UX research methods — heuristic evaluation¶
Purpose: Systematically review UI against proven heuristics before launch or in audits.
Harvest / review: 2026-05
Agent hint: Load for design critique · Pair with ../ui-design/psychology.md
Nielsen's 10 heuristics (summary)¶
| # | Heuristic | One-line test |
|---|---|---|
| 1 | Visibility of system status | User always knows what's happening |
| 2 | Match system ↔ real world | Language and concepts feel familiar |
| 3 | User control and freedom | Easy undo / exit |
| 4 | Consistency and standards | Same patterns everywhere |
| 5 | Error prevention | Prevent mistakes before they happen |
| 6 | Recognition over recall | Options visible, not memorized |
| 7 | Flexibility and efficiency | Shortcuts for experts, simplicity for novices |
| 8 | Aesthetic and minimalist design | No irrelevant noise |
| 9 | Help users recover from errors | Clear errors + fix path |
| 10 | Help and documentation | Findable help when needed |
How to run a heuristic evaluation¶
- Scope — key flows only (homepage → donate, checkout, signup).
- 2–5 evaluators independently note issues per heuristic.
- Severity — cosmetic / minor / major / catastrophic.
- Prioritize — fix majors on primary flows first.
- Re-test after changes.
Use NN/g's free workbook for structure.
Patterns¶
- Evaluate mobile and desktop separately.
- Include keyboard-only pass for operability.
- Capture screenshots with heuristic tag per finding.
Anti-patterns¶
| Avoid | Why |
|---|---|
| One person, one pass | Misses ~30% of issues vs multiple evaluators |
| Heuristics as legal checklist | Judgment still required |
| Only happy-path flows | Edge cases break trust |
Sources¶
- NN/g — 10 Usability Heuristics — 2026-05
- NN/g — How to conduct heuristic evaluation — 2026-05