Card-sorting Tools by Tom Tullis Posted on February 15, 2009 Last updated December 30, 2009

Card-sorting is a great way to learn how users, or prospective users, of a website perceive the logical relationships among the various elements of the site. There are three basic steps to any card-sorting project:

1. Define the cards to be sorted. These are commonly brief descriptions of the candidate pages of a website, not including pages that are purely for navigation.
2. Have representative users sort the cards. In most cases, the participants are allowed to create and name their own categories.
3. Analyze the results. Two of the common techniques for analyzing card-sorting data are hierarchical cluster analysis and multi-dimensional scaling.

Card-sorting is one of the "Special Topics" we discuss in Chapter 9 of our book, Measuring the User Experience. Specifically, we discuss the analysis of data from two different types of card-sorting studies: Open Card Sorts, where participants are allowed to create and name their own categories for the cards, and Closed Card Sorts, where participants are given the names of the categories that they are to sort the cards into. Both methods are useful and have their place. Open card sorts are particularly useful near the beginning of a project. Closed card sorts are more useful later on when you have some candidate organizations that you want to evaluate.

Here are some of the examples from the book of the analysis of open card-sort data:

• Table 9.8. Sample data from one participant in a card-sorting exercise involving fruit.
• Table 9.9. Overall distance matrix for all 20 participants in the sample involving fruit.
• Figure 9.2. Results of hierarchical cluster analysis of the sample data.
• Figure 9.3. Results of multi-dimensional scaling analysis of the sample data.

Now let's look at some of the tools that can help you do card-sorting studies.

	CardSort A Windows application. Created by Steffen Schilb in 2003 as part of his diploma thesis. Currently limited to 100 cards. Each card description is limited to 18 characters. Analysis is done using EzCalc, a program previously available from IBM. Free.
	OptimalSort A web-based service. Developed by Optimal Usability, a consulting company in New Zealand. Includes a variety of analysis features. Free version allows for up to 10 participants per project, 30 cards per project, and 3 projects. Licenses available for unlimited use. Also see the free tool for analysis of OptimalSort data by Aapo Puskala.
	uzCardSort Runs on Windows, Macintosh, and Linux versions of Mozilla. Provides basic browsing of the data. All data is saved in XML files. Open Source.
	WebCAT A web-based tool you can install on your server if you're a techno-geek! Developed by the Web Metrics team at NIST. Includes hierarchical cluster analysis. Free.
	WebSort A web-based service. Supports sorting of images as well as textual cards. Developed by Larry and Jed Wood of Parallax LLC. Provides a variety of data analysis options. Free version supports one study with 10 participants. Licenses available for unlimited use.
	xSort A Macintosh application. Developed by EnoughPepper, a company in Lisbon, Portugal. Provides a variety of reports and analysis tools, including hierarchical cluster analysis. Free.

Comments? Contact Tom@MeasuringUX.com.

Measuring UX Homepage