Card Sorting in UX: What It Is and How to Run a Session That Actually Works

Ever watched someone navigate your website and think "Why are they looking there for that information?" You're not alone. The disconnect between how we organize content and how users expect to find it is one of the biggest usability killers out there.

Card sorting is your secret weapon for bridging that gap. It's a research method that reveals how people naturally think about and group information: giving you the blueprint for an information architecture that actually makes sense to your users.

What Is Card Sorting, Really?

Think of card sorting as organized mind-reading. You take all the content, features, or topics from your website or app, write them on individual cards (physical or digital), and ask people to group them in ways that feel logical to them.

The magic happens when you see patterns emerge. When 8 out of 10 people put "Contact Us" and "Support" in the same pile, that's not coincidence: that's insight into how your users' brains are wired.

The beauty of card sorting is its simplicity. No complex prototypes, no lengthy surveys. Just cards, categories, and human intuition doing its thing.

The Three Flavors of Card Sorting

Open Card Sorting

This is the "blank canvas" approach. You hand someone a pile of cards and say, "Group these however makes sense to you, and give each group a name."

Open sorting is perfect when you're starting fresh or when you suspect your current organization is completely off-base. I've seen open card sorts completely flip website structures on their head: in the best possible way.

When to use it: Early in projects, when redesigning from scratch, or when you need fresh perspective on familiar content.

Closed Card Sorting

Here, you provide predefined categories and ask people to sort cards into those buckets. It's like saying, "We think these are the right categories: do you agree?"

Closed sorting validates your hunches. If people struggle to fit cards into your categories, or if certain categories stay empty, you know something's not clicking.

When to use it: Testing existing navigation, validating proposed structures, or when you have business constraints that limit your organizational options.

Hybrid Card Sorting

The best of both worlds. You provide some categories but let people create new ones when your suggestions don't fit their mental model.

Hybrid sorting often reveals the gaps in your thinking. Those new categories people create? They're showing you what you missed.

When to use it: Refining existing structures, when you're mostly confident but want to catch blind spots, or when you need to balance user needs with business requirements.

Setting Up a Card Sort That Actually Works

Step 1: Get Your Content Ready

Start by listing everything that needs organizing. For a website, that might be all your main pages and key features. For an app, it could be every major function or screen.

Write clear, jargon-free labels. Instead of "Customer Relationship Management Portal," try "Manage Customer Info." Your participants shouldn't need a business degree to understand what's on the cards.

Step 2: Find the Right People

You need real users, not your colleagues' opinions. If you're designing a healthcare app, get actual healthcare workers. Building an e-commerce site? Talk to people who actually shop online.

Aim for 15-30 participants per sort. Fewer than that and you might miss important patterns. More than that and you're probably not going to learn much new.

Step 3: Choose Your Format

In-person sessions give you rich qualitative data. You can ask "Why did you put those together?" and watch body language when someone hesitates over a decision.

Remote digital tools like OptimalSort or Miro let you reach more people and make analysis easier. The trade-off is less insight into the "why" behind decisions.

Step 4: Set Clear Instructions

Keep it simple: "Group these cards in ways that make sense to you. There are no right or wrong answers: we want to understand how you think about this information."

For open sorts, add: "When you're done grouping, give each group a name that describes what's in it."

For closed sorts: "Put each card in the category where you'd most expect to find it."

Running the Session Like a Pro

Create the Right Environment

Whether it's in-person or remote, minimize distractions. Give people enough time to think: rushing leads to lazy groupings that don't reflect real mental models.

For in-person sessions, use a large table with plenty of space. For digital sessions, make sure your tool is intuitive and doesn't become the focus instead of the sorting task.

Ask the Right Questions (For Moderated Sessions)

• "Talk me through why you grouped these together"

• "What would you call this group?"

• "Where would you expect to find [specific item] if you were looking for it?"

• "Is there anything that doesn't seem to fit anywhere?"

Capture the Thinking, Not Just the Sorting

The groupings matter, but the reasoning behind them matters more. When someone says, "I put billing and payments together because they're both money stuff I don't want to think about," that's gold. It tells you these items need to be easily findable but not prominently featured.

Common Mistakes That Kill Your Results

Mistake 1: Too Many Cards

More than 50-60 cards and people's brains start to shut down. They'll resort to arbitrary groupings just to finish. If you have more content than that, consider multiple smaller sorts.

Mistake 2: Vague Card Labels

"Portal" could mean anything. "User Dashboard" is better. "View Your Account Info" is best. Specificity prevents confusion and leads to more accurate sorting.

Mistake 3: Leading the Witness

Don't explain your current organization or hint at preferred groupings. Comments like "We usually put these together" will bias results.

Mistake 4: Ignoring the Weird Results

When someone creates a category that makes no sense to you, dig deeper. Sometimes the "weird" results reveal mental models you never considered.

Real-World Example: Government Service Redesign

I once worked on a project for a city's online services. The existing site organized everything by department: Planning, Parks & Recreation, Public Works. Makes sense from a government perspective, right?

Wrong. Our card sort revealed that residents think in terms of life events and tasks: "Moving to the City," "Starting a Business," "Dealing with Property Issues." Nobody cares which department handles what: they just want to solve their problem.

The redesign organized content around these user journeys instead of organizational charts. Permit applications went from a 15-minute hunt to a 3-click find.

Analyzing Your Results: Turning Data Into Direction

Look for Strong Agreements

When 80% of participants put the same items together, that's a no-brainer grouping. These become your core categories.

Identify the Confused Items

Items that get scattered across different groups by different people need attention. They might need clearer labeling, better descriptions, or perhaps they belong in multiple places.

Extract Category Labels

Pay attention to what people name their groups. Often, these labels are better than anything you could come up with internally. "Stuff I Need to Do" might not sound professional, but if that's how users think about it, maybe "Tasks" or "Action Items" is your answer.

Create Your Information Architecture

Use the patterns to build your navigation, categorization, and content hierarchy. The strongest groupings become your main categories. The category labels become your menu items.

Your Card Sorting Checklist

Before the session:

• Content cards written in plain language

• 15-30 representative participants recruited

• Clear instructions prepared

• Tools tested (if going digital)

• Quiet, distraction-free environment arranged

During the session:

• Instructions clearly communicated

• Participants working without time pressure

• "Why" questions asked (for moderated sessions)

• Participant reasoning captured

• Unusual groupings explored

After the session:

• Strong agreement patterns identified

• Problematic/confused items noted

• Category labels extracted from participants

• New information architecture drafted

• Results validated with team/stakeholders

Card sorting isn't magic, but it's pretty close. It transforms guesswork into insight and assumptions into evidence. Most importantly, it ensures your information architecture works for the people who actually have to use it: not just the people who built it.

When you organize information the way your users expect to find it, everything else gets easier. Navigation becomes intuitive. Search results make sense. And those confused "Why are they looking there?" moments become a thing of the past.