You have to provide your friend with driving directions to come to your home. All of a sudden, your are struggling. Is it two or three stops before turning right? Providing a direction can be very difficult even if you take that route everyday.
Cognition is distributed. Because our memory is optimized, our knowledge is ingrained within our environment. We retain only what is necessary and use the environment to retrieve sequences.
Recalling the direction to drive home is no problem when you are driving home. On the other hand, trying to recall an experience out of context is prone to error. Consequently, trying to gather information about a process and human work or evaluate a solution out of real context of use is wrong.
That’s why we have to be very careful about laboratory studies. In one of our recent usability studies for an online bill payment application, we initially found users were able to accomplish their tasks efficiently and they expressed their appreciation. At a later phase, during testing in real work setting, we discovered users still had to reconcile check numbers with bank statements. In this context, online payment created more trouble than benefit. Initial laboratory studies didn’t provide the right insight into these problems.
Evaluating usefulness (real practical advantage) out of context is as difficult as giving driving directions out of context.
Since users cannot really recall or imagine how they will use a product in real work setting as demonstrated above, it is only by observing usage in real work setting we can understand how a product will be used.
Many disciplines and authors such as Dan Diaper (by the way, he is currently looking for a job?) stress how important it is to study work and behavior in real work setting. Our practice has demonstrated over and over how essential it is. You may also read about the Cognitive Approach.
While usability can be evaluated in a laboratory, usefulness has to be evaluated in real work setting.