Observation of Mental Models

This was, for me anyway, a very "real" session, in that I was observing a user interacting with a system I'd created for a user community. Although Kurt isn't really from my user community, observing him gave me many insights I would never have gotten from reasoning about the organization of the web site or from introspecting about my own experiences with it.

As someone intimately familiar with the site, I usually don't bother to search but follow links directly to where I want to be. I had assumed, however, that most beginning users would be best served by "search", though Kurt's experience suggests it's not difficult to form a model of the site sufficient for effective browsing.

Some conclusions about the web site:

• The fact that there is a naming convention is generally helpful, but can totally mess things up when it isn't followed 100%
• A naive user will quickly form models of what is where (even if, as is probably the case with many web sites, he has no particular model when he first visits the site).
• If these preliminary models are reasonably accurate, the user will do well. If not, he will be somewhat crippled until he unlearns those models and learns more "correct" ones.
• Therefore, the site should be structured so that initial impressions tend to accord well with the underlying organization. I had thought that this was the case, but not always. For example, when he found listings for several manufacturers under "information", Kurt seemed to conclude that the information area was organized entirely by manufacturer. He later had to unlearn this to find information on drum machines.