When I evaluate website-styled graphic interfaces, I apply the Bastin/Scapin and Nielson heuristic methods, among others. I also analyze the system’s structure using an information architecture. Thus, every on-screen page has an equivalent schematic in the architecture.
The product and its model therefore correspond. Both are two-dimensional, and the application and its image can be navigated in similar ways; the user’s actions may not follow one-another or be interconnected.
We were recently asked to review a voice interface system: an automated telephone information system.
Unlike graphic applications, which come with a patent visual support, the information users receive while listening to vocal segments on a telephone system are not as easy to fully retain. Graphic representation, therefore, is a must. By building schematics for the messages’ structure, navigation and content, we were better able to understand the user’s experience throughout the process.
That being said, a conflict can arise at this point of the evaluation: the product and its model no longer correspond, as the voice application is being studied using a visual support.
Thus, what we see in the schematics is heard sequentially by the user; the instructions follow one another, so callers can’t navigate the application like they would on a website.
It could almost be said that schematics tend to reduce the auditory, mnesic and attentive effort users make, and to restrict the application’s flow.
Given the independent nature of auditory applications, I always try to remember that every message or option I propose will require the user to apply more cognitive effort to hear, listen, assimilate and understand. The same logic applies to the design phase: all changes to the application should be tested on real users to determine their “auditory relevance”.
In summary, we should always give careful consideration to the type of support used during evaluations, and customize our analysis to the type of interface being studied and the cognitive demands it places on its users.
0 Comment(s)