I was at the IMC last Thursday and Friday. One of the sessions I attended was called “Usability Testing Without the Lab Coats”. The session was led by Andre Charland of Nitobi.
Andre effectively hammered home some points I thoroughly agree with — concepts similar to those advocated by Steve Krug of “Don't Make Me Think” fame:
- You don't need a fancy lab to do usability testing. In fact, since labs are unnatural settings, they can make test subjects nervous. A regular office setting (or even a home-like setting) is arguably better.
- You don't need to run dozens of tests. Five is usually enough. And even one test is better than none.
- You don't even need subjects that are carefully screened to match your demographics. All that really matters is that you recruit “passionate users”, i.e. people who would actually use the site or application you're testing. (For example, if you're testing an online banking site, don't test on people who would never do online banking.)
- Test early. There's little point in testing a site after it's been coded. By then, changes are too expensive and time-consuming.
Test early, test often. Any test is better than no test, so test no matter what. Good advice all-round.