Heuristic evaluation (expert review)

Heuristic evaluation is a method of finding problems in an interface, where several experts independently assess a site or app against a set of principles (heuristics), most often the 10 Nielsen heuristics. It is also called an expert usability review.

In plain words

Instead of inviting customers to a test, UX specialists walk through the site themselves and mark the places that break well-known principles of ease of use: an unclear button, no confirmation when an item is added to the cart, a form error described in a confusing way. Nielsen Norman Group defines the method as a way to find design problems in an interface, where evaluators compare the design against heuristics (NN/g).

How many experts you need

NN/g recommends 3-5 evaluators working independently: each one alone will almost certainly miss some problems (NN/g). One expert is fast, but incomplete.

How it is done

According to NN/g, the process has three stages:

1. Preparation. Choose and brief the team, agree on how to record findings (a spreadsheet, a board), and limit the scope: specific tasks, sections, a user group, a device. For a store, for example, "finding a product and checking out on a smartphone".
2. Independent evaluation. Each person first gets to know the product, then goes through it a second time and records heuristic violations. The guideline is 1-2 hours.
3. Consolidation. Group similar problems, discuss disagreements, set priorities.

The NN/g severity scale (0-4)

To know what to fix first, each problem is rated (NN/g):

• 0: I don't agree this is a usability problem;
• 1: a cosmetic problem: fix if time allows;
• 2: a minor problem: low priority;
• 3: a major problem: high priority;
• 4: a usability catastrophe: must be fixed before launch.

Severity combines frequency (how often it happens), impact (how hard it is to overcome), persistence (whether it bothers once or every time) and market impact. Nielsen advises collecting ratings after the sessions, using a questionnaire with the full list of problems, each person rating independently, which takes about 30 minutes. The ratings of a single expert should not be trusted; the average of about three evaluators is reliable enough for practice.

Example (made-up): the store has no Nova Poshta branch selection on mobile because the list button does not open. This blocks the order for every mobile customer, so the rating is 4. Uneven spacing between product cards is a 1.

How it differs from usability testing

NN/g says directly that heuristic evaluation does not replace research with users. Problems are fully revealed only by watching people, while an expert review helps decide what to focus the next test on (NN/g). For tests, Nielsen recommends rounds of 5 users (NN/g).

When to use it

• Early in a project: on prototypes and mockups, when it is too early to test with people.
• On a limited budget: to find obvious problems without recruiting participants.
• Before a usability test: to remove the obvious and spend the test on harder questions.
• Before a redesign or after big changes: as a quick audit of the current state.

Common mistakes

• One evaluator instead of 3-5.
• Evaluators discuss findings before each has finished their own evaluation.
• No severity scale: a list of a hundred items with no priorities.
• Treating every heuristic violation as a mandatory fix.
• Seeing the review as a replacement for testing with customers.

FAQ

What is a heuristic evaluation of a site? An expert review of a site's ease of use against a set of principles (heuristics), carried out independently by several UX specialists.

How many experts does a heuristic evaluation need? NN/g recommends 3-5 independent evaluators.

How is heuristic evaluation different from usability testing? In an evaluation, experts assess the site against principles; in testing, you watch real users. The first is cheaper and faster, the second shows real behavior.

Related terms

• 10 Usability Heuristics of Jacob Nielsen for E-commerce: How to Increase Conversion
• Usability
• UX and UI: what they are and how they differ

Sources

• Moran K., Gordon K. How to Conduct a Heuristic Evaluation. Nielsen Norman Group — nngroup.com/articles/how-to-conduct-a-heuristic-evaluation
• Nielsen J. Severity Ratings for Usability Problems. NN/g — nngroup.com/articles/how-to-rate-the-severity-of-usability-problems
• Nielsen J. Why You Only Need to Test with 5 Users. NN/g — nngroup.com/articles/why-you-only-need-to-test-with-5-users