What an expert review can and can't tell you

A heuristic evaluation is the cheapest research there is, which is both its strength and its risk. No participants, no recording, no recruitment. One person who knows the principles walks your journey and checks it against a short list of rules of thumb: make the state visible, match the system to the real world, give the user control, keep things consistent, prevent errors before they happen. Each mismatch is a finding, and an afternoon of this produces a prioritised list of problems faster than any other method.

Nielsen’s ten heuristics are the standard checklist, and they’re worth knowing by heart, because they name things that are invisible until you have the word for them. “Match between system and the real world” is why a button labelled “Submit” fails where “Get my quote” works. “Recognition rather than recall” is why a checkout that makes you remember your order number is worse than one that shows it to you. “Error prevention” is why a form that blocks an impossible postcode as you type beats one that tells you after you hit submit. The heuristics show you where to look, but they won’t tell you what to build.

Treat a review as a map of where to look next. It finds the problems worth investigating, but it can’t prove those problems matter to your customers, and it won’t tell you how to fix them. Once a review flags a problem, the next step is to watch a real person hit it, or to test a fix, before anyone redesigns on the strength of one person’s checklist. A review that ends in “we fixed everything on the list” has skipped the part where you check whether the list was right.

This is where the method over-reaches. Because it’s fast and the person doing it is confident, the findings feel like evidence, when they’re one expert’s opinion, structured but still opinion. Expertise also has a known bias. The more you know about how interfaces work, the less you resemble the person who’s confused by yours. The reviewer who can name all ten heuristics is exactly the person who won’t struggle with your checkout the way a tired parent on a phone at 11pm will. A review can’t see that person, so you have to go and watch them.

So use it for triage, to turn a mountain of possible improvements into the five worth testing first. It costs almost nothing and catches the obvious stuff, dead links, invisible buttons, jargon nobody outside the office understands. The trouble starts when it becomes the whole programme, a confident report standing in for observation. You can usually spot it: lots of recommendations, no participants, and nobody ever checked whether any of it was true.

Put it to work

Take your highest-value page and check it against three usability principles you can name. List every mismatch, then mark each as "fix now" or "test first".

Your answer

Show the worked example

There is no right answer. Write yours first, then compare.

Sources

1. Jakob Nielsen, 10 Usability Heuristics for User Interface Design
2. NN/g, Heuristic Evaluation