Open, closed and hybrid card sorting
| Type | Categories | Answers the question |
|---|---|---|
| Open | Participants create and name them | How do people group this content? |
| Closed | Fixed in advance | Does our structure make sense? |
| Hybrid | Fixed, plus people can add new ones | What is missing from our structure? |
The Askyo block is a closed sort: you define the categories, and participants place every card into one of them. For an open sort, run a few moderated sessions or use a dedicated tool first, then check the resulting structure with a closed sort.
A card sorting exercise example
An online shop is redesigning its catalogue menu. Cards: “Kettles”, “Blenders”, “Air purifiers”, “Hair dryers”, “Smart plugs”, “Coffee machines”. Categories: “Kitchen”, “Home climate”, “Personal care”, “Smart home”. The question text:
- Sort each product into the section where you would look for it first.
If a third of participants put coffee machines under “Smart home”, that is a navigation problem you found before launch.
How many cards and participants?
Keep it to about 15 to 30 cards. More than that tires people and the last cards get dropped anywhere. Common UX guidance suggests at least 15 participants per user group before patterns become stable; more is better when the results are close.
How to analyse a closed card sort
agreement for a card = participants who put it in its most popular category ÷ all participants × 100%Build a card by category matrix: rows are cards, columns are categories, cells show how many people chose that pairing. Cards with high agreement are placed well. Cards split between two categories need a clearer name, a cross-link in the menu, or a rethink of the categories themselves.
Mistakes in card sorting studies
- Card labels that repeat category names, so people match words instead of meaning.
- Jargon from inside the company on the cards.
- Treating one closed sort as proof: it tests your categories, it does not invent better ones.
- Recruiting only colleagues, who already know where everything lives.
Card sorting vs tree testing
The two methods are often confused. A card sort asks people to build or fill a structure; a tree test gives them a finished menu and a task (“where would you find the warranty terms?”) and checks whether they get there. Run the card sort to shape the categories, then a tree test to confirm people can find things in them.
Building it in Askyo
The Card sorting block has two lists: cards and categories, at least one of each. Participants drag cards from an “Unsorted” zone into category zones; it works with a finger on a phone and from the keyboard. Categories stand in two columns on wider screens. The answer is complete only when every card has been placed.
Analytics shows the card by category matrix and how many cards were placed. The block is free on every plan, alongside the other research methods.
What you can set
- cards to sort
- categories to sort them into