Task-based usability tests are perhaps the most important type of usability measurements available. Simply put, they require users to attempt to actually use the item in question. Their failure or success demonstrates the weak or strong qualities of the item.

As always, though, there are many other aspects to this type of measurement. The first requires that a task be devised for the participant. In other words, the participant has to actually do something with the product. This is essential. You would gain nothing by simply handing the item to the participant and then expect them to deliver data about the items usability. However, if you have the participant do something that is realistic (that is, do something that real users would do), then THAT will give you a ton of information.

Tasks could be a variety of things depending on what you wanted to measure. Having a participant install a software product, for example, measures that portion of the software, but it doesn't measure the software's use. Asking the participant to use the software, likewise, will only measure the particular use you are asking them to carry out. Thus, the test is a bit narrow, but the data is normally very rich about that one small portion.

After identifying the task you are interested in, you must find participants (and this really goes for all measures of usability). The participants must closely resemble the real users of the item in question. If the participants do not, the data is faulty and could be called into question.

The final portion is in carrying out the actual test. Here there are many things to consider, but many of these considerations depend on what you are interested in finding out. Common ways of testing people include video and audio recording of the participants as they try to complete the task. While doing this researchers often ask participants to talk aloud while performing the task. Thus, if the participant says "I don't understand" during one portion of the test, you gain a better understanding of where real users may become confused.

For testing software there are many special usability software products on the market. Most of them perform some type of recording of the keystrokes and navigation a participant employs during the test. So, if they must select a specific icon on the screen in order to do something, but they can't find the icon, it will be recorded by the specialty usability software.

Task-based usability testing is a huge subject area, and I'm sure I'll post more on it. But for now, understand that it is one of the most important methods for creating a usable product.