Cost-Justifying Usability: An Update for an Internet Age

Deborah J. Mayhew Deborah J. Mayhew & Associates
Keystroke level modeling (KLM) is one of a variety of cognitive modeling techniques that have been reported in the literature over the last two decades. It is the original modeling technique described by Card, Moran, and Newell in their classic book The Psychology of Human-Computer Interaction (1983). It is the simplest and most basic of a family of modeling techniques (e.g., GOMS, which stands for goals, operations, methods, and selection rules) that have evolved from it over the past two decades (John, 1990, 1995; John and Kieras, 1996; Atwood et al., 1996). Another book provides a good summary of past work in cognitive modeling, including the KLM technique (Carroll, 2003).
KLM, simply put, involves identifying and counting all of the discrete human operations physical (e.g., mouse click, keystroke, or moving the hand from the mouse to the keyboard), cognitive (e.g., read a syllable of text or make a mental comparison) and perceptual (e.g., locate something on screen) that a user must execute to most efficiently accomplish a specific task on a specific user interface design. System response time operators are then added to the model where appropriate and when there are data to estimate them accurately. Time parameters per operator (available in the literature for human operators) (Table 16.1) are then plugged into a task model to predict a total task time. The total task times generated by such models predict the fastest time (on average) that highly trained...