March 24, 2005
Garrick writes about using fake data to see if anyone's paying attention.
When putting together a prototype for usability testing, it's best to use realistic data. If you're evaluating the readability of a search results screen, put in the actual results. If you're evaluating a check-out process, make all the information throughout the entire process real.
Then, after tweak the data just slightly. Make it humorous, make it unrealistic, throw in a knock-knock joke.
I go back and forth on this. When using real data I've certainly had my share of "but that's not Jane's phone number!" moments, and it is fun when someone barks with laughter in the middle of a meeting because they just got a joke in the mockup. (Although y'know, no one has yet asked why Takeshi Kaneshiro and Yuen Biao are rooming together at Alexandria Technical College … maybe I should have picked a school that has campus housing. :) Not surprisingly, I've found that the type of data I use really depends on what I'm doing.
Dan Brown has put together a fantastic (and large!) poster about Representing Data in Wireframes (PDF) that touches on the nuances involved in choosing what kind of data to use, the risks of using each type, and how to apply each technique. This one's a keeper. (Via UXCentric.)
Oh. When I say large poster, I mean large: three by six feet. Not quite a Unicode chart, but whew!
