Thatâs where usability consultants and software come in. Consultants such as the Nielsen Norman Group, Fremont, Calif., and Human Factors International, Inc., Fairfield, Iowa, observe how real users interact with your site and suggest changes to make it easier to use. Pricing depends on the level of detail. For example, for $99, consultant Abacus Web Services of Bradenton, Fla., will conduct a quick evaluation and make some suggestions. At the other end of the spectrum, for $1 million a year, Zentropy Partners, Los Angeles, helps big e-commerce sites improve their usability.
In addition, there are software tools that companies can buy to automate usability testing. These systems, which test factors such as download times and site search engines, range from $299 to about $40,000.
Staples Inc., Framingham, Mass., signed on with Human Factors International in 1999 and, with its guidance, relaunched its Web site in May 2000. In the following quarter, revenue was up almost five-fold on a year-over-year basis. Also, Staples.com attracted 300,000 repeat customers in that quarter, compared with 180,000 in the previous quarterâan increase of 67%. âI think that HFI, working with our internal team, was responsible for those sorts of customer metrics,â said Colin Hynes, director of usability for Staples.
Staples has now hired another consultancy, behavioral and social science researchers American Institutes for Research, Washington, D.C., to visit customer workplaces to find out how individuals use Staples.com to order supplies, Hynes said.
But hiring outside consultants is not for everyone.
âThat functionality is too critical to us; we never look outside for it,â said Phil Gibson, VP-Web business for National Semiconductor Corp., Santa Clara, Calif. The technology company relies on detailed e-mail surveys of customers to track usability. This keeps the information in-house, he explained.
Watching the user
User testing is central to usability testing. In some cases, usability experts gather eight to 10 users in a room to observe their activities on the site. The users are given assignments representative of tasks that would typically be performed on a site, such as placing an order or looking up some information. Then, the experts observe how quickly the users are able to perform the tasks. The test subjects also are interviewed.
Software tools may automate some of this process. For example, a system from Maxamine Inc., San Ramon, Calif., downloads an entire Web site, then measures the site for factors such as how long each page takes to download. Also measured is the âdepthâ of a siteâhow long it takes to get from the home page to any particular page within the site.
Maxamineâs research indicates that shallow sites are better for e-commerce, because they make it easier for customers to make purchases. So-called deep sites are better for training, taking users step-by-step through a process.
Vividence Corp., San Mateo, Calif., provides software that allows test subjects to download an Internet Explorer plug-in that, for the duration of the test, tracks each page they visit, takes a screenshot of each page and allows the user to write comments about the pages.
As with other usability testing, consultants give subjects a task, then observe how well they perform it. In this case, the users work at their home or office PCs, and the tests are conducted over the Internet.
Because these tests are conducted remotely and automatically, Vividence says it can draw from a larger group of test subjects than other methods. Also, the remote nature of the tests allows analysts to assess differences among users located in different parts of the world, Vividence said.
Consultant Jakob Nielsen, a principal of Nielsen Norman Group, cautioned against relying on usability software.
âMost automated tools are voodoo usability [mechanisms] that detract attention from the simpler and cheaper idea of just sitting next to customers and watching them,â he said. âSome companies will do anything to avoid having a customer in the building. But ultimately you are better off getting in direct touch with real human beings instead of sticking pins into dolls or other abstractions of reality.â
But John Marshall, CEO of usability software provider ClickTracks Inc., Sunnyvale, Calif., argued that software provides lots of useful information relatively cheaply. The $500 ClickTracks system graphically shows usersâ traffic patterns, including which links visitors click on and where they leave a site entirely. This information can tip Webmasters off to potential usability problems.
âItâs not a replacement for a Jakob Nielsen for $40,000 [analysis]. Weâre not going to tell you everything you need to know,â Marshall said. âBut people can be productive with ClickTracks on their lunch break. They can download it, install it, import their log files and get some productive data within 30 minutes with a sandwich in their hands.â