What needs to be finished first
A list of pages, labels, topics, or items exists that can be grouped without mixing different levels of granularity.
What needs to be ready before start
Card set or digital sorting board; label stickers or text cards; timer; notes for clusters and edge cases; camera or export tool.
One facilitator; five to fifteen participants or test users; one observer or note taker; optionally one person who watches for label ambiguity.
Target audience, list of content items, and the decision to run an open or closed sort; any existing navigation or taxonomy assumptions.
30-45 min
Prepare one card per item. Keep the wording short and consistent. Decide in advance whether participants create their own labels or sort into predefined groups.
The one question this method answers
How do people naturally group and label these items?
Marker: Minute
| Step | Duration | Action | Hint |
|---|---|---|---|
10-5 min | 5 min | Explain the item set, the goal, and the sort mode. Make sure everyone understands whether the sort is open or closed. | If the sort mode is unclear, participants will invent different rules. State the mode before the first card is moved. |
25-15 min | 10 min | Let participants move cards into groups or categories. Observe patterns, hesitation, and cards that are difficult to place. | Do not guide the grouping too early. The point is to learn the user's mental model, not to coach the result. |
315-25 min | 10 min | Review the clusters, ask for labels, and capture why certain items were grouped together or kept apart. | Ambiguous labels are often more valuable than perfect agreement. Record the uncertainty instead of forcing a false consensus. |
425-35 min | 10 min | Compare the results across participants, highlight stable clusters, and mark items with inconsistent placements. | Look for patterns, not just majority votes. Repeated disagreements are clues for weak labels or a wrong taxonomy. |
535-45 min | 10 min | Translate the findings into navigation or label hypotheses and note which items need a second test such as tree testing. | A sort result is a hypothesis, not the final IA. Always end with the next validation step. |
What comes out at the end
Cluster map or card sort export with labels, stable groups, ambiguous items, and a shortlist of information architecture hypotheses.
Record date, target audience, and sort mode. Keep separate versions for open and closed sorts, and note which items changed between runs.
Compact working template for Card Sorting with context, input, output artifacts, and next step.
Shows how users group and label content.
When and for what do we use this method?
Which data, observations, decisions, or materials are available?
Short notes along the runsheet.
Owner, date, and success signal.
Concrete filled scenario, fictional example
Concrete filled scenario, fictional example
Setup: 28 content items, open sort, 8 participants.
Stable clusters
Ambiguous items
Next step: Run tree testing on the top two label sets before updating the navigation.
Recognize symptoms and steer against them
Some cards are full pages while others are tiny subtopics, so participants cannot sort them consistently.
Normalize the item list before the session. One card should mean one comparable unit.
Participants start inventing labels in a closed sort or use predefined categories in an open sort.
Explain the mode clearly and show one example. The sort mode must stay stable from start to finish.
The result reflects only one team's internal language.
Use people who match the intended audience. If the audience is broader, run more than one session.
The team writes a label and stops without checking whether it really fits the grouped cards.
Test every label against the items in the group. If it feels forced, rewrite it.
The navigation is changed directly from one card sort without follow-up testing.
Treat the result as a hypothesis. Follow it with tree testing or first-click testing before making structural changes.
Done signals checkable in under a minute
Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.