Usability Heuristics
UI/UX Design Principles
Chapter 7 · Usability Heuristics
Chapter 2 covered testing with real users. This chapter covers a real, different, and complementary method: expert review against a fixed, well-established checklist — Nielsen's own 10 usability heuristics.
A Real, Dated History
Jakob Nielsen and Rolf Molich originally developed a set of usability heuristics in 1990. Nielsen refined them into their current, final form in 1994, via a real factor analysis of 249 actual, documented usability problems. The resulting 10 heuristics have remained unchanged since 1994 — a genuinely long track record for a design-research artifact.
The 10 Heuristics
Visibility of System Status
Keep users informed about what's happening, with timely feedback.
Match Between System & Real World
Use familiar language and real-world conventions, not internal jargon.
User Control & Freedom
A clearly marked exit for unwanted actions — real undo and redo.
Consistency & Standards
Don't make users wonder whether different words mean the same thing.
Error Prevention
Design out error-prone conditions rather than relying on error messages.
Recognition Rather Than Recall
Make options visible — don't force users to remember information.
Flexibility & Efficiency of Use
Accelerators for expert users, without punishing novices.
Aesthetic & Minimalist Design
Every irrelevant unit of information competes with the relevant ones.
Help Users Recognize & Recover From Errors
Plain-language errors that state the problem and suggest a fix.
Help & Documentation
Searchable, task-focused help, available in context when needed.
Heuristic Evaluation as a Method
A heuristic evaluation is the real, structured process of one or more expert reviewers independently assessing an interface against these 10 heuristics, flagging every real violation they find. Nielsen's own real, specific recommendation: three to five independent evaluators, since any single reviewer — no matter how experienced — reliably misses a real share of the issues another evaluator would catch.
Two Methods, Compared
| Method | Who's involved | What it's good at | What it structurally misses |
|---|---|---|---|
| Heuristic Evaluation | 3-5 UX experts, independently | Fast, cheap, catches well-known usability patterns early | Real user context, motivation, and genuinely novel problems no heuristic anticipates |
| Usability Testing (Ch. 2) | Real users, attempting real tasks | Real behavior, real confusion, real task success/failure | Slower to run; needs a real or prototype interface to test against |
Hands-On Exercises
An interface shows the exact same word ("Delete") for two functionally different actions on two different screens — one permanently removes data, the other only hides it temporarily. Explain, in your own words, which of Nielsen's 10 heuristics this violates, and why.
📄 View solutionA manager says "we don't need to bother with real user testing — we already had one senior designer do a heuristic evaluation and they're very experienced." Explain, in your own words, why this reasoning doesn't hold up, using this chapter's own real recommendations.
📄 View solutionA colleague says "testing with 5 users and doing a heuristic evaluation with 5 experts are basically the same thing, since they use the same number." Explain, in your own words, why this is incorrect, citing the real, different thing each method actually measures.
📄 View solutionChapter 7 Quick Reference
- Nielsen & Molich's real 1990 heuristics, refined by Nielsen in 1994 via factor analysis of 249 real usability problems — unchanged since
- 10 heuristics: system status, real-world match, user control, consistency, error prevention, recognition over recall, flexibility, minimalism, error recovery, help/documentation
- Heuristic evaluation — 3-5 independent expert evaluators reviewing against the heuristics, since any one evaluator misses real issues
- Real, important distinction: heuristic evaluation's "3-5 evaluators" and Chapter 2's "5 test users" are different concepts measuring different things — heuristic evaluation can't replace real user research