**Redesign of Web Site May Be Critical For SuccessNEW YORK -- The relaunch of a Web site often results in a 50 percent increase in the number of visitors to the site, according to Jakob Nielsen, principal and co-founder of the Nielsen Norman Group, a user-experience think tank and consulting firm in Silicon Valley. Nielsen made the announcement yesterday at the User Experience World Tour 2000/2001 in New York City.
Even the slightest change to a site results in improved usability by as much as 50 percent, Nielsen told a crowd of more than 400 people at the Grand Ballroom at the Manhattan Center. In some cases, revamped e-commerce sites saw increases of 400 percent in sales.
"The point is, it doesn't go up a little, it goes up a lot," Nielsen said of revamped sites. "You get a little better [and] you're getting a lot more traffic. It's due to that competitive [Web] environment. If you have two sites and one is 10 percent better, you spend 100 percent of your time there and no time at the site that's 10 percent worse."
When Web sites make their initial launches, they often present designs -- though not deliberately -- that are confusing for users, who get lost and cannot find products. The objective is to give good customer service and make the site simple, Nielsen said.
"The first Web site a company does is almost always bad, and that's why we get this need for redesign," Nielsen said. "They had to first release the bad site that didn't work and then realize that, 'Oops, there is such a thing as usability as well, and we have to make it easier to use and appealing.' "
Finding out what works on the Web has revealed a new set of problems that traditional marketers are not accustomed to facing. They have to research not only whether visitors like a particular product, but also whether the Web site is usable. To do that, marketing managers must conduct test groups, Nielsen said. He recommended testing users in small groups of four to five people for a shorter period of time to get faster and better results.
"The question is, how fast can you get the information and how often can you do the testing? If you have a budget to test 40 or 100 people, that's wonderful," Nielsen said. "However, if you do have a budget for 100 people, I would rather do 20 studies each with five people. In between each study you change the design, because what happens is, after the first five people you do know that there are weaknesses in the design."
A marketer can conduct one test with 100 people to get a precise measure of how unusable a site is. But the marketer already knows this, Nielsen said. What is needed is an understanding of what needs to be changed to improve the site.
"When you think you improved the site, maybe you didn't, and that's why you need to test again," Nielsen said. "It's not enough to say, 'Let's test it, let's get the data and let's fix it.' No, it is, 'Let us hope we fixed the problem and then test again to see if we have.' Every single time you test, you get better and better."
Frequent testing is a superior way of getting better data on improving a Web site rather than one large study that gives a bad report with detailed numbers and takes months to do when the data is no longer relevant.
"A lot of companies are beginning to do this. They are really getting more focus on what needs to be done," Nielsen said. "On the other hand, it's still a very small percentage of the overall Internet community that's doing these types of testing."