In this video interview, usability pioneer Jakob Nielsen explains how to conduct a simple user test. He observes that both quantitative and qualitative types of testing are helpful, but that one can gain a great deal of insight from just five user testing sessions.
First, screen for target subjects that are likely to use your site. Next, ask subjects to perform realistic tasks, one task at a time. The test subject sits at the computer, with the facilitator next to him or her and a bit back, so as not to intrude too much.
Test subjects can asked questions, but the facilitator can't answer. If they ask, "Can I do such and such....?" you answer, "You can do whatever you want." If they ask, "Is such and such the case?" you answer, "Let's try to find out." Encourage test subjects to think out loud with a running monologue.
Though larger studies record keystrokes and visual expressions, and even use two-way mirrors, all you really need is a pencil and a notepad -- which you're likely to fill with ways you can improve your site!
In response to Dr. Wilson's question about the pros and cons of the online testing tool UserTesting.com, Dr. Nielsen notes an advantage that it's easier to do, since people don't have to come into your office. However, you have less ability to screen subjects for your target audience and don't learn as much remotely as you would sitting right next to the subject, observing reactions, etc.
Jakob Nielsen is founder of the Nielsen Norman Group (www.nngroup.com) provides usability research on both mobile and regular websites, including actual user testing. Then information is shared through consulting and seminars. Nielsen publishes a bi-weekly blog "Alertbox" (www.useit.com) that provides results from research on usability. This interview was taped at the Conversion Conference West in San Jose, CA, May 4, 2010.