Articles & Examples

Card Sorting - Part 1

Card Sorting is a user testing method for organising data into structure. There’s a lot of information about on what they are, how to conduct them. Problem is, they’re all over the place and mostly they’re written by scientists so tend to be a little difficult to grasp and bogged down in analysis (which can take over your life if you let it!) I’ve decided to document my understanding of how to plan, conduct and analyse a card sort, from a practitioners point of view.

The basics

Why use them?

  • Cheap and easy to conduct
  • Identifies patterns in grouping across multiple users
  • Categorises items
  • Highlights terminology which is likely to be misunderstood

Card sorting will help you to...

  • Organise information on a page level
  • Create top level navigation
  • Organise content
  • Gather user-centred labelling
  • Validate thinking

Type of card sort

Closed card sort

  • Pre-defined category labels
  • Helps identify where things belong in a structure
  • Forces participants to think carefully about different categories
  • Great validation tool

Open card sort

  • users build architecture from individual items of content upwards
  • Provides user generated labels
  • Usefull in early stages of a project for exploration

Designing your card sort

Establish objectives
Before you begin your card sort you need to ask yourself a number questions to ensure it's the correct task to perform to get the information you want to know.

Ask yourself:

  1. What do you want to find out?
  2. Have I got a structure in mind, or just loads of content?
  3. are there any particular pieces of content I have doubts about?
  4. Will this help us, or is it too late to change the structure or labelling?

Choose card sort method
An open card sort will help you...

  1. Establish general section headings
  2. Check labelling of content
  3. Spot problem areas
  4. Spot affinities

A closed card sort will help you...

  1. See how users classify content
  2. Identify problems with an existing information architecture
  3. Identify ambiguous labels or bits of content
  4. Test sections

Make cards

  • As a rule of thumb, use no more than 30 cards. Every five additional cards adds between 5 and 10 minutes to the length of the test.
  • If you've got lots of content, collect a representative sample.
  • Work with others to agree the labels for the cards. The labels need to be as clear as possible.
  • Use card, paper's rubbish.
  • Have some spares
  • Number each card on the back so results can be compiled easily.

Organising the card sort

  • Make sure you've got a room with a big table, or the floor.
  • Recruit users. This could be through an agency, but allow at least 2 weeks for this. Create a schedule with slots and define the profile of your users. This can be done through the use of Personas.
  • Get a digital camera to record each sort.
  • Get observation. At least one person to take notes on the session.

Articles

Examples

Legal disclaimer: Any redesigns of existing sites may use names, logos, graphics, and/or images which are trademarked or copyrighted by the owning organisation. These items do not fall under any copyright claim or association of Mark Boulton.

Search

A picture of Mark BoultonI'm a graphic designer from near Cardiff in the UK. I've been a designer for over ten years now and primarily work on the web. I'm still partial to a bit of print every now and then though. I used to work for Agency.com in London as an Art Director before working as a Senior Designer for the BBC in sunny Cardiff. This was all before I took leave of my senses and formed my own design consultancy, Mark Boulton Design Ltd.

I've got a thing about grids and typography and occasionally ramble on about them to anyone who will listen.

If you're after simple, clean and effective web design; let me know.