A UX audit is a structured evaluation of a product's experience, measured against real user behaviour, usability heuristics, and business goals, to find what's hurting users and prioritise what to fix. It's diagnosis, not redesign.
Most UX audits fail at one thing: being acted on. The output is a 60-slide deck that gets admired once and never reopened. A UX audit that earns its keep does the opposite. It finds the handful of issues actually costing you signups, transactions, or retention, ranks them, and tells you what to fix first.
This is the guide we wish we'd had when we ran our first audit back in 2011. It covers the full process (our 7-lens framework), a 50-point checklist you can use today, what's genuinely changed in the AI era, an honest take on cost, how to choose a partner, and the part almost every guide skips: what to do after the report. It's written for the product managers, founders, and design leads deciding whether to run an audit, and how.
What a rigorous UX audit includes. Use this to run one, or to judge a provider:
- Evidence over opinion. Behavioural and attitudinal data, not a list of personal preferences.
- Business-goal alignment. Findings tied to activation, conversion, or retention.
- A named, repeatable framework. Not an ad-hoc walkthrough.
- Severity scoring and prioritisation. What to fix Monday, and what to ignore.
- Accessibility built in. WCAG is part of the audit, not an upsell.
- A roadmap with a next step per finding. And a way to measure it.
If you'd rather hand the whole thing off, that's our UX Audit Report service.
What is a UX audit?
A UX audit systematically inspects a product, live or in progress, to surface usability, accessibility, and experience problems, then turns them into a prioritised, evidence-backed list of fixes. The good ones blend quantitative signals (analytics, funnels, heatmaps, session recordings) with qualitative judgement (heuristic evaluation, expert review, user research). Numbers tell you where it hurts. Judgement tells you why, and what it's worth.
People use three terms interchangeably that aren't the same thing:
| What it is | When to use it | What you get | |
|---|---|---|---|
| UX audit | Holistic expert review of the whole experience against heuristics, data, and goals | You sense UX is costing conversions but can't pinpoint where | Prioritised findings and a roadmap |
| Usability testing | Watching real users attempt real tasks | You need evidence on a specific flow or hypothesis | Observed task success or failure |
| Heuristic evaluation | Scoring an interface against fixed principles (e.g. Nielsen's 10) | A fast, low-cost expert pass, often one lens inside an audit | A list of heuristic violations |
A UX audit usually contains a heuristic evaluation and may commission usability testing. It's broader than either. (Some auditors also use a cognitive walkthrough, stepping through a task as a first-time user, as an alternative lens to heuristic evaluation.) Jakob Nielsen's ten usability heuristics remain the field's common language, which is why we treat heuristic evaluation as one lens of a fuller audit rather than the whole job.
And what a UX audit isn't: a visual-design opinion, a brand refresh, or a redesign. It tells you what's broken and what fixing it is worth. Skip straight to the redesign without the audit and you'll spend three months making the wrong screen prettier.
Why run a UX audit?
Friction is expensive, and most of it is invisible to the team that built the product. You stop noticing the confusing step you designed eighteen months ago. Your users never do. A UX audit puts a number and a priority on that friction.
The cost of not looking is concrete. The 2026 WebAIM Million found detectable accessibility failures on 95.9% of the top one million home pages, an average of 56 per page — both numbers rising, not falling, from the year before. Most products are leaking users in ways nobody has measured.
You probably need an audit if you recognise a few of these:
- Analytics show people arriving and leaving at the same step, repeatedly.
- Activation or trial-to-paid has flattened and you've run out of obvious fixes.
- Support keeps answering the same "how do I…" question.
- You're about to invest in a redesign and want to know what to keep.
- You've shipped fast for a year and never stopped to look at the whole.
- Your conversion rate is below benchmark and you can't say precisely why.
The point isn't to collect problems. It's to find the two or three that, once fixed, move a metric you care about, and to ignore the rest with confidence.
When (and how often) to run one
The best time to audit is before you spend money acting blind. The moments with the most upside:
- Before a redesign, so you keep what works and only rebuild what's broken.
- After a launch. Real traffic exposes problems no internal review predicted. Give it a few weeks of data.
- At a growth plateau. When acquisition is fine but conversion or retention has stalled, the leak is usually in the experience.
- After a year of fast shipping. Velocity quietly fragments an experience, and an audit re-aligns it.
- Before a funding round or major sales push. Find the holes before they're scrutinised.
As a cadence, most teams shipping at pace benefit from a focused audit every three to six months. Not a giant annual review nobody acts on. A tighter, recurring look at the flows that matter most. The trigger is change: every significant redesign, new core feature, or shift in audience is a reason to re-audit the affected flow.
