UX Foundations · Step 3 · 4 min read
Heuristic evaluation
Reviewing a design against ten well-known rules of thumb finds many problems before any user sees it. Nielsen's ten usability heuristics in their current wording, with an everyday example of each, how to run an evaluation with a few colleagues, and how to rate how serious each problem is.
You don't always need users to find usability problems. Some are so common that experts have written them down as rules of thumb: show people what's happening, speak their language, let them undo, prevent errors. Checking a design against these heuristics is fast, cheap and finds many problems early. It doesn't replace testing with people, but it means your tests can find the harder problems instead of the obvious ones.
What you'll be able to do after reading:
- Name Nielsen's ten usability heuristics, with an example of each
- Run a heuristic evaluation with three to five evaluators
- Rate how severe each problem is
- Know what heuristic evaluation can't tell you
1. The ten heuristics
Jakob Nielsen published them in 1994; NN/g revised their names and descriptions in 2020 (NN/g, 10 Usability Heuristics for User Interface Design). The method itself comes from Nielsen and Molich (1990).
| # | Heuristic | In everyday terms | An example of breaking it |
|---|---|---|---|
| 1 | Visibility of system status | Show what's happening | A "Save" button that gives no sign anything was saved |
| 2 | Match between the system and the real world | Speak the user's language | "Error 0x80070005" instead of "You don't have permission" |
| 3 | User control and freedom | Let people undo and leave | No way to cancel a long form without losing everything |
| 4 | Consistency and standards | Same words and patterns everywhere | "Delete" on one screen, "Remove" on another, for the same action |
| 5 | Error prevention | Stop mistakes before they happen | Letting people book a return flight before the outbound one |
| 6 | Recognition rather than recall | Show options, don't make people remember | Asking for an order number that's only on a previous screen |
| 7 | Flexibility and efficiency of use | Shortcuts for experts, simple for beginners | No way to repeat last week's order |
| 8 | Aesthetic and minimalist design | Only what matters, nothing extra | A login page with six banners and two pop-ups |
| 9 | Help users recognize, diagnose, and recover from errors | Clear error messages with a way out | "Invalid input" with no hint which field or why |
| 10 | Help and documentation | Help when needed, where needed | A complex setting with no explanation anywhere |
2. Running an evaluation
- 1Choose
A task and the screens it uses.
- 2Review alone
3 to 5 evaluators, each with the ten heuristics.
- 3Merge
Combine the lists; remove duplicates.
- 4Rate
Severity 0 to 4 for each problem.
- 5Fix
The worst first; then test with users.
NN/g's advice: "Ideally, three to five people should independently evaluate the same interface", because one person finds only some of the problems (NN/g, Moran and Gordon, How to Conduct a Heuristic Evaluation, 2023). Each evaluator reviews alone first; then the group merges the lists. NN/g offers a free workbook with that article.
3. Rating severity
Nielsen's severity scale helps decide what to fix first (NN/g, Severity Ratings for Usability Problems):
| Rating | Meaning |
|---|---|
| 0 | "I don't agree that this is a usability problem at all" |
| 1 | "Cosmetic problem only" |
| 2 | "Minor usability problem" |
| 3 | "Major usability problem" |
| 4 | "Usability catastrophe: imperative to fix this before product can be released" |
Consider how often it happens, how much it hurts, and whether people can work around it once they know.
4. What it can't tell you
Heuristic evaluation finds problems experts can predict. It can't tell you whether people need the feature, whether they understand your domain words, or how they behave in their real context. Use it before testing, not instead of it. See Usability testing with five people.
5. Common mistakes
- One evaluator, so many problems are missed.
- Listing problems without the heuristic, the location and a suggested fix.
- Treating every finding as equally important.
- Using heuristics as personal taste ("I don't like the colour").
- Skipping user testing because "we already did a heuristic review".
6. Vocabulary
| Term | Meaning |
|---|---|
| Heuristic | A rule of thumb for good usability |
| Heuristic evaluation | Reviewing a design against a set of heuristics |
| Evaluator | A person who reviews the design |
| Severity rating | How serious a problem is, from 0 to 4 |
7. Try it
Do the Heuristic review of a real app lab: pick an app you use, check three key screens against the ten heuristics, and rate what you find.
Reflection questions: Which heuristic did the app break most often? Which problem would you fix first, and why?
8. If you already practise
- Add your design system's rules as extra "heuristics" for your own product.
- Run a heuristic review before every usability test, and check later whether users hit the same problems.
- Keep a shared list of recurring problems; it often reveals a pattern in your design system.
Checklist
Learn more
- NN/g: 10 Usability Heuristics for User Interface Design (1994, revised 2020) · How to Conduct a Heuristic Evaluation (2023) · Severity Ratings for Usability Problems (1994)
Books, courses and videos to go deeper
- Usability EngineeringJakob NielsenBook · Morgan Kaufmann, 1993 (Amazon)
- Don't Make Me Think, RevisitedSteve KrugBook · New Riders, 2014
Book links go to Amazon. As an Amazon Associate I earn from qualifying purchases, at no extra cost to you.
Related
How people think · Usability testing with five people · Heuristic review of a real app · AI design review process