KLM-GOMS
Keystroke-Level Model - Goals, Operators, Methods, Selection Rules · Also known as: GOMS Model, KLM
The Keystroke-Level Model (KLM), part of the Goals-Operators-Methods-Selection rules (GOMS) framework, is a computational method for predicting how long a user will take to accomplish a routine task using an interactive system. Developed by Card, Moran, and Newell in 1983, KLM decomposes user actions into primitive operators (keystrokes, mouse clicks, mental preparation, system response waits) with empirically derived execution times, enabling designers to estimate task performance without running user studies.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use KLM for early-stage interface design comparison when you need quick, quantitative predictions without user testing. Best applied to routine, procedural tasks with clear, error-free execution paths. Most reliable for experienced users on familiar interfaces. Do not use for novel, learning-intensive, or error-prone tasks, or when user behavior is highly variable or exploratory.
Strengths & limitations
- Rapid, low-cost prediction of task execution time before prototyping or user studies.
- Produces quantitative, comparable estimates across design alternatives.
- Decomposes tasks into interpretable, reusable cognitive and physical operations.
- Effective for routine, procedural interactions where optimal strategy is clear.
- Limited to routine tasks; does not model learning, exploration, or error recovery.
- Assumes expert, error-free execution; does not reflect novice or realistic variability.
- Difficulty modeling complex cognitive processes; mental operator (M) is coarse-grained.
- Requires accurate task decomposition; errors in step-by-step specification propagate into predictions.
Frequently asked
How do I estimate the mental operator (M) time for complex decisions?
The standard M time is 1.35 seconds, but for complex cognitive operations (e.g., deciding between many options or remembering information), you may extend it to 2–5 seconds or break it into multiple M operators. Document your reasoning and validate against pilot task observations.
Can I use KLM for mobile or touch interfaces?
Yes, but adjust operator times: touch-screen P (point) is typically faster (~0.70 sec) than mouse P, and there is no H (home) operator. Gesture operations may require custom timing derived from pilot studies.
What should I do if my task has branches or error recovery?
KLM is designed for routine, error-free paths. For conditional steps, create separate models for each path and weight by estimated probability. For error recovery, model separately and add as a contingency; this is beyond classical KLM's scope.
Sources
- Card, S. K., Moran, T. P., & Newell, A. (1983). The Psychology of Human-Computer Interaction. Lawrence Erlbaum Associates. ISBN: 0898592437
- Kieras, D. E. (1997). A Guide to GOMS Task Analysis. Technical Report. University of Michigan. link ↗
How to cite this page
ScholarGate. (2026, June 3). Keystroke-Level Model - Goals, Operators, Methods, Selection Rules. ScholarGate. https://scholargate.app/en/human-computer-interaction/klm-goms
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Cognitive WalkthroughHuman Computer Interaction↔ compare
- Heuristic EvaluationHuman Computer Interaction↔ compare
- System Usability ScaleHuman Computer Interaction↔ compare
- Think-Aloud ProtocolHuman Computer Interaction↔ compare