Card sort is used to find trends in how website visitors expect navigation to work. At its most basic, your research group takes a stack of index cards and sorts the terms on the cards into piles based on navigation structure.
Quantitative analysis
“I have a stack of cards. These cards represents terms of links you’ll see on a website. Sort them into stacks based on topic. If one card describes the pile, put it on top. At the end if you don’t have a top card, use a blank card to give it a name. For the “junk” cards, just put a blank card on top."
80-90 cards are ideal, 100 ok.
Software
Software can automate this so that it isn’t an actual stack of cards. Right now it’s desktop, not web based. There’s a new open source tool in development. Should be here around the end of the year.
- cardzort and cardcluster (no support)
- IBM – Ezsort
10-15 people is plenty of data
Samples: 10 students who are NOT engaged with you (not potential students)
10 HS Students who were scholarship applicants (highly engaged with the U)
Results
“About” was not a top level choice. “General Information” was the label 11/20 students chose.
Use results to inform your navigation, but it doesn’t have to match lockstep.
Interesting non-results:
FAQ – had no meaning to the group
About
Computing service
Prospective students (?!?!?!)
Results of card sort qualitative research
Asked the following after the sorting was done:
What does the phrase “future students” mean to you?
- Students that are going to that college
- Already decided
What does the term “prospective students” mean to you?
- nothing
Task based navigation gets around this problem- “become a student” versus prospective student
Term “admissions” tested okay. The problem is that there are tasks associated with that key word, but there are also tasks outside of admissions that still fall within prospective students.
Popularity: 5% [?]
No comments yet.