How to run one: the WowMakers 7-Lens Framework
The fastest way to a useful audit is a consistent process. We run every product through the same seven lenses. We call it the 7-Lens Audit, and it's the exact framework behind our UX Audit Report. Each lens answers one question and surfaces one kind of finding.
Lens 1: Business-objective alignment
What is this product actually for? Before touching the interface, we anchor on the one or two metrics that matter: activation, trial-to-paid, checkout completion. Every later finding is scored against these, so "nice-to-have" never sneaks onto the critical list. In practice this is the lens that saves the most wasted effort. Teams routinely arrive with a list of cosmetic gripes, and the moment you ask "which of these moves activation?", half the list falls away and the real problem (usually a single step on the money path) comes into focus. Common finding: design decisions lovingly optimised for a metric nobody is accountable for.
Lens 2: User & buyer personas
Who uses this, and who decides to pay? In B2B those are different people, and an interface tuned for the daily user can still fail the economic buyer, or the other way around. We audit for both. Where journey maps or personas already exist, we layer them over the behavioural data to see where real drop-off aligns with documented friction. More tellingly, we look for where it doesn't, because the gap between the map and the data is usually where the product drifted. Common finding: an experience that delights users but never makes the value legible to the person signing off.
Lens 3: Behavioural metrics
Where does it actually hurt? GA4 funnels, drop-off points, rage clicks, heatmaps, and session recordings. This is where the evidence comes from, and it tells the heuristic review which screens to scrutinise instead of guessing. We look for the single steps quietly shedding a third of users, the rage-click clusters that mean "I tried and it didn't work", and the dead zones people never discover. Common finding: one step in an otherwise healthy funnel doing most of the damage. Invisible in aggregate, obvious in the recordings.
Lens 4: Heuristic product evaluation
Does the interface respect how people work? A structured pass against Nielsen's ten usability heuristics: visibility of system status, match to the real world, user control and freedom, consistency and standards, error prevention, recognition over recall, flexibility and efficiency, aesthetic and minimalist design, help users recognise and recover from errors, and help and documentation. (These principles sit alongside the broader laws of UX that shape good interface decisions.) Common finding: across our own audits, the one we see most often is invisible system status during asynchronous operations. A save, a payment, an upload that gives no sign it's working, so users retry, refresh, or leave.
Severity scoring. A violation isn't worth much as a bare observation. It needs a severity so the roadmap can sort itself. We use Nielsen's 0-to-4 severity scale: 0 (not a problem), 1 (cosmetic), 2 (minor), 3 (major), 4 (catastrophe), judged on frequency, impact, and persistence. A catastrophe on the checkout flow outranks a dozen cosmetic issues on a settings page, and severity is what makes that obvious to a stakeholder skimming the report.
Lens 5: Attitudinal metrics
What do users say, versus what they do? SUS (System Usability Scale) scores, in-product surveys, NPS verbatims, and the patterns hiding in support tickets. Cheap signal that most audits skip entirely. The most useful moment is the contradiction. A high satisfaction score sitting on top of a low completion rate usually means people want it to work, and are forgiving a flow that's quietly failing them. Common finding: a likeable product with a broken job-to-be-done.
Lens 6: Design-system consistency
Does the product feel like one product? Inconsistent components, type scales, and interaction patterns tax users invisibly. Every "wait, this button behaves differently here" is a tiny tax on trust, and they compound. We audit the system itself, not the screens alone: tokens, components, states, and the patterns that should have been reused but were reinvented. Common finding: three buttons that do the same thing and look different, or four that look the same and don't.
Lens 7: Prioritised recommendations
What do we fix, and in what order? Every finding goes on an impact/effort matrix (or gets scored with MoSCoW), tied back to the Lens 1 metrics, so you leave with a sequence: fix this sprint, this quarter, never. Not an undifferentiated list. When the matrix is ambiguous, the Kano model helps. A "delighter" (an excitement attribute) ranks below a threshold item that blocks core task completion, even if the delighter is more fun to build. The output is the whole point of the audit. A roadmap, not a worry list.
The 50-point UX audit checklist
Here's the working checklist we use, grouped into eight categories. It's deliberately ungated, so use it today.
