Guerilla User Testing
Comments(0) Since many libraries don’t have official usability positions or even committees, the idea of conducting user tests can be quite intimidating. But they don’t have to be! We do a lot of what we call Guerilla tests (a form of “discount usability”). There are a few methods- heuristic evaluation, paper prototypes, quick questionnaires, scenarios, etc.
If you are specifically looking for user feedback, paper prototypes/questionaires are ideal. The basic idea is to tackle one, maybe two key questions you might have about what terminology to use, order or placement of links on a website etc. The results of this type of test, like many other types of usability tests, can only indicate that there may be a problem with the interface element you’re testing. If it’s obvious that there are bigger problems with the interface, more extensive testing may be in order.
For example, our Usability Working Group decided to test the effectiveness of the link label “Get books, articles, and more,” that links to the library’s various delivery services. It was fairly clear that the current label was misleading (users could easily misinterpret it as being about circulation, research, how to find books etc.). So we looked at about 30 other library websites to see what other’s use and then made a list of these and a few other options to use for the test. We then basically split the test in 2 and asked one set of participants a set of questions and the other set different types of questions. The first set of participants were shown a printout of the library website, the current link was pointed out, and they were asked what they thought they’d find if they clicked the “Get books, articles, and more” link. The 2nd set of participants were shown the list of delivery services and asked if they could think of a name that would describe the contents of that list. They were then shown the list of alternate labels and asked to pick the one that best fit their understanding of the listed services. This test both verified that the current label was problematic and also gave us a sense of what other terms best fit the user’s expectations.
Tips:
- Ideal time for entire participant interaction should be 5, 10 minutes tops.
- Because the test is fast, try to test as many people as possible to get a decent consensus
- Choose a simple issue and make sure wording is clear. If questions are simple, you can often fit 5-7 questions into one test
- Demo the test with a co-worker or test subject to fine tune wording.
- Use a printout of the interface (using a live version or computer based mockup can be too cumbersome to do quickly)
- Choosing participants: we typically go to locations where there are many students - like the reference area, student union, etc. After approaching a few students, it becomes easier to tell who might and might not be willing to participate.
Some links:
- Usability on the cheap By Suneet Kheterpal
- Guerrilla HCI: Using Discount Usability Engineering to Penetrate the Intimidation Barrier by Jakob Nielsen
- Usage and Usability Assessment: Library Practices and Concerns by Denise Troll Covey
