Research: Card Sorting and Tree Testing

Card Sort

Redesign of customer-facing Executive MBA website (Fall 2015)

For a major redesign of the Executive MBA program’s website, I conducted in-person interviews with current students, using them as proxies for prospective students who are the real targets of the site. Questions ranged from how they had decided to become students in the program to what their recollections about finding information on the website had been.

white index cards in groups on a dark surface — Letting users create affinity groups of content items eliminates internal silo-speak

As the conclusion of the interview, I conducted an open card sort exercise: the content strategist had prepared short snippets of content from the site-in-progress, and we asked users to group the cards and then name the groups they had created. The number of users was small, so even analysis using the Best Merge Method didn’t yield a strong result, but the resulting discussion between myself, the content strategist and the stakeholders was fruitful in designing the site navigation.

Tree Test

Executive Ed redesign – navigation Tree Test (Optimal Workshop TreeJack)

This project was a complete redesign of pages for the Executive Education program at Kellogg. While an outside firm had been hired to design the pages, I was to provide user research to inform the vendor’s user experience design. There was discussion and concern about the navigation proposed by the vendor, so it was decided to execute a tree test to compare two—and ultimately a third—navigation strategies.

Pie Tree Image — Tree test “Pie Tree” chart shows paths users take

Having already conducted the original research, I described 10 tasks done by site users, discovered in those interviews. These tasks were presented to study participants along with one or the other of the competing navigation schemes and the users were asked to identify where in the navigation they would be able to complete each task. Success, failure, false success and time taken were used to evaluate the options.

I conducted interviews that had surfaced the original needs, prepared the test including programming it in TreeJack, and led discussions of the results.

During the first round of testing neither option was a clear winner, and in fact, neither option did particularly well, leading to further discussion among the stakeholder team. Ultimately a third navigation option did fare better in subsequent testing and was adopted for the site.