run-heuristic-evaluation

Use 3-5 evaluators who each complete two independent passes: a familiarization walkthrough, then a screen-by-screen hunt using violation questions for each of Nielsen's 10 heuristics. Merge and deduplicate findings afterward, then have every evaluator severity-rate every issue.

A heuristic evaluation broadly sweeps an interface against 10 usability principles, while a cognitive walkthrough tests whether a first-time user can complete one specific task by asking four questions per step. Expert reviews typically combine both methods.

No. Inspection predicts problems while testing observes actual user behavior, and the two methods find different issues. Heuristic evaluation works as a pre-test sweep or a fallback when user access is impossible, and findings must be phrased as predictions.

Nielsen's research recommends 3-5 evaluators because a single evaluator finds only about 35% of problems and different evaluators find different issues. Solo evaluators should disclose the coverage caveat and split passes across days.

Avoid it when you need observed user behavior, WCAG accessibility compliance audits, or severity ratings defined outside the inspection. Accessibility findings go to a WCAG-grounded audit, and severity scoring uses a separate canonical 0-4 scale.