Online vs In-Person Card Sorting: Which Should You Pick?

The short answer: it depends on what you need to learn.

Online card sorting lets you collect data from dozens (or hundreds) of participants quickly and cheaply. In-person card sorting gives you something online tools can't -- you get to watch people think out loud, ask follow-up questions on the spot, and understand why they grouped things the way they did.

Most teams do well starting online and layering in a few moderated sessions when the stakes are high. But let's break down each approach so you can decide for yourself.

Online Card Sorting

Online card sorting is the workhorse of most UX research teams. You set up a study, share a link, and participants sort cards on their own time from wherever they happen to be. No scheduling back-and-forth, no booking a conference room.

The big win is volume. You can easily get 30, 50, or 100+ participants, which means your similarity matrices and cluster analyses actually hold up. The tradeoff is that you're flying blind on the "why" -- you see what people did, but not the reasoning behind it.

Pros

• Much larger sample sizes than in-person methods
• Significantly cheaper per participant
• Results come back fast -- often within a few days
• Participants sort in their own environment, on their own schedule
• Built-in analysis tools handle the number crunching for you
• No geographic limits on who can participate

Cons

• You can't ask "tell me more about that" in the moment
• No way to observe body language or hesitation
• Some participants will rush through without much thought
• You miss the context behind sorting decisions

Best For

• Validating patterns across a broad audience
• Teams spread across different locations
• Projects with tight budgets or timelines
• Studies that need participants from multiple countries

In-Person Card Sorting

In-person card sorting is where the rich, messy, human insights live. You sit across from someone, hand them a stack of cards, and watch what happens. They pause, they backtrack, they mutter "well, this could go in either group." That's gold.

The real value is the conversation. You can probe decisions in real time -- "What made you put those two together?" -- and catch the mental models that never show up in quantitative data. You also build empathy fast, which is useful if you need to bring stakeholders along on the journey.

Pros

• Deep qualitative insights you can't get any other way
• Real-time follow-up questions as participants sort
• You can watch the thinking process unfold
• Clears up confusion immediately
• Great for building stakeholder empathy (invite them to observe)
• Full context around every decision

Cons

• Practical limit of roughly 5-15 participants per study
• Considerably more expensive per session (scheduling, space, facilitator time)
• Takes weeks to recruit, schedule, run, and analyze
• Limited to people who can show up in person
• Manual analysis is time-consuming
• Facilitator presence can influence behavior

Best For

• Exploring a new or unfamiliar domain
• Complex information architecture with lots of ambiguity
• Understanding the reasoning behind sorting decisions
• Studies where stakeholders need to see real users firsthand
• Recruiting from a very specific or specialized user group

The Hybrid Approach

Honestly, the best results usually come from doing both -- just at different scales.

Start with a handful of in-person sessions (5-10 participants) to understand how people think about your content. Listen for patterns in their language, watch where they hesitate, and note the categories they invent on their own. Then take what you've learned and run a larger online study to see if those patterns hold up across a wider audience.

This gives you the qualitative "why" and the quantitative "how many," without blowing your budget on dozens of moderated sessions.

1. Run 5-10 moderated sessions to surface mental models and language patterns
2. Design an online card sort informed by what you heard in person
3. Collect online responses from a larger group to validate (or challenge) your early findings

Tools for Each Method

Online Card Sorting Platforms

• CardSort (unlimited free usage)
• Optimal Workshop (professional features)
• Maze (integrated user testing)

In-Person Materials

• Physical index cards or sticky notes on a wall
• A quiet room and a note-taker (or recording setup)
• CardSort with screen sharing works well for remote-moderated sessions

Data Quality: What You're Actually Getting

These two methods give you fundamentally different kinds of data, and neither is "better" -- they answer different questions.

Online data is broad. You get clean similarity matrices, dendrograms, and statistical patterns across a large group. What you don't get is any sense of why those patterns exist.

In-person data is deep. You get transcripts full of reasoning, moments of confusion caught in real time, and a feel for how people actually think about your content. But you're working with a small sample, so you can't generalize with confidence.

The sweet spot is using each type of data for what it's good at. Let the quantitative data tell you what's happening; let the qualitative data tell you why.

Our Recommendation

For most teams, start online. It's faster, cheaper, and gives you a solid foundation of data. If you see surprising patterns or confusing clusters in the results, follow up with a few moderated sessions to dig into the "why."

Start online when you want to:

• Get a broad read on how users group your content
• Move quickly and stay within budget
• Validate or pressure-test an existing IA

Start in-person when you need to:

• Explore a brand-new domain where you don't know what to expect
• Get stakeholders in the room watching real users
• Understand specialized or expert-level mental models

Further Reading

• What is Card Sorting? Complete Guide
• Card Sorting (UX Glossary)
• Information Architecture (UX Glossary)
• How To Run Your First Card Sort Study

Frequently Asked Questions

How many participants are needed for reliable card sorting results? It depends on the method. For online card sorts, aim for 30-50 participants if you want patterns you can trust statistically. For in-person sessions, you'll typically hit a point of diminishing returns around 5-15 people -- after that, you start hearing the same themes over and over. Your choice really comes down to whether you need broad validation or deep understanding.

Which card sorting method provides better cost efficiency? Online, by a wide margin. You skip the scheduling overhead, the facilitator time, and the physical space. In-person sessions cost more per participant, but the qualitative depth you get from each one is much richer -- so it's not really an apples-to-apples comparison. Think about what kind of insight you need, not just the price tag.

Can online card sorting completely replace in-person research sessions? Not entirely. Online tools are great for spotting patterns at scale, but they can't replicate the experience of watching someone think through a tricky sorting decision in real time. If you only have budget for one method, online is a solid default -- but you'll miss the nuance that comes from face-to-face observation.

What is the typical timeline difference between online and in-person card sorting? Online studies can wrap up in a matter of days once you share the link. In-person studies usually take a few weeks when you factor in recruiting, scheduling, running sessions, and analyzing notes. The actual sorting task takes about the same amount of time for participants either way -- typically 15-30 minutes.

Which method works best for international user research projects? Online is the clear winner here. Participants can join from anywhere without travel costs or timezone headaches. If budget allows, you can supplement with in-person sessions in key regions to pick up on cultural differences that might not surface in an unmoderated study.