Modeling Tools and Software
Most of the important usability work can be completed with a simple office suite. It may help to have a flowcharting package, and you also need software for graphics work, but that's about all the software you need. You also need to be able to use a word processor to document meetings and descriptions, and you need a tool to mock up screens and pages. Which tool you use is not as critical as making sure that the usability staff members are comfortable with the tool and that they do not get distracted or waste time writing "code" to make the screen mock-ups work. Some people prefer a graphics program like Adobe Photoshop, but a presentation tool like Microsoft PowerPoint works just as well. Some usability staff are already facile in a tool such as Microsoft Visio. Whatever tool your staff members already know how to use that allows them to quickly mock up screens and pages is the best tool to use.
Sophisticated modeling tools may or may not be necessary. Available software can assist in the development of very large taskflows and the modeling of taskflow behavior. An example of this type of software is Micro Saint (a product made by Micro Analysis and Design, Inc.), which supports task modeling. I have seen this software used to good effect in very complex and critical applications, especially in the military design arena. However, I have yet to see this software used to make a difference to commercial Web sites or applications.
Limited modeling tools are available for usability work. You may wish to create your own. At our company, we built an application called the Task Modeler. It is basically a specialized spreadsheet that helps add up the number of clicks, mouse movements, and keystrokes used to complete a task. Using this application on a group of tasks representative of the work to be completed on a given interface provides a good indication of the time it will take an expert user to use the software. This data is important because when measuring the speed of task completion during a usability test, you're measuring only how fast users are during their initial usage, not how fast they would be after extended experience with the interface. During the test, users spend only minutes with the software, so they will not be experts on using the interface. Yet there are many cases when you are designing for expert users. Also, you don't want to make the classic blunder of designing for first usage only. For example, a menu design that can be used easily and quickly by novices is a much better alternative to using commands initially, but you may then find the commands are faster once learned. If the software will be used full time, going with the menu can be a million-dollar mistake. In this case, we built our own tool.
Many companies have purchased tools to track Web sites and provide feedback and statistics on usage. Some of these tools claim to provide usability information and are useful for performing quick checks and validation. For example, there are tools that let you know if your alt text tags are missing (accessibility tools) or if you are using too long a line length. Be wary, however, of tools that track download times for a page, or how many users clicked on a page, or how much time people spent on a Web page, independent of other information. While this information can be useful to know, it can also be misleading. Why did a user spend only 3 seconds on a page? Was it because (a) the page is poorly designed, (b) the page is well designed and the user got what he or she wanted right away, or (c) the page before was poorly designed, so the user clicked on the wrong link? You cannot tell any of this just by reading a report on where people went and how long they stayed. Nothing can replace a trained usability professional evaluating a screen or page or watching and interpreting users performing a task.