What Heuristic Evaluation Actually Buys You
A single evaluator reviewing an interface against a set of usability principles finds about 35% of its usability problems, according to Jakob Nielsen's research across six projects. Three evaluators together find roughly 65%. Five reach about 75% - and returns drop off sharply after that, since different evaluators tend to catch different problems rather than the same ones twice. That curve is the entire case for heuristic evaluation: no participants, no recruiting, no scheduling - just 3–5 UX professionals independently reviewing a product against a fixed checklist.
Nielsen's 10 Usability Heuristics
Jakob Nielsen and Rolf Molich published the original heuristics in 1990; Nielsen refined them into the ten below in 1994, based on a factor analysis of 249 real usability problems. They're still the reference list three decades later.
- 1. Visibility of system status: the design should always keep users informed about what's going on, through appropriate feedback within a reasonable time.
- 2. Match between the system and the real world: speak the users' language - words, phrases, and concepts they already know, not internal jargon.
- 3. User control and freedom: give users a clearly marked emergency exit for actions taken by mistake, without forcing them through an extended process.
- 4. Consistency and standards: users shouldn't have to wonder whether different words, situations, or actions mean the same thing. Follow platform and industry conventions.
- 5. Error prevention: good error messages matter, but the best designs prevent the problem from occurring in the first place.
- 6. Recognition rather than recall: minimise memory load by keeping elements, actions, and options visible instead of requiring users to remember information across screens.
- 7. Flexibility and efficiency of use: shortcuts - invisible to novices - let expert users speed up interactions the design still supports for beginners.
- 8. Aesthetic and minimalist design: interfaces shouldn't contain information that's irrelevant or rarely needed. Every extra unit competes with the units that matter.
- 9. Help users recognise, diagnose, and recover from errors: error messages should be plain language, precisely state the problem, and suggest a fix - no error codes.
- 10. Help and documentation: ideally the system needs no explanation, but when documentation is necessary, it should be searchable and focused on the user's task.
How to Run a Heuristic Evaluation
- Step 1 - Recruit evaluators: 3 is the minimum worth running; 5 is the point of diminishing returns. Fewer than 3 leaves too much coverage on the table.
- Step 2 - Define scope: the entire product or one specific flow? Define exactly which tasks each evaluator should attempt.
- Step 3 - Independent review: each evaluator goes through the product alone and logs issues against the ten heuristics before comparing notes with anyone. Independent review is what prevents groupthink from collapsing your 65–75% coverage back down toward one person's 35%.
- Step 4 - Consolidate findings: bring evaluators together, group issues by heuristic, and eliminate duplicates.
- Step 5 - Severity rating: rate each issue 0–4 - 0 means you don't agree it's a problem, 1 is cosmetic, 2 is a minor usability problem, 3 is a major problem worth fixing, 4 is a usability catastrophe. Prioritise the 3s and 4s.
- Step 6 - Recommendations: propose a concrete fix for each issue, not just a description of what's wrong.
Limitations to Know
Heuristic evaluation finds problems experts recognise; it doesn't show you how real users actually behave, where their attention goes, or which problems they'd never notice because they'd work around them without complaint. It should complement user testing, not replace it. Evaluators also bring their own biases - the issues they flag aren't automatically the ones that matter most to your actual users.
Key Takeaways
- Coverage scales with evaluator count: 35% for one, ~65% for three, ~75% for five - then it flattens.
- Nielsen's 10 heuristics, refined in 1994 from a factor analysis of 249 real problems, remain the reference checklist.
- Independent review before consolidation is what protects the method's coverage advantage.
- Pair heuristic evaluation with user testing: experts find what's likely wrong, users confirm what actually trips them up.
