A comparison of broad versus deep auditory menu structures

The primary purpose of this experiment was to gain a greater understanding of the utilization of working memory when interacting with a speech-enabled interactive voice response (IVR) system.

A widely promoted guideline advises limiting IVR menus to five or fewer items because of constraints of the human memory system, commonly citing Miller's (1956) paper. The authors argue that Miller's paper does not, in fact, support this guideline. Furthermore, applying modern theories of working memory leads to the opposite conclusion--that reducing menu length by creating a deeper structure is actually more demanding of users' working memories and leads to poorer performance and satisfaction.

Participants took a working memory capacity test and then attempted to complete a series of e-mail tasks using one of two IVR designs (functionally equivalent, but one with a broad menu structure and the other with a deep structure).

Users of the broad-structure IVR performed better and were more satisfied than users of the deep-structure IVR. Furthermore, this effect was more pronounced for those with low working memory capacity.

Results indicate that creating a deeper structure is more demanding of working memory resource than the alternative of longer, shallower menus.

This experiment has important practical implications for all systems with auditory menus (particularly IVRs) because it provides empirical evidence refuting a widely promoted design practice.