UX glossary
Heuristic Evaluation
In a heuristic evaluation, experts review an interface against established rules of thumb for good usability. It helps you find typical weak spots quickly and cheaply, even before real people test the product.
Imagine a city council that wants to launch a new online form for residential parking permits in two weeks. There’s no time or budget for a thorough test with residents. Still, the form shouldn’t contain any obvious stumbling blocks. This is where a heuristic evaluation comes in.
Explained simply
A heuristic is a rule of thumb, a tried-and-tested guideline that holds true most of the time. In a heuristic evaluation, experts walk through an interface step by step and check where it breaks those rules. Real users aren’t involved.
Think of a car inspection: before a car goes in for its official roadworthiness test, a garage checks it against a fixed checklist. Brakes, lights, tires. That catches the typical defects, even though nobody recreates every driving situation. The method goes back to Jakob Nielsen and Rolf Molich, who introduced it in 1990.
Nielsen’s ten heuristics
The most widely used set is the ten usability heuristics that Nielsen refined in 1994. They include “Visibility of system status” (show what’s going on), “Match between the system and the real world” (speak your users’ language), “Error prevention” and “Recognition rather than recall” (make options visible instead of relying on memory). You’ll find an overview in the resource 10 Usability Heuristics.
The process matters: several experts first review the interface independently and only then combine their findings. Across six of his projects, Nielsen found that single evaluators caught only about 35 percent of the problems on average. That’s why he recommends three to five evaluators, since each one spots different issues.
An example
Back to the parking permit. Before: after submitting, the form shows nothing but a blank screen, error messages appear bundled at the top of the page, and the field labeled “Registered keeper details” uses bureaucratic jargon. Three evaluators independently find different issues and then agree on the most important ones. After: a confirmation clearly shows that the application was received, errors appear right next to the affected field, and the label now reads, in plain words, “Who is the car registered to?”
Common misconceptions
A heuristic evaluation is no substitute for usability testing. It shows where an interface breaks proven rules, but not how real people cope with it in their own context. And not every violation is automatically a problem: sometimes it’s a deliberate trade-off. To learn how to use expert judgment well, read How to Use Expert Evaluation Correctly.
In practice
- Review alone first, then together. If you see the others’ findings early, you’ll miss your own. Only compare notes once everyone is done.
- Limit the scope. Focus on key flows such as sign-up or checkout instead of reviewing the entire product at once.
- Rate the severity. For each finding, note how badly it gets in people’s way. That tells the team what to fix first.
Related terms
Usability Testing, System Usability Scale, Mental Model, Cognitive Load