Usability heuristics

Jakob Nielsen's ten heuristics are the standard checklist for finding usability problems without running a user test. They are rules of thumb, not laws: each one names a way that interfaces commonly fail people.

Where this sits

• jobs-quote-ux comes first. It names the person and what they are trying to do. A heuristic pass with no person in mind produces a list of nitpicks. If nobody has named the person yet, do that before anything below.
• closed-loop-visual-feedback supplies the evidence. Inspect the rendered interface, at the size and on the device the person will use. Reading the source is not an inspection.
• inverse-set-kondo is the eighth heuristic turned on everything else. Aesthetic and minimalist design, asked of rules, checks, config, and code instead of the screen.

The ten

Walk them in order, once per screen or per step of a flow.

1. Visibility of system status. The person can always tell what is happening. A dashboard says how fresh its data is. A filter that is on looks on. A load that takes more than a second shows that it is working.
2. Match between the system and the real world. It speaks the person's language, in the order they think. "Late trips" rather than dq_flag_3. Dates the way the audience writes them.
3. User control and freedom. There is an obvious way back out of anything. Clear all filters. Undo. Close the drill-down and land where you were.
4. Consistency and standards. The same thing looks and behaves the same everywhere, and follows conventions the person already knows. One colour means one thing across every chart. Red is not "good" on one tab and "bad" on the next.
5. Error prevention. The design makes the mistake hard to make, which beats a good error message. A date picker that cannot select an end before the start.
6. Recognition rather than recall. Nothing has to be remembered from another screen. Chart titles carry the active filters. Legends sit next to what they explain.
7. Flexibility and efficiency of use. A newcomer can get through it, and a regular is not slowed down by the newcomer's path. Sensible defaults, shareable URLs, the last view remembered.
8. Aesthetic and minimalist design. Every element on screen competes with the ones that matter. Cut the KPI nobody acts on, the gridlines, the third decimal place.
9. Help users recognize, diagnose and recover from errors. When it goes wrong, say what happened in plain words and what to do next. "No trips match these filters. Clear the borough filter?" rather than an empty chart.
10. Help and documentation. Best if none is needed. Where it is, it sits at the point of need: a one-line definition on the metric, not a manual somewhere else.

Rating what you find

Give each finding a severity, so the list can be acted on in order:

• 4, blocker: the person cannot finish the task, or leaves believing something false.
• 3, major: they finish, but slowly or only after a wrong turn. Fix before sharing.
• 2, minor: a stumble they recover from alone. Fix when convenient.
• 1, cosmetic: fix only if time allows.

A wrong number that looks right is always a 4. On a dashboard, being trusted is the task.

Output

For each finding: the heuristic by number and name, where on the screen it is, what the person experiences, the severity, and the smallest fix. Order by severity, highest first.

Do not force a finding for every heuristic. Say which ones the screen passes and move on. If the screen passes all ten, say so plainly and stop.

Created by Exagrow AI Consulting. Provided under the MIT License.