Usability studies provide a way to determine whether the software works as you had intended by observing what happens when users actually attempt to use the product. Where do they make mistakes, and what kinds of mistakes occur? Where do they get confused or become hesitant?

Usability testing should generally be used for any novel or complex product. However, any complex activity can be submitted to usability testing. For example, it may be useful to do a usability test on assembly instructions or installation procedures.

Here are some suggested procedures for usability testing:

1. Find or create a laboratory. It would be useful to have (a) one-way glass for unobtrusive observation, (b) videocameras to record both actions and facial expressions, and (c) timing devices. The purpose of using a laboratory is to ensure that each user in the study confronts the same testing conditions.

2. Determine your tasks. You cannot just tell people to “please interact with the product.” Instead, you need to decide on a set of tasks that the user will attempt. These tasks should be grounded in the primary applications for the product or should relate to troubling or resolved design issues, as identified by prior market research.

3. Recruit users. As always, the sample of people studied should reflect the intended population of users. Do not rely only on company employees who are comfortable with the product. Select for a wide range of  skills and experiences.

4. Evaluate your results. Delays, errors or unexpected responses can all be used to diagnose problems with the design. Have the design people involved with the evaluation of the results, as it is they who will have to make the changes.

More companies would benefit from submitting their new products/services to usability testing.