Testing 1-2-3 - Internet/Web/Online Service Information

Industry Standard, The, March 5, 2001 by Alexei Oreskovic

Easy-to-use Web sites aren't just "designed" -- they're the result of extensive usability testing. The good news: It's easier than you think.

One day every two weeks, Eric Burnette welcomes strangers into his office and videotapes them as they perform various acts. There was a 70-year-old lady in spiked heels, for instance. "She was very forthcoming," recalls Burnette. "She took no prompting. I just sat back and watched." Then there was a woman Burnette describes as having very little confidence. "One of the worst," he remembers.

As the usability expert for Motley Fool, a financial information service based in Alexandra, Va., Burnette watches people as they click through the company's Web site. By carefully noting how users surf -- paying attention to where they stumble and where they succeed -- Burnette collects valuable information about the site's design.

It wasn't long ago that many businesses, in haste to establish a presence online, put up badly designed sites. Today, the emphasis is on fine-tuning those sites, making them easy to use. So more and more corporate sites, like patients in a sick ward, are undergoing a battery of tests designed to shed light on problem areas.

Naturally, there are a range of products, vendors and consultants promising miraculous cures for poorly designed sites. For the owners of plagued pages, this kind of intervention-from-above can seem the easiest, if not the only, option. But according to many site experts who have tested and fixed their own designs, improving a site's usability requires neither deep pockets nor expensive technology. "The dirty little secret of usability testing is that you can actually spend very little money and get lots of valid and useful info;' says Burnette.

Many people associate usability testing with high-tech labs complete with two-way mirrors and fingernail-size audiovideo equipment. All that's nice, but it's hardly essential. Motley Fool conducts its experiments in an old conference room using a consumer-grade VHS camcorder to tape participants' onscreen activity.

High-tech eye-tracking equipment, which pinpoints exactly where a person's eyeball gazes on an individual Web page, is another unnecessary luxury. It can be useful (a recent study by the Poynter Institute found that online newspaper viewers looked at text first and images later), but it's usually excessive. "Frankiy, it's more worth my while to go to the Web and read somebody else's research than it is to spend $50,000 and start an eye-tracking lab," says Burnette.

The real key to improving a site's usability lies not so much in equipment as in the actual process. Most experts recommend testing early and often so that the results can help drive site design. "Many firms take the approach of testing for usability very late in the cycle. That's a real flaw," says Jeff Rubin, managing partner of Usability Group and author of The Handbook of Usability Testing. "There's very little you can change at that late date."

The answer, Rubin believes, is to begin testing as soon as a Web site hits the concept stage. It's a belief that's shared, and practiced, by many established sites. At Motley Fool, says Burnette, usability testing starts before a project is approved.

That's also how it works at Terra Lycos, an international network of Web sites based in Barcelona, Spain. For sites that are not yet live, a favored technique at Terra Lycos is card-sorting, in which the tester jots down various ideas and terms on index cards. Participants then are asked to organize the cards. "Out of that comes a hierarchy about how things should be put together, which helps with making menus and taskbars," explains Dave Hendry, the company's network manager for user interface research.

Another early-stage technique is to run so-called low-fidelity designs past test subjects. These designs are basically nonworking prototypes of a site -- anything from paper drawings of a homepage to a digital mock-up. While such experiments aren't very useful for learning about specific details, they can flag potential problem areas and offer valuable insights into how people perceive basic concepts.

The most critical phase of usability evaluation, however, comes after a site is functional. This is when individuals are closely monitored as they perform a series of specific tasks. A financial site, for instance, might ask test participants to find various stock ticker symbols or set up an online portfolio.

Proponents tout the advantages of this type of hands-on exercise over question-based focus groups. "What [focus groups] say isn't necessarily what they do," explains Keith Pelczarski, another usability expert at Motley Fool. "They tell you what you want to hear." With task-oriented testing, on the other hand, there's less room for confusion -- people either succeed at using the site or they don't.

For all its effectiveness, task-oriented testing is surprisingly affordable. Terra Lycos and Motley Fool both test their sites on only five people (all of whom can be recruited by a focus-group firm for less than $1,000). With 16 test participants, Lego Group, which recently launched a commerce version of its site, is at the high end. "If you do a comprehensive test lasting anywhere from 60 minutes to two hours, drilling down to a lot of areas on the site," explains Usability Group's Rubin, "they'll expose 90 percent of your usability issues." Rubin advises running these tests four to five times throughout a Web site's development cycle.