Heuristic evaluation (Wikipedia)
Wikipedia’s article on heuristic evaluation, the usability inspection method Rolf Molich and Jakob Nielsen developed from their usability-consulting experience: expert evaluators review an interface against a list of established usability principles rather than testing it with real users.
License: CC BY-SA 4.0 — https://creativecommons.org/licenses/by-sa/4.0/
Key points
Nielsen’s ten usability heuristics: visibility of system status; match between system and the real world; user control and freedom; consistency and standards; error prevention; recognition rather than recall; flexibility and efficiency of use; aesthetic and minimalist design; help users recognize, diagnose, and recover from errors; and help and documentation.
Conduct: Nielsen recommends three to five evaluators. Each works independently before results are aggregated, specifically to reduce group confirmation bias — evaluators who discuss findings together before recording their own tend to converge on the same subset of problems rather than surfacing a diverse set individually. Nielsen’s original report found individual evaluators were “mostly quite bad” at spotting problems alone, which is the reason the method insists on several independent passes rather than one.
Evaluator roles: without an observer present, evaluators write up their own findings as a report — more work per evaluator, but cheaper to run. With an observer present, evaluators narrate their reasoning aloud while the observer records it — less effort and interpretation time for the evaluator, at the cost of the observer’s own time.
Severity rating: problems found are typically categorized on a numeric scale by their estimated impact on user performance or acceptance, so a team can triage which findings to fix first.
Other named heuristic sets exist beyond Nielsen’s: Gerhardt-Powals’ cognitive-engineering principles (emphasizing cognitive load and meaningful information presentation), and the Weinschenk and Barker (2000) classification of 20 heuristic types. Domain-specific and culture-specific heuristic sets have also been developed, since a generic list can’t catch problems specific to one specialized domain or culture.
Advantages and limitations: the method needs no user recruitment and is cheap relative to formal testing, but its results are only as good as the evaluators doing it — quality is expert-dependent, qualified evaluators can be hard to find, and findings reflect personal judgment rather than empirical data. It also surfaces a different set of problems than performance testing does, rather than a strictly overlapping one.