Heuristic Evaluation: Nielsen’s 10 Usability Heuristics With Real Examples
3 Views
Summarize Article
Related Blogs
Summarize Article
Three to five evaluators, each spending one to two hours alone with an interface. That is the setup the Nielsen Norman Group recommends for expert reviews, and it finds a large share of usability problems before a single user test.
A heuristic evaluation is an expert review in which trained evaluators inspect an interface against a set of usability principles, called heuristics, and log every place the design breaks one. The most widely used set is Jakob Nielsen’s 10 usability heuristics. He developed them with Rolf Molich in 1990 and refined them in 1994.
This guide explains each heuristic with a real product example, then shows how to run the review, rate severity and write up findings with a template you can copy.
A heuristic evaluation in UX is a fast, low-cost inspection method where several reviewers check an interface against established principles, working on their own before merging results. It needs no participants or lab. The output is a list of problems, each tied to a heuristic and rated for severity.
The method sits inside a wider review. A full UX audit combines this expert review with analytics and accessibility checks. The review explains the design causes behind problems that analytics can only point at.
Usability itself has a formal definition: ISO 9241-11 describes it as the extent to which specified users can reach specified goals with effectiveness, efficiency and satisfaction in a specified context. Heuristics are shortcuts for judging those three qualities without running a study.
Nielsen’s 10 usability heuristics are broad rules of thumb for interface design, first published in 1994 and still used largely unchanged. They cover everything from system feedback to help content. The table pairs each with a familiar example, so you can spot the same patterns in your own product. Keep it open during reviews.
| # | Heuristic | Plain meaning | Real example |
| 1 | Visibility of System Status | Keep users informed about what is happening | Upload progress bars that show percent complete |
| 2 | Match Between the System and the Real World | Use the user’s words and conventions | A trash can icon for deleted files |
| 3 | User Control and Freedom | Offer a clear way out of mistakes | An undo option right after sending an email |
| 4 | Consistency and Standards | Follow platform and internal conventions | The cart icon sitting top right on most stores |
| 5 | Error Prevention | Stop errors before they happen | Disabling a submit button until required fields are filled |
| 6 | Recognition Rather than Recall | Make options visible so users need not remember them | Recently viewed items on a shopping site |
| 7 | Flexibility and Efficiency of Use | Offer shortcuts for frequent users | Keyboard shortcuts in email and design tools |
| 8 | Aesthetic and Minimalist Design | Remove anything that competes with key content | A checkout page without the main site menu |
| 9 | Help Users Recognize, Diagnose, and Recover from Errors | Explain errors in plain words with a fix | “Card declined. Try another card or contact your bank.” |
| 10 | Help and Documentation | Provide help that is easy to search | In-app help panels tied to the current screen |
The NN/g heuristics article gives the full wording and was last reviewed on January 30, 2024. The Nielsen heuristics are deliberately general, which is why they still apply to apps, voice interfaces and AI chat tools.
Run a heuristic evaluation in five steps, from scoping to a joint severity rating, with 1 to 2 hours of solo review per evaluator before anyone compares notes. Independent review comes first on purpose. Evaluators who work together early tend to anchor on the first problem anyone spots.
Reviewers can come from outside UX. Experience still helps. A mix of one or two designers with a developer or product manager often catches different classes of problems.
Plan for about a week end to end on a medium-size product. The independent reviews take an afternoon each, and merging plus rating fits into a 4-hour workshop.
Want an expert heuristic review of your product?
WebOsmotic designers review your key flows independently against all ten heuristics and deliver a severity-rated fix list in about 5 working days.
Rate each problem on Nielsen’s 0 to 4 severity scale, considering how often it occurs, how badly it hurts users and whether it persists after users learn the interface. A rating of 4 means fix before release. A rating of 1 means cosmetic. Agree the rating as a group after merging.
| Rating | Meaning in Nielsen’s severity scale | Typical action |
| 0 | Not a usability problem | Drop it |
| 1 | Cosmetic problem | Fix if time allows |
| 2 | Minor problem | Low priority |
| 3 | Major problem | High priority, next release |
| 4 | Usability catastrophe | Fix before release |
Average the ratings when evaluators disagree. Discuss any problem where scores differ by 2 or more points, and budget about 5 minutes per disputed item. Those debates often reveal that evaluators pictured different users.
A heuristic evaluation checklist turns each heuristic into a question an evaluator can answer on every screen. The template then records each problem with its location, the heuristic broken, a severity and a suggested fix. Use one shared sheet so findings merge cleanly.
A starter heuristic evaluation checklist, one question per heuristic:
| ID | Screen | Heuristic | Problem | Severity | Suggested fix |
| H-01 | Checkout, payment | 9 | Error reads “Error 402” with no next step | 3 | Plain-language message with a retry option |
| H-02 | Sign-up | 5 | Password rules shown only after a failed attempt | 2 | Show rules beside the field from the start |
Accessibility needs its own pass. The heuristics do not test contrast ratios or keyboard access, so pair the review with a check against W3C’s WCAG 2.2.
A heuristic evaluation uses expert reviewers and a set of principles, while usability testing watches real users attempt tasks. The expert review is faster and cheaper, and it can run on early prototypes. Testing reveals problems experts miss, especially around user goals and language. Teams get the best results using both.
| Factor | Heuristic evaluation | Usability testing |
| Who takes part | 3 to 5 evaluators | 5 or more real users |
| Typical time | About 1 week | 2 to 4 weeks |
| Finds | Breaks from known principles | Real confusion and unmet goals |
| Best stage | Prototypes and live products | Prototypes and live products |
| Misses | Problems specific to your users | Rare edge cases and consistency issues |
Run the expert review first and fix the obvious issues within 1 or 2 sprints. Then test with users, so their 30 to 60 minute sessions surface the deeper problems.
For help running both, our UI UX services team pairs expert reviews with moderated tests. The UX fundamentals guide on our blog covers the principles behind both methods.
Ready to turn heuristic findings into better screens?
Our design team fixes the severity 3 and 4 problems first, then tests the new screens with real users before release.
Use 3 to 5 evaluators. One reviewer misses many problems, and returns fall off quickly past five. Ask each reviewer to work alone for 1 to 2 hours before the group merges findings, so no one anchors on another person’s first impressions.
Yes. The Nielsen heuristics describe human behavior more than any specific technology, so they apply to mobile apps, voice interfaces and AI chat tools. NN/g last reviewed the list on January 30, 2024. Teams sometimes add domain checks, such as AI output transparency.
No. It finds breaks from known design principles, but it cannot reveal what your specific users want or the words they use. Pair it with at least 5 usability test sessions before any major release.
Usability heuristics are broad principles, such as “prevent errors”. Design guidelines are specific rules, such as a 44-point minimum touch target on iOS. An expert review checks the broad principles, while guidelines are checked line by line.
About one week for a medium-size product. Each evaluator spends 1 to 2 hours on the independent review, the team spends half a day merging and rating, and writing the report takes another 1 to 2 days.