Navigation & information architecture
1. Is the primary navigation visible and consistent across pages?
2. Can a user always tell where they are?
3. Are labels written in the user's language, not internal jargon?
4. Is the most important action obvious on each screen?
5. Are there dead ends with no clear next step?
6. Is search present where the catalogue or content warrants it?
7. Is the back and exit behaviour predictable?
Forms & inputs
8. Is every field actually necessary?
9. Are labels persistent (not placeholder-only)?
10. Is validation inline and specific ("enter a valid email", not "error")?
11. Are errors shown next to the field, in plain language?
12. Is autofill and autocomplete supported where useful?
13. Is the input type correct on mobile (numeric keypad for numbers, and so on)?
14. Is progress shown for multi-step forms?
Visual hierarchy & typography
15. Does the eye land on the most important element first?
16. Is body text comfortably readable (size, line length, contrast)?
17. Is there a consistent type scale?
18. Is whitespace used to group and separate, not to decorate?
19. Are interactive elements visually distinct from static ones?
20. Is the brand expressed without sacrificing clarity?
Accessibility (WCAG 2.2)
21. Does colour contrast meet AA (4.5:1 body text, 3:1 large text and UI components)?
22. Is every interactive element keyboard-reachable, with no keyboard traps?
23. Are focus states visible?
24. Do images have meaningful alt text?
25. Are form fields properly labelled for screen readers?
26. Is information ever conveyed by colour alone?
27. Does the page have a logical heading structure?
28. Do media have captions or transcripts where needed?
Mobile & responsive
29. Are tap targets large enough (WCAG 2.2 sets a 24×24px minimum; ~44px is the comfortable target)?
30. Does content reflow without horizontal scroll?
31. Are gestures discoverable, not hidden?
32. Is critical content above the fold on small screens?
33. Do forms and modals behave on mobile keyboards?
34. Is performance acceptable on a mid-range device?
Error handling & feedback
35. Does every action confirm it worked?
36. Are destructive actions confirmed or undoable?
37. Are loading and empty states designed (not blank)?
38. Do error messages say what happened and what to do?
39. Is the system status always visible during long operations?
40. Are success states clear and reassuring?
Content & trust
41. Is microcopy clear, concise, and human?
42. Are prices, terms, and next steps unambiguous?
43. Is social proof or credibility present where decisions are made?
44. Is help reachable without leaving the task?
45. Is onboarding teaching by doing, not by tour-overload?
Conversion & flow
46. Is the path to the primary goal as short as it can be?
47. Are there unnecessary steps, fields, or decisions on that path?
48. Is friction front-loaded (asking for too much, too early)?
49. Does the flow recover gracefully when a user makes a mistake?
50. Does the experience match the promise that brought the user in?
Working through the checklist? See what a finished UX audit looks like (free, no email required)
Free UX Audit Report
UX audit in the AI era: what's different in 2026
The biggest change since most UX-audit guides were written is that the interface is no longer fixed, and the audit has to account for it. Four things changed.
Auditing AI-generated interfaces. When layouts are assembled or personalised by a model, "the screen" stops being a fixed artefact, and a component that passes review on one render can fail accessibility on another. Auditing now means testing ranges of generated states, not a single snapshot, and checking the guardrails that keep generated UI inside your design system. The question shifts from "is this screen right?" to "is every screen this system can produce acceptable?"
AI dark patterns. Classic dark patterns are static: one manipulative layout shown to everyone. AI-era ones are per-user, a nudge tuned to each person's hesitation, harder to catch in a single review pass because no two users see the same thing. Auditing them takes scenario-based thinking and a deliberate hunt for personalised pressure.
AI-assisted audit tools. Heuristic-evaluation assistants can pre-scan an interface for common issues quickly and at breadth: contrast failures, missing labels, obvious inconsistencies, faster than a human. They're genuinely useful for the first pass. They're not a substitute for judgement on the findings that matter.
Knowing when AI is enough. AI is good at surfacing candidate issues across many screens. It's poor at severity and trade-offs, which problems cost conversions, and what to fix first. Use AI for breadth and a senior human for judgement. Any audit that pretends a model can rank business impact on its own is selling you a longer list, not a better decision.
Accessibility is now a compliance layer, not a nice-to-have
Accessibility moved from "good practice" to "legal exposure" in the last year. The European Accessibility Act began enforcement on 28 June 2025. The Act covers a defined list of products and services rather than software generally. Its service requirements apply to listed services provided to consumers, including e-commerce, consumer banking, electronic communications, passenger transport, e-books, and services providing access to audiovisual media. Micro-enterprises providing covered services are exempt. Non-EU businesses may still be covered when they place covered products on the EU market or provide covered services to consumers in the EU. Enforcement isn't theoretical anymore, either: in June 2026 a French court ordered Carrefour to make its e-commerce site and app fully accessible within six months, and regulators in the Netherlands and Sweden have moved from warning letters to active audits. Then in October 2025, WCAG 2.2 was approved as an international standard (ISO/IEC 40500:2025), turning the guidelines from a W3C recommendation into a formal benchmark regulators and procurement teams can cite directly. Meanwhile the 2026 WebAIM Million found 95.9% of top home pages still carry detectable WCAG failures, up from 94.8% the year before — accessibility debt is compounding, not shrinking. That gap is a UX problem and a legal-risk problem at once.
