UX Audit
A UX audit is a bounded, evidence-led review of an existing product that identifies experience risks, explains their likely impact, and prioritizes further action.
Download Markdown (.md)Soroush’s Product Wiki · https://wiki.soroush.uk/topic/ux-audit/
01What it means
A UX audit combines expert inspection with available product evidence. Depending on scope, it may examine key journeys, heuristics, accessibility, content, analytics, support themes, consistency, and design-system use. Its output is a reasoned inventory of findings and hypotheses, not a universal usability score or a substitute for observing users.[1][2][3]
02Why it matters
Mature products often contain problems spread across teams and releases. An audit creates a common view of recurring friction, distinguishes isolated defects from systemic causes, and helps leaders choose where deeper research or remediation will have the greatest value. It can also establish a baseline before a redesign or acquisition integration.
03When to use it
Use an audit when a product has accumulated inconsistency, performance signals lack explanation, a redesign needs focus, or direct research must be prepared efficiently. Define the audience, journeys, platforms, and decision horizon first. Avoid promising complete coverage when access, evidence, or time is constrained.
04How it works
- Set the audit question, scope, representative users, critical journeys, environments, and evidence available.
- Choose explicit criteria such as usability heuristics, accessibility requirements, content principles, and system conventions.
- Walk through realistic scenarios and triangulate observations with analytics, research, complaints, and operational data where available.
- Document each finding with location, evidence, affected users, likely consequence, confidence, severity, and systemic pattern.
- Prioritize validation and remediation, then separate quick corrections from questions requiring research or structural investment.
05A practical example
A university audits its course-registration service after rising support volume. Expert review finds inconsistent prerequisite messages and poor recovery from timetable conflicts; analytics show repeated backtracking at the same steps, while support logs reveal confusion about saved choices. The team fixes terminology immediately and schedules moderated testing for the unresolved planning model.
06Where it breaks
An audit reflects its criteria, evidence, evaluator expertise, and sampled scenarios. Experts may miss domain knowledge, assistive-technology behavior, or motivations visible only in research. Analytics show events rather than intent. Severity estimates are hypotheses until validated, and a long issue list can distract from the few systemic changes that matter.
Common mistakes
- Auditing an undefined product surface and implying that sampled screens represent the whole experience.
- Presenting evaluator opinion without criteria, evidence, confidence, or affected scenarios.
- Turning every observation into a redesign recommendation before validating causes and constraints.
07Sources
- How to Conduct a Heuristic EvaluationNielsen Norman Group · Professional guidanceHow to Conduct a Heuristic Evaluation
- 10 Usability Heuristics for User Interface DesignNielsen Norman Group · Professional reference10 Usability Heuristics for User Interface Design
- Conformance Evaluation and ReportsW3C Web Accessibility Initiative · Standards guidanceConformance Evaluation and Reports