Learnability of user interfaces

This week's readings touched a very interesting topic - The learnability of a user interface. I believe that among all the metrics of an interface that are studied and analyzed, learnability is perhaps the most difficult one to measure.

In my opinion, the diffculty in measuring learnability can be attributed to several reasons

The innovativeness of in interface. The logical grouping of interface elements. The dependency of one section of the interface on the other - e.g. contextual information. The familirarity of the user with similar interfaces. The expectations of the user. The different interpretations of the different terms in the interface to different types of users. The personalizability of an interface. The resemblance of portrayal of the information in the interface to the way the task is performed in real life - e.g. following the same sequence of actions. Many of the reasons stated above cannot be directly measured in usability tests. Sometimes it may so happen that after using a very simple and innovative interface for a while, users may come up with ideas to combine the different services of the interface to make things even simpler. For example, using macros in excel, or in text editing softwares like Notepad++.

For situations where learnability of an interface is critical, it is imperative to release a product in phases and gather feedback in the form of suggestions from the users. This is exactly the strategy that is adoped by several companies when they release alpha, beta and release candidate versions of their applications. Doing so gives them three advantages

They are able to market the product due to its free nature in the beta phase. It gives their prospective clients and early adopters a chance to learn the product well before hand. It helps in gathering valuable feedback and make enhancements before the product goes mainstream.

Learnability is tough to measure, no matter what technique you adopt. However, in my opinion, iterative development based upon feedback is one of the best practical ways to make something learnable gradually. The other way of course would be - to innovate, but then again, you would still be unsure if your innovation is actually as usable at you claim it to be until it reaches the end user.