Information Design Using Card Sorting

At the beginning of any information design exercise, it is normal to be confronted by a very long list of potential subjects to include. The challenge is to organise this information in a way that is useful and meaningful for the users of the system.

While careful investigation and analysis of the information may reveal some clues, it can be virtually impossible to determine which topics should be grouped together.

The difficulty in organising the content stems from a lack of knowledge about how real users make use of this information. Without this, any exercise in information design is a purely theoretical one.

A card sorting session can go a long way towards resolving this problem.

What is card sorting?

In essence, card sorting works as follows:

1. Write down each topic on a filing (index) card.

2. Give the pile of cards to a number of users and ask them to group the cards into piles.

3. Collate the results, and make use of them when completing the information design.

This is a very simple, and often very effective, method of working with users to come up with a usable design.

It should be noted, however, that a card sorting exercise does not produce a finished information design. There are many other inputs into the information design process, such as:

• business requirements
• strategic directions
• technical goals and limitations
• usability guidelines

All of this information must be taken into account when structuring your content. That being said, the card sorting will give you an excellent starting point for this work, as it will inform you of how real users think.

Why use this method?

There are a number of benefits to using card sorting:

• Simple, and well understood.
• Cheap to use (the only real cost is the preparation time).
• Quick to apply, which allows more users to be involved.
• It draws out the natural groupings of information in a way that avoids directly asking users.
• Involves the users in the design process, and helps to demonstrate that the system will be created with the users' needs in mind.

These last two points are further discussed in the following sections:

Improving on the questionnaire

Surveys are an oft-used method to obtain information on user behaviour and opinions. While surveys are simple to construct, they suffer from the same limitations as the other methods that directly query the users.

That is, you will tend to obtain:

• what the users believe, not what they actually do
• what you want to hear, not the truth
• too much information from those who are coping with their jobs, and too little from those who aren't.

By replacing direct questioning, card sorting generally avoids these issues (although they are never fully resolved). It is for this reason that the approach is so widely used.

The use of a group format also has considerable benefits. Often, the participants will bring to the session quite different opinions. Through the discussion and eventual resolution of these differences, it becomes possible to identify a workable structure.

The group also provides an energy and enthusiasm that makes light work of an otherwise tedious task.

Involving the users

Card sorting sessions are an important opportunity to involve the actual users in the design process. This makes them feel involved in the project, and emphasises that the end product will be built to meet their needs.

It also provides a forum in which the project can be 'marketed'. Done successfully, this increases user enthusiasm and reduces any resistance to change when the system is implemented. It also allows realistic and accurate information to be conveyed regarding the timetable and scope of the project.

About this paper

This paper outlines a step-by-step approach to preparing and running card sorting sessions. Guidelines on how to analyse and make use of the results are also provided.

What is written here is based upon our practical experience of using card sorting to design intranets and online documentation. Hopefully these tips will allow you to learn from our mistakes and discoveries.

Preparing the topics

The first step to conducting a card sorting is to determine the list of topics. This list should be drawn from a wide variety of sources:

• existing documents and manuals
• existing online content
• descriptions of business groups and processes
• planned applications and processes
• potential future content.

The net should be thrown as wide as possible. In this way, it becomes possible to create a structure that will not just work for the present content, but for future material. Adding new content should require minimal reworking if the structure is correctly designed.

Creating the list of topics

It is actually quite tricky to come up with a workable list. There are a number of issues that need to be covered:

• The length of the list needs to be manageable. Too few items, and there is little scope for the users to come up with a structure. Too many items, and the task becomes daunting and confusing.

• Existing structures should not be reflected in the description and selection of the topics. It is important not to "pre-judge" the structure by leaving clues that will lead the users back to the current way of arranging the content.

• Words such as "manual" and "guide" should not be used in the topic descriptions. These reflect the current formats, and don't relate to the actual content.

• The hardest challenge is to find the right level of detail. That is, how much information is encompassed by each item on the list. A topic such as "human resources" is perhaps too broad, while "rate of leave accrual" is probably too specific. As a general guide, each item on the list should be self-contained, and cover one logical group of content.

• Avoid including "grouping terms" (such as "Company News"). The whole purpose of the card sorting exercise is to make the users group the cards.

  Any headings that you place in the list will tend to encourage the users to place other cards under them, instead of other groupings which may be more natural.

• The terms used in the list must be meaningful to the participants in the session. This should be checked by people knowledgable about the working environment and subject matter.

The general rule when selecting topics for inclusion is to ensure that they are "neutral", and they don't direct the users towards a particular structure. After all, there is no point in running a card sorting exercise if all you are going to get back is what you starting with.

It is always a good idea to have several people collaborate on the list. This will help to double-check that the entries are meaningful and workable.

Also, if you are running a number of card sorting sessions with different user groups, make sure that you revise the list between sessions to take into any account lessons learned.

Go to page: 1  2  3 

Author
James Robertson is the managing director of Step Two Designs, an online development company based in Sydney, Australia. James specialises in XML development, information management and systems design.