Navigation on the home page is going to set the tone for your whole site. Is your site well classified? If people are looking for a Palm PDA, will they find it in Computers, Electronics, or Office Gear? Where are people most likely to look for it?

In their paper "SunWeb: User Interface Design for Sun Microsystems' Internal Web" (www.useit.com/papers/sunweb), Jakob Nielsen and Darrell Sano described four different usability studies to help them determine the navigational tools for the Sun Microsystems internal Web site:

1. Card sorting to discover categories

2. Icon intuitiveness testing

3. Card distribution to icons

4. Thinking aloud walk-through of page mock-up

The team wrote up fifty-one types of information that might be found on the internal Web site. They wrote each on note cards and scattered them across a table. The experiment subjects were asked to sort these cards into piles based on similarity. As a result, they created proper groupings for information categories for the home page.

The resulting categories required icons or pictograms to represent them on the home page. Once the first draft of these icons was created, it was submitted to the icon intuitiveness test. Subjects were asked to interpret each unlabeled icon to see how understandable it was. Most were good or excellent, but several had to be substantially altered.

The two tests were then combined. The new icons were printed and handed to subjects, who then had to place them in the proper categories on the table. This test ensured that users would associate the correct concepts with each of the general groups previously defined. The "thinking aloud" study simply asked subjects what sort of information they would expect to find behind each of the icons.

The conclusions from this project are that a uniform user interface structure can make a Web significantly easier to use and that "discount usability engineering" can be employed to base the design on user studies even when project schedules are very tight. That was back in 1994 and it is still a valid method today. It's cheap, too. On the other hand, you could wander over to the National Institute of Standards and Technology Web site and get your hands on a WebCAT (http://zing.ncsl.nist.gov/WebTools/index.html). As the site describes:

The Web Category Analysis Tool (WebCAT) allows a web designer/usability engineer to test a proposed or existing categorization scheme of a Web site to determine how well the categories and items are understood by users.

Functions and Features

WebCAT is a variation on traditional card sorting techniques. Its process of categorizing and analyzing information is interactive. First, the usability engineer uses the WebCAT setup interface to specify the design of an experiment; she provides the names of the items, a category method, instructions for the subjects, a description of the study, and a variety of other options. WebCAT then creates the exercise for the user which is implemented as a Java applet that provides a satisfying interactive experience. The subject merely drags items from a list to one of several category bins. Items may be moved back and forth between bins until the subject is satisfied with his results, then enters his name, makes an optional comment, and presses "Done."

The usability engineer can view the collected data at any time using a new interface which implements a clustering algorithm.

It's a good idea to measure where your visitors think content should reside.

If you took the time to ask each visitor, you would find that almost every one cringes at being subjected to your all-singing, all-dancing, animated greetings on the pre-home page, also known as the splash page.