Usability heuristics

When to use

• A finding is about to be justified by naming Nielsen, Norman or Shneiderman, or by a term that moved between editions — affordance, signifier, progressive disclosure — or by a number: five users, 3-5 evaluators, 85%, 7±2, three clicks.
• Before running a heuristic evaluation, or reporting one as findings.

When NOT to use

• Reaching for an empirical principle instead. Fitts, Hick, Miller, Gestalt, von Restorff and the aesthetic-usability effect belong to empirical-design-principles, linked below: those predict, and fail by predicting wrongly.
• Picking token values, or judging craft quality — see Cross-references.

The gate: provenance, then reliability