A compliance-aware audit checks the things that get sites into trouble:
- Colour contrast. Body text at 4.5:1, large text and UI components at 3:1.
- Visible focus states. Every interactive element shows where keyboard focus is.
- Full keyboard operability. Nothing requires a mouse, and there are no keyboard traps.
- Screen-reader labelling. Form fields, buttons, and icons have accessible names.
- Logical structure. A sensible heading order and landmark regions.
- No meaning by colour alone. Status and meaning carry a second signal.
If the product or service falls within that defined scope, accessibility belongs in the compliance review as well as the UX review. An audit can surface gaps before they lead to complaints or enforcement.
UX audit tools (by job)
You don't need all of these. You need one or two per job. Pick by the question you're answering, not the longest feature list.
- Behavioural analytics, for where do people drop off? GA4, PostHog, Mixpanel, Amplitude for funnels, retention, and event analysis.
- Session replay & heatmaps, for what does the struggle look like? Hotjar, FullStory, LogRocket, UXCam for real sessions, rage clicks, and dead clicks.
- Usability testing, for can users complete the task? Maze, Lookback, UserTesting for moderated and unmoderated studies. (See our take on usability testing tools.)
- Design & inspection, for is the system consistent? Figma to audit components, tokens, and patterns at source.
- Accessibility, for does it pass WCAG? axe, WAVE, and Lighthouse for an automated first pass, then manual keyboard and screen-reader testing for the rest.
- AI-assisted, for what are the obvious issues, fast? AI heuristic-evaluation tools for breadth, always human-reviewed before they reach the report.
The tool is never the audit. A heatmap shows you that people hesitate. Only a person can tell you why, and whether it's worth fixing.
What goes into a UX audit report?
A report nobody acts on is a failed audit. A useful one has four parts: an executive summary a stakeholder can read in two minutes; findings grouped by severity (critical, major, minor), each with evidence attached; prioritised recommendations sequenced by impact and effort; and a clear next step for each, so "fix this" never becomes "fix this, somehow, sometime."
In practice the deliverable that gets acted on is concrete. Three artefacts do most of the work:
- Annotated screenshots. The problem circled on the actual screen, with the heuristic it violates and a one-line fix.
- Annotated session clips. Ten seconds of a real user hesitating is worth a page of prose to a sceptical stakeholder.
- An impact/effort matrix. Every finding plotted, so the roadmap is a glance, not a debate.
Curious what a finished one looks like? See an example of our UX Audit Report deliverable →
How much does a UX audit cost?
Honestly, "it depends", and anyone quoting a flat number before seeing your product is guessing. Cost scales with scope: the number of platforms (web plus iOS plus Android triples the surface), whether new research is required or existing data is enough, the depth of deliverable (a findings list versus a prioritised, costed roadmap), and turnaround. A focused audit of a single critical flow is a very different exercise from a full multi-platform review.
Rather than anchor you on a misleading figure, we'd rather scope it to what you actually need. Tell us the product and the question you're trying to answer, and we'll come back with a clear, scoped proposal. Start here.
See what you'd actually get: a finished UX audit report (free, no email required)
Free UX Audit Report
Should you run your own audit, or hire a studio?
Every guide tells you to run an audit. Almost none tell you when not to run your own, which is the question that actually decides whether you need help.
When DIY works: an early-stage product, one clear flow, an in-house designer with fresh eyes and the time to use the checklist above. You'll catch real, obvious wins, and build a habit of looking.
When DIY misleads you: when you're auditing decisions you made. Founder bias and anchoring are real. Teams consistently rate their own flows as clearer than users find them, and "we already know what's wrong" is usually right about the cheap half and blind to the expensive half. No fresh perspective produces false confidence, which feels like diligence and isn't.
What a professional audit adds: independent eyes, a repeatable framework, behavioural and attitudinal data triangulated rather than cherry-picked, and a prioritised, costed roadmap instead of a list. If you need that independent eye, our UI/UX design team has run audits across fintech, SaaS, and consumer products since 2011. And that uncertainty about where your product stands is exactly what a scoped UX Audit Report is for.
