Quick answer: What is a heuristic evaluation?
- What: a review of an interface against usability guidelines.
- Usual guidelines: Nielsen's 10 usability heuristics (1994).
- Who: three to five evaluators, working independently.
- How long: about 1–2 hours each, per Nielsen Norman Group.
- Output: merged findings, each rated 0–4 for severity.
Key takeaways
- A heuristic evaluation finds likely problems cheaply; it doesn't replace testing with users.
- One reviewer misses issues; independent reviewers catch more.
- Narrow the scope: one task, one user group, one device.
- A violation isn't automatically a bug; context decides.
- Severity ratings turn a long list into a fix order.
Founders rarely have budget for a usability lab, and they rarely need one to find their first ten problems. A heuristic evaluation, invented by Jakob Nielsen and Rolf Molich in 1990, lets a small team find likely usability issues in an afternoon by checking the product against a short list of proven principles. This guide covers Nielsen's 10 heuristics with startup examples, how to run an evaluation the way Nielsen Norman Group recommends, and how to rate what you find, as read on 17 November 2026. Jump to the 10 heuristics, how to run one or severity ratings.
- Usability heuristics (Nielsen, 1994)
- 10
- Independent evaluators recommended
- 3–5
- Per evaluator, timeboxed
- 1–2 h
- Nielsen's severity scale
- 0–4
What is a heuristic evaluation?
A heuristic evaluation is a method for finding design problems in an interface: evaluators judge it against a set of guidelines (heuristics) that make systems easy to use. It works on almost anything people interact with, from prototypes to live apps, and it's especially useful early, when fixes are cheap.

Nielsen's 10 heuristics, with startup examples
| # | Heuristic | A startup example of a violation |
|---|---|---|
| 1 | Visibility of system status | "Generate report" clicked; nothing moves for 20 seconds |
| 2 | Match between system and real world | Onboarding asks for a "workspace slug" |
| 3 | User control and freedom | No way to undo deleting a project |
| 4 | Consistency and standards | "Save" is blue on one screen, grey on the next |
| 5 | Error prevention | A date field that accepts 31 February |
| 6 | Recognition rather than recall | Users must remember an API key shown once, two screens ago |
| 7 | Flexibility and efficiency of use | No keyboard shortcuts in a tool used all day |
| 8 | Aesthetic and minimalist design | Pricing page with five banners competing with the plan cards |
| 9 | Recognise, diagnose, recover from errors | "Something went wrong" with no next step |
| 10 | Help and documentation | No help link on the integration screen where everyone gets stuck |
Table: Nielsen's 10 usability heuristics with startup examples (heuristics from Nielsen Norman Group; examples are ours)
Each heuristic has a behavioural reason behind it. Recognition rather than recall, for example, follows from limited working memory; see recognition over recall and the rest of the Laws of UX.
How to run a heuristic evaluation

- Pick the heuristics — Nielsen's 10 for usability; add domain ones if you need them.
- Train the team — everyone reads the heuristics; do a practice round on a simple app.
- Narrow the scope — one task, one section, one user group or one device.
- Evaluate independently — 1–2 hours each; first pass to learn, second to find issues.
- Note each issue against a heuristic — with a screenshot and a possible fix.
- Merge findings — affinity-map duplicates, discuss disagreements.
- Rate severity — 0 to 4, then fix from the top.
Nielsen Norman Group stresses that evaluators shouldn't see each other's notes until they finish: the point of several reviewers is independent observations.
Rating severity
Nielsen's 1994 severity scale combines how often a problem happens, how hard it is to overcome and whether it keeps bothering users:
| Rating | Meaning |
|---|---|
| 0 | Not a usability problem |
| 1 | Cosmetic: fix if there's time |
| 2 | Minor: low priority |
| 3 | Major: high priority |
| 4 | Catastrophe: fix before release |
Table: Severity ratings for usability problems (Jakob Nielsen, Nielsen Norman Group, 1994)
Mistakes

- One evaluator — everyone misses things; use three to five
- Treating every violation as a bug — NN/g's example: hamburger menus violate #6 but often make sense on mobile
- Scope too wide — a whole product in an hour finds only the obvious
- No severity — a flat list rarely gets fixed
- Skipping real users — heuristics find likely issues; testing confirms them

