- Plan
- Explore
- Concept
- Evaluate
- Launch
Heuristic Evaluation
Overview
How do you catch usability problems before spending the time and money to put a design in front of real customers?
Heuristic evaluation is a usability inspection method Jakob Nielsen and Rolf Molich developed from their usability-consulting experience: three to five expert evaluators independently judge an interface against a fixed checklist of usability principles, then merge their findings into one prioritized list. It’s the “expert review” technique covered in Usability Evaluation Methods, and the most widely used member of the usability-inspection-method family — run, in the large majority of cases, against Nielsen’s own ten heuristics below.
Three to five expert evaluators independently judge an interface against a fixed checklist of usability principles, then merge their findings into one prioritized list.
Evaluate independently, then merge
Each evaluator reviews the interface alone and records their own findings before anyone compares notes — evaluators who discuss problems together before writing anything down tend to converge on the same subset of issues rather than surfacing a genuinely diverse set. Nielsen’s own early studies found individual evaluators working alone were “mostly quite bad” at spotting problems, which is the reason the method insists on aggregating several independent passes rather than trusting a single reviewer. A full evaluation typically wraps up within days rather than weeks. See Usability Evaluation Methods for how the recommended evaluator count actually scales with expertise (three true experts vs. five or six less-experienced team members) and for the method’s own false-positive caveat.
Evaluator roles: with and without an observer
Running the evaluation without an observer present means each evaluator formalizes their own findings into a written report afterward — more work and interpretation time per evaluator, but no observer to schedule or pay. Running it with an observer present shifts that cost: evaluators narrate their reasoning aloud as they go, and the observer transcribes it in real time, cutting the evaluator’s own workload and write-up time at the expense of the observer’s.
Rate each finding’s severity
A found problem is typically rated on a numeric severity scale by its estimated impact on user performance or acceptance, which is what lets a team triage the merged list — fixing what’s severe first — rather than treating every flagged issue as equally urgent.
Nielsen’s ten heuristics
1. Visibility of system status
Keep users informed about what’s happening through timely, appropriate feedback.
2. Match between system and the real world
Speak the user’s language, with familiar words, phrases, and concepts rather than system-oriented terms, and follow real-world conventions so information appears in a natural, logical order; see Interface Metaphor for the specific case of borrowing an entire real-world domain’s visuals and behavior wholesale.
3. User control and freedom
Support undo and redo, and give users a clearly marked “emergency exit” to leave an unwanted state without an extended process.
4. Consistency and standards
Don’t make users guess whether different words, situations, or actions mean the same thing; follow platform and industry conventions.
5. Error prevention
Eliminate error-prone conditions where possible, or check for them and present a confirmation before users commit to an action. See Error Prevention and Recovery for this heuristic in full — forcing functions, constraints, and detection.
6. Recognition rather than recall
Make objects, actions, and options visible so users don’t have to remember information from one part of the interface to another.
7. Flexibility and efficiency of use
Let experienced users speed up routine actions (accelerators, shortcuts) invisible to novice users, so the interface serves both.
8. Aesthetic and minimalist design
Interfaces shouldn’t contain information that’s irrelevant or rarely needed; every extra unit of information competes with the relevant ones.
Interfaces shouldn’t contain information that’s irrelevant or rarely needed; every extra unit of information competes with the relevant ones.
9. Help users recognize, diagnose, and recover from errors
Error messages in plain language (no codes), that precisely indicate the problem and suggest a constructive solution. See Error Prevention and Recovery for recovery mechanisms beyond messaging, like Undo and reversible defaults.
10. Help and documentation
Even though a system should ideally work without documentation, provide it where necessary: easy to search, focused on the user’s task, listing concrete steps, and not too large.
Mapping to the interface design principles
Most of these map directly onto the seven interface-design principles, which were synthesized from this list alongside Ben Shneiderman’s eight golden rules and Edward Tufte’s writing on information presentation: heuristics 1, 3, 4, 5, 6, 8, and 9 each correspond closely to a same-numbered principle there. Two heuristics don’t have a direct one-to-one match in that shorter list and are worth keeping in view on their own — match between system and the real world (heuristic 2), and help and documentation (heuristic 10).
Most of these map directly onto the seven interface-design principles, which were synthesized from this list alongside Ben Shneiderman’s eight golden rules and Edward Tufte’s writing on information presentation: heuristics 1, 3, 4, 5, 6, 8, and 9 each correspond closely to a same-numbered principle there.
Other named heuristic sets
Nielsen’s ten aren’t the only checklist in use. Gerhardt-Powals’ cognitive-engineering principles take a more holistic approach centered on reducing Cognitive Load and presenting information meaningfully, and the Weinschenk and Barker (2000) classification catalogs 20 separate heuristic types. Neither has displaced Nielsen’s list as the default, but a generic checklist — Nielsen’s included — can’t catch problems specific to one specialized domain or culture, which is what has driven the development of domain-specific and culture-specific heuristic sets alongside the generic ones.
Advantages and limitations
Heuristic evaluation needs no user recruitment, scheduling, or payment, which keeps it cheap and fast relative to the formal studies covered in Usability Evaluation Methods — a real advantage for early-stage design, when a running prototype to test with real users may not exist yet. The tradeoff is that results are only as good as the evaluators doing the judging: quality is expert-dependent, genuinely qualified evaluators can be hard to find, and a finding reflects the reviewer’s personal judgment rather than empirical data gathered from the people who’ll actually use the system. The method also tends to surface a different set of problems than performance testing does, rather than a strictly overlapping one — one more reason Usability Evaluation Methods treats expert review, informal evaluation, and formal studies as complementary rather than substitutes.
Related Concepts
Patterns
Principles
- Interface Design Principles
- Undo
- Recognition Over Recall
- Keyboard Shortcuts
- Interface Metaphor
- Error Prevention and Recovery
- Cognitive Load
Processes
Further reading
Jakob Nielsen’s “10 Usability Heuristics for User Interface Design” (NN/g, updated 2020 — commercially published, no stated open license) is the canonical statement of the ten heuristics this page evaluates against, illustrating each with a physical-world analogy — a mall’s “You Are Here” map for visibility, stovetop-control layout matching burner placement for match-between-system-and-the-real-world — and tying the slip/mistake distinction directly into the error-prevention heuristic.
Sources
Heuristic evaluation (Wikipedia) is this page’s sole source — Nielsen and Molich’s ten heuristics, the method’s independent-evaluation-then-merge procedure and its “mostly quite bad alone” finding, the with/without-observer roles tradeoff, severity rating, the other named heuristic sets (Gerhardt-Powals, Weinschenk and Barker), and the method’s advantages and limitations.