Card Sorting

An exercise in which a group of users generate a category tree or folksonomy.

Usually the activity will proceed by shuffling cards with the various options printed on. Then laying them onto a large table in "natural groups". Once all the cards are on the table each group is described and these names are noted as a possible classification scheme. It is common to then repeat the process a number of times and note the similarities and differences between the groupings that emmerge.

It is normal to find that each run through the shuffled cards will highlight a different aspect that is important for the information being analysed.

Teaching Information Architecture

If your client is attempting to impose a too simplistic structure to your information this can be a good exercise to help them appreciate that there is no just a single answer.

Select a collection of playing cards, with various suits and values from a variety of packs (small, large, round, triangle the more distinct the better). You should have 10-20 cards evenly spread between the suits, values and pack sizes (select a grouping using sampling theory).

Now go through the classification exercise three times, you should find that the groupings reflect distinct aspects. For example the first may be based on suit, the second on value and so on.

Point out that a simple adoption of the first approach would have ignored the aspect revealed by the second. An that different users will naturally group their information in distinct ways.

Links to this page

The following pages link to here: Gathering Data