Testing Web sites for usability is largely recognized as a requirement for launching top-quality sites. Advocates of usability testing have emphasized the need for sites to evaluate how successfully they've communicated their message to their target audience. And the results of usability testing are extremely valuable, not only in terms of fixing user interface flaws, but also in uncovering users' needs.
Still, something's missing. It seems that many Web sites consider usability testing to be a luxury, an unnecessary expense that can be ignored. The belief that companies need to hire an outside agency to run usability testing also causes them to shy away. Really, though, you don't need much time or money to test your site. Spending just a couple of days with a dozen or so subjects, you can run a successful, professional usability test for less than $1,000.
The first thing you'll need in order to begin testing is, well, a room with a computer. You'll want to equip the machine with a 33.6 Kbps modem (don't forget, this is how most of the Web sees your site) and a 14- or 15-inch monitor. It's possible to test usability with nothing more than this and a moderator, if you like, but there are some extras you may want to consider. Programs such as Virtual Network Computing (VNC) or Microsoft NetMeeting -- both freely available -- can let your development team watch the users' actions, including their mouse movements, from a computer in another room. (You don't want users to feel like they're performing in front of an audience.)
For a few dollars more, you can add an $80 Web camcorder to watch the users' facial expressions. These often speak louder than words. You can even save all this information for further review with a television, VCR, and a VGA-to-TV converter, which can be purchased for less than $200. Remember, this equipment is reusable, so a $300 initial outlay is pretty reasonable.
It's usually not hard to find subjects for usability testing. You'll need 8 to 12 subjects to produce reliable test results. Often you can post a message to a newsgroup or mailing list, or even your own site, soliciting participants. Offering cash -- say, $50 -- in exchange is a popular tactic but, depending on your audience, you may do just as well giving away gift certificates, T-shirts, or something else.
It's important to be an objective moderator in usability testing. Be extremely careful with your mannerisms and reactions: Anything you do may skew the sample you're collecting. Encourage your subjects to be as descriptive, honest, and candid as possible. Welcome all types of feedback, including comments on your choice of color scheme and font. These comments are all valuable. Ask subjects to describe every step they're taking as they navigate through the site. Have them make notes on everything that's awkward or confusing.
Create a set of opening questions, focusing on the users' general experience with computers and the Net, previous experiences with the site, and what they want from it in the future. Prepare a set of simple tasks related to common features on the site. Allow users to stray from the path: This can help you determine exactly where you should place certain features of your site.
When a task is completed, ask subjects what they thought of the process and how it can be improved. (In the event that a user fails to complete a task, don't force the issue. Move on. You're testing the site, not the user, and a user failure may indicate something that needs to be reworked.) When the testing is complete, ask the subject for an overall opinion. Focus on trouble spots.
In the final analysis, you'll need to look at your notes, user responses, and any video you've gathered, and try to determine patterns. The most important details to be worked out are failed tasks. They're likely to be problem areas for many users, and they should be addressed immediately. Take note of any other common problems users had, and address them as soon as you can. Even after the usability testing, you can take e-mail feedback and see if user troubles have continued.
What do you do when you're done with usability testing? Run another session! I recommend running usability testing at least every quarter, in order to catch new problems as they surface and any changes in your target users and their needs.