Heading into a major site redesign, with a product team split among several divergent and firmly held opinions, I decided it was time to break out a card sort.
Card sorting is an old, and often overlooked, UX method for organizing information. In its simplest form, you hand a test subject a stack of index cards representing the site’s content and ask her to sort them into piles and name each pile. This is what’s known as an “open card sort” — the naming of the piles is left up to the test subject. This is a great way to discover mental models of an information space, natural groupings of content, and labeling. Once you have derived a set of generally accepted categories and labels, you can move on to a closed card sort. In a closed card sort, you will provide a set of category names and ask test subjects to place the content cards into the categories where they seem to fit best. This is a good way to test your nascent navigation scheme.
These tests are pretty straightforward. The difficulty comes when you go to collate the data. Measuring the results of a card sort comes down to similarity scores. Similarity scores measure the number of times different test subjects place the same card in the same pile. If all your test subjects sorted two cards into the same pile, then the two items represented by the cards would have 100 percent similarity. If half the users placed two cards together and half placed them in separate piles, those two items would have a 50 percent similarity score. Multiply that by 40 or 50 cards and a dozen or so test subjects, and you’re looking at a long, costly process of collation and counting.
That’s why I was very happy to find Optimal Sort, an online card sorting tool with great analysis components. Part of the Wellington-based Optimal Workshop, Optimal Sort allows you to create cards representing your content and then invite test subjects to sort and categorize the content by following a link to your survey. The real value comes with the analysis tools. Optimal Sort provides a clutch of analysis tools, the most useful of which are the Similarity Matrix (shown below) and the Dendograms. Both quickly highlight the obvious content clusters and suggest generally understood labeling. For me, those two features were worth the $110 monthly fee.
So far, we’ve had over 100 responses to our survey and some very clear patterns are emerging. It seems that when the humble old card sort is merged with some smart data analysis and presentation, we’ve got a great new UX tool.
No Comments yet »
Leave a comment
You must be logged in to post a comment.