Card sorting for website navigation works best when it is part of a broader user-centered design process. Here is where it typically fits:
- Content audit: Inventory everything on your site.
- Card sorting: Discover how users group that content.
- Sitemap creation: Draft your new information architecture.
- Tree testing: Validate the draft structure.
- Wireframing and prototyping: Design the navigation UI.
- Usability testing: Test the full experience with real users.
Skipping steps 2 and 4 is one of the most common reasons website redesigns underperform. At webdesign-by-osz.com, we include card sorting and tree testing as standard practice in our web design projects because the data consistently leads to better outcomes for our clients.
Frequently Asked Questions About Card Sorting for Website Navigation
How long does a card sorting session take?
For participants, an unmoderated online card sort typically takes 10 to 20 minutes to complete. Moderated sessions may take 30 to 45 minutes because of think-aloud discussion. The overall study, including setup, recruitment, and analysis, usually takes 1 to 3 weeks.
Can I do card sorting for free?
Yes. Tools like UXtweak and Optimal Workshop offer free plans that support small studies. You can also run a manual card sort using physical sticky notes or a whiteboard tool like Miro or FigJam at no cost.
What is the difference between open and closed card sorting?
In an open card sort, participants create and name their own groups. In a closed card sort, participants sort cards into predefined categories that you set up. Open sorts are exploratory. Closed sorts are evaluative.
How many cards should I use?
Aim for 30 to 60 cards. This range gives enough data to reveal patterns without overwhelming participants. If your site has more content than that, group closely related pages into single cards.
How many participants do I need for a card sort?
A minimum of 15 participants is recommended for reliable patterns. For open card sorts or high-stakes projects, 20 to 30 participants will give you more confidence in the results.
Can card sorting be done remotely?
Absolutely. Remote, unmoderated card sorting is the most common approach today. Participants complete the exercise from their own device at a time that suits them, and specialized tools handle data collection and analysis automatically.
What should I do after the card sort?
Analyze the results to identify grouping patterns and label preferences. Build a draft sitemap based on those findings. Then run a tree test to validate that users can find content within your new structure before committing to design and development.
Is card sorting only for new websites?
No. Card sorting is equally valuable for website redesigns, adding new sections to an existing site, or reorganizing a site that has grown organically over time and become difficult to navigate.