A UX audit examines how an existing experience helps or obstructs the people using it. Its value is not the length of the report. A useful audit tells a team which problems to address, why they matter, and how to check that the change helped. Start with one customer journey and follow it beyond the attractive first screen.
Follow a task from arrival to confirmation
For a fictional language school, the task might be finding an evening class and requesting a place. Begin on the page a visitor actually reaches from search or a campaign. Find the schedule, understand the level, check the location, complete the form, and read the confirmation.
Repeat the journey on a phone and with a keyboard. Include the less comfortable routes: a full class, an unavailable date, an invalid email, and a slow response. Note whether the person can recover without losing their work. This gives the audit a consistent boundary instead of a random collection of visual opinions.
Write findings another person can reproduce
Each finding needs a location, a situation, the observed behaviour, and its consequence. “The form is confusing” gives a developer little to work with. “After submitting without a phone number, the error appears above the visible screen and focus stays on Submit” describes something the team can investigate.
Attach a screenshot without personal information and write the steps needed to reach the problem. Distinguish observed failures from suspicions. If you believe the course description is too technical but have not tested it with prospective students, label it as a question rather than a proven cause of lost enquiries.
Rank consequences before cosmetic preferences
Group issues by what they prevent. A broken submission blocks the task. An unclear fee may prevent a confident decision. Uneven spacing may reduce polish without blocking either. Consider how often the situation occurs, how serious the consequence is, and the effort required to fix it.
Do not turn a rough priority score into scientific certainty. A rarely used route can still matter if failure causes a serious problem. Give each high-priority issue an owner and a suggested acceptance check. Keep visual refinements in the list, but make the trade-off visible.
- 1Checkout button hidden on mobile
- 2Phone number is not tappable
- 3Error doesn’t say which field is wrong
- 4Inconsistent button colours
- 5Typo on the About page
Score = impact × how many people meet it. The fix list reorders itself.
Retest the journey after the changes
An audit is incomplete if the report is delivered and never revisited. Repeat the original steps on the updated website. Confirm that an error message is now visible, the keyboard can reach the control, and the success screen explains what happens next.
If you measure enquiries, compare equivalent traffic and time periods and note other changes such as campaigns or pricing. A before-and-after improvement is not automatically proof that a single interface change caused it. Use the result to choose the next focused review.
Before you build
- Audit one complete customer task.
- Document steps, evidence, and consequences.
- Separate confirmed problems from hypotheses.
- Assign owners and repeat the same checks after fixes.