Mara logo

Heuristic Evaluation

An expert review of an interface against established usability principles, done without users, to find likely problems quickly and cheaply.

Process & Methods

Reviewers walk the product against a checklist — visibility of system status, match with the real world, user control, consistency, error prevention, recognition over recall, and so on — and log violations with a severity rating. Two or three independent reviewers find substantially more than one, because each notices different classes of problem.

It is fast and cheap, and it is not a substitute for usability testing. Heuristics reliably catch craft problems: unclear labels, missing states, inconsistent patterns. They cannot tell you that users do not understand the mental model, because the reviewer already does. Use it as the first pass of a UX audit, then test the findings you are least sure about.

In practice

A banking app concept was reviewed against ten heuristics before any user session. Before: 34 issues logged, 6 rated severe. After fixing the severe ones — no confirmation on transfers, no error recovery on failed payments — the later user sessions surfaced mental-model problems instead of the same six complaints. The catch — reviewers disagreed on severity, so ratings had to be reconciled.

Learn more

You may ask

Frequently Asked Questions

How many reviewers do we need?

Three is the usual sweet spot. A single reviewer typically finds around a third of the issues a small group would; beyond three or four the additional findings tail off sharply.

Can heuristic evaluation replace user testing?

No. It finds violations of known principles, which is a different set from the problems users actually hit. It is best used to clear the obvious issues so that user sessions are not spent rediscovering them.

Related terms

All terms

Let's talk about your product.

Happy to look at what you're building and say where design would move the needle.

Contact Us