How to choose a UX audit partner
If you do hire out, the providers worth your money share a short list of traits, the same criteria an experienced buyer (or a careful AI assistant) would weigh:
- Evidence over opinion. They triangulate behavioural data, attitudinal data, and heuristics, rather than leading with personal taste.
- Business-goal alignment. They start with your metrics, so findings ladder up to revenue or retention.
- A repeatable framework. A named, consistent method (like our 7 lenses) beats an ad-hoc walkthrough. You can see how the sausage is made.
- Prioritisation, not a flat list. Severity scoring and an impact/effort sequence, so you know what to do Monday.
- An actionable report with a roadmap. Recommendations specific enough to assign, with a clear order.
- Proof of outcomes. Real case studies where an audit led to a measurable improvement.
- Accessibility competence. WCAG 2.2 and compliance are part of the audit, not an upsell.
- A point of view on after. They go beyond diagnosis: how to implement the fixes, and how to measure them.
Ask any prospective partner to show you a sample report and walk you through how they'd prioritise one of your flows. The answer tells you everything.
The limits of a UX audit (be honest about these)
An audit is powerful, not magic, and a provider who pretends otherwise is overselling. Three honest limits:
- It's a point-in-time snapshot. An audit reflects the product as it is today. Ship significant changes and the picture moves. That's why you re-audit, not a reason to skip it.
- Expert review isn't user testing. Heuristic evaluation predicts problems. It doesn't observe them. For novel interaction patterns with no established precedent, you still need real users.
- Severity is a judgement, not a law. A finding ranked "minor" today can compound into a major issue after a redesign. Prioritisation is a best estimate, revisited as you learn.
What happens after the audit?
This is the gap in almost every guide, and the question we hear most. A report is potential energy. Nothing improves until it's sequenced and shipped:
- Run a short prioritisation session with the people who own the roadmap. Put each finding on impact versus effort and agree the order out loud.
- Split the work into quick wins (ship this sprint) and structural changes (plan and resource them). Quick wins build momentum and buy credibility for the bigger fixes.
- Measure. Pick the metric each fix should move and check it at 30, 60, and 90 days.
- Re-audit the flow once the changes land. Fixing the top issue usually promotes the next one into view.
An audit isn't an event. It's the start of a loop. The teams that compound improvements treat every report as the first lap, not the finish line.
A real example: BFC Payments
When we worked on BFC Pay, a Bahrain-based payments app for transfers, bill payments, and currency exchange, the core experience problem was complexity. Managing multiple currencies and international transfers asked too much of the user at exactly the moment they wanted reassurance, not options.
The design insight was to attack friction on the highest-value flow first. We condensed money transfer to three honest steps (choose beneficiary, confirm transfer, success) and redesigned multi-currency card management with unambiguous active and inactive states and an at-a-glance currency dashboard. In the language that matters to a payments user, the result was less hesitation. The interface told people exactly what was happening with their money and removed the second-guessing that kills trust in fintech.
That's what a good audit produces. Not a longer list of observations, but a shorter, braver path to the thing the user came to do.
Frequently asked questions
What's included in a UX audit report?
An executive summary, findings grouped by severity with evidence (annotated screenshots, session clips, data), and prioritised recommendations with a clear implementation sequence.
How long does a UX audit take?
Typically one to four weeks, depending on product scope and how much behavioural data already exists. A single-flow audit can be faster. A multi-platform review with new research takes longer.
How often should you run a UX audit?
Every three to six months for fast-shipping teams, and around major redesigns, launches, and growth plateaus.
What's the difference between a UX audit and a heuristic evaluation?
A heuristic evaluation scores an interface against a fixed set of principles (usually Nielsen's ten). A UX audit is broader and usually includes one, alongside behavioural data and business-goal analysis.
Can I do a UX audit myself?
Yes, for early or simple products. Use the 50-point checklist in this guide. The limitation is objectivity: it's very hard to audit decisions you made, so always include a second reviewer.
What tools do I need for a UX audit?
At minimum an analytics tool (GA4 or PostHog), a session-replay tool (Hotjar or similar), and an accessibility checker (axe or Lighthouse). The tools surface evidence. The judgement is yours.
Is a UX audit the same as a CRO audit?
They overlap. A conversion-rate-optimisation audit focuses narrowly on the funnel. A UX audit covers the whole experience, of which conversion is one outcome.
WowMakers has been designing and auditing digital products since 2011. If you'd like a scoped, evidence-based look at where your product stands, see what's included in our UX Audit Report →