Ready for a second pair of eyes?
How we know
The method, evaluator numbers and timings come from Nielsen Norman Group's "How to Conduct a Heuristic Evaluation" (2023); the heuristics from Jakob Nielsen's article on NN/g; the severity scale from Nielsen's 1994 article; all read on 17 November 2026. The startup examples are our own illustrations.
Key terms, defined
- Heuristic: a broad rule of thumb for usable design.
- Heuristic evaluation: judging an interface against heuristics.
- Evaluator: a person reviewing the interface.
- Severity rating: a 0–4 score for how serious a problem is.
- Affinity diagram: grouping similar findings to merge them.
Frequently asked questions
What is a heuristic evaluation?
A method for finding usability problems by judging an interface against a set of guidelines called heuristics, usually Jakob Nielsen's 10 usability heuristics. Evaluators review the product themselves rather than watching users.
What are Nielsen's 10 usability heuristics?
Visibility of system status; match between the system and the real world; user control and freedom; consistency and standards; error prevention; recognition rather than recall; flexibility and efficiency of use; aesthetic and minimalist design; help users recognize, diagnose and recover from errors; and help and documentation.
How many evaluators do you need?
Nielsen Norman Group recommends three to five people evaluating the same interface independently, because any single evaluator, however experienced, misses some problems.
How long does a heuristic evaluation take?
Nielsen Norman Group suggests timeboxing each evaluator's session to about one to two hours, then a team session to merge and prioritise findings.
Is a heuristic evaluation the same as a UX audit?
It's usually one part of a UX audit. A full audit also adds accessibility checks against WCAG, speed measurements, analytics and the whole user journey.
Can a heuristic evaluation replace user testing?
No. Nielsen Norman Group says it can't replace user research, because UX is contextual. It's a cheap way to find glaring problems and decide what to test.
Sources
- Nielsen Norman Group — How to Conduct a Heuristic Evaluation (Kate Moran, Kelley Gordon, 2023)
- Nielsen Norman Group — 10 Usability Heuristics for User Interface Design (Jakob Nielsen)
- Nielsen Norman Group — Severity Ratings for Usability Problems (Jakob Nielsen, 1994)
- W3C — Web Content Accessibility Guidelines (WCAG) 2.2
- Science of Apps — Product investigation
- Checked November 17, 2026. Figures that change — fees, salaries, prices — are reviewed every six months; if something is out of date, tell us.
Where to go from here
- What is a UX audit? — the bigger picture
- How to do a UX audit — a one-day process
- Laws of UX — why users behave as they do
- Science of Apps - LaunchpadThe UX audit checklist: a free, journey-based template for websites and apps, built on Nielsen's heuristics and WCAG 2.2A free UX audit checklist for websites and apps: 60 checks across the user journey, from first impression to payment, using Nielsen's heuristics and WCAG 2.2.
- Science of Apps - LaunchpadHow to do a UX audit of your website in one day: the eight steps, the standards to check against, the free tools, and how to turn findings into a fix listHow to do a UX audit in one day: set goals, walk the journey, check Nielsen's heuristics and WCAG 2.2, measure speed, read analytics and rank the fixes.
- Science of Apps - LaunchpadThe best UX audit tools for startups, free and paid: what each one actually checks (behaviour, speed, accessibility, research) and the gaps no tool fillsThe best UX audit tools for startups, free and paid: Clarity, PageSpeed Insights, WAVE, Hotjar, Maze and an end-to-end product test, with what each checks.
- Science of Apps - LaunchpadHow to get your first 100 users without paid ads: a channel-by-channel plan founders can start todayHow to get your first 100 users without paid ads in 2027: your network, communities, launch sites, directories, content and one-to-one outreach, step by step.
- Science of Apps - LaunchpadA go-to-market strategy for startups on one page: the eight boxes of a GTM canvas, a worked example, the B2B and self-serve motions, and a 90-day launch planA go-to-market strategy template for startups: an eight-box GTM canvas covering customer, problem, promise, price, channels and motion, plus a 90-day plan.
- Science of Apps - Launchpadllms.txt for startups: what the file is, what Google says about it for Search and AI Overviews, and when it's still worth addingWhat llms.txt is, why Google says Search and AI Overviews don't use it, and when a startup should still add one, with a simple template.
Get the ideas, not the noise.
Sign up with Google and we’ll send you short UX psychology tips by email — why people tap, stall and leave, and what to do about it. Unsubscribe anytime. Prefer something to keep? Get the free Laws of UX guide as a PDF.
Launch day ends. Your listing shouldn't.
List free on the Launchpad and, once live, get a free report on your landing page, discoverability and speed; paid reports review every screen against the Laws of UX.





