User Experience Audit: How to Run One and Act on the Results

A user experience audit is an expert-led, evidence-based review of a live digital product, measured against established usability principles, your own analytics, and the business goals the product is meant to serve. The output is not a verdict; it is a ranked list of problems with a recommended action for each. Most guides that rank for this term spend nine tenths of their words on how to run the audit and then trail off at “write a report”. That is the part that actually decides whether the work was worth paying for, so this guide weights the back half: scoring findings, prioritising them, turning them into a tracked backlog, re-measuring, and keeping the product audited after the consultant has gone home.

Audit versus usability testing: not the same thing

These get used interchangeably and they should not be. The distinction is about where the evidence comes from.

A UX audit is top-down. An evaluator infers problems by walking the product against a fixed set of standards (Nielsen’s heuristics, accessibility rules, your analytics). Usability testing is bottom-up: you watch real users attempt real tasks and infer problems from where they stall. Audits suit products that have been live a while and carry a backlog of data; testing suits new features and flows that have no history yet. The strong teams run both: an audit to find systemic issues fast and cheaply, then targeted testing to validate the fixes before they ship. If your interest is specifically the funnel where money leaks out, our guide to conversion funnel optimisation covers that narrower diagnosis in detail.

How to run a UX audit, step by step

Treat this as a ux design audit spine. Each step feeds the next.

1. Define goals and scope. Pick one contained flow, the checkout or the signup, not “the whole site”. A scoped audit produces fixable findings; an everything audit produces a 60-page document nobody reads.
2. Identify stakeholders. Product, design, support, sales and an exec sponsor. Support tickets in particular are a free, pre-sorted list of what already annoys users.
3. Gather the data. Quantitative first: drop-off and bounce, task success, session length, the steps where users vanish. Then qualitative: support tickets, interviews, personas, and the real user flows people take rather than the one you designed.
4. Run a heuristic evaluation. Walk every page in the flow against Nielsen’s ten usability heuristics, logging each issue with a screenshot and the heuristic it breaks. Use three to five independent evaluators if you can; around three reviewers catch roughly 60 percent of the issues a much larger study would surface, which is why a lone auditor misses things a small panel would not.
5. Severity-rate and prioritise. Covered in full below; this is the step that earns the fee.
6. Write the report. Findings and recommendations, ranked. Not an exhaustive log of every cosmetic nitpick.
7. Re-measure. After fixes ship, re-run the same baseline metrics you captured in step three. No baseline, no proof.

Nielsen’s ten heuristics, in plain terms

First published in 1994 and still the standard checklist:

• Visibility of system status
• Match between the system and the real world
• User control and freedom
• Consistency and standards
• Error prevention
• Recognition rather than recall
• Flexibility and efficiency of use
• Aesthetic and minimalist design
• Help users recognise, diagnose and recover from errors
• Help and documentation

Walk the flow ten times, once per heuristic, rather than trying to hold all ten in your head on a single pass. You will catch more.

What a UX audit costs and how long it takes in the UK

These figures come from UK agency and consultancy pricing pages rather than any official body, so treat them as indicative, not gospel. A comprehensive audit in Britain typically lands somewhere between roughly one and five thousand pounds, with the broader market stretching from a few hundred for a light review to ten thousand for a large, complex estate. Consultant and agency day rates sit roughly in the range of one hundred and twenty to two hundred and fifty pounds an hour, and a thorough audit usually runs to about 20 to 40 hours of work. The single biggest cost driver is scope: a five-page checkout audit is a different job from a 200-page marketplace.

You can do a basic audit yourself with the heuristic checklist, your analytics and a free accessibility checker. The reason teams pay for one anyway is independence: in-house evaluators are blind to conventions they invented, and an outsider catches the assumptions nobody questions.

What to do with the results: severity scoring

This is where the genuine value is, and where most articles wave their hands. Raw findings are useless until they are ranked, because you will always find more problems than you have time to fix.

Use Nielsen Norman Group’s 0 to 4 severity scale. It is the canonical reference and worth citing in your report so nobody argues the scoring is arbitrary. The full definitions live in NN/g’s severity ratings for usability problems.

Severity is not one number plucked from the air. NN/g defines it as a blend of three factors: how frequently users hit the problem, how hard it is for them to overcome, and whether it persists once they have seen it or trips them every single time. A rare, recoverable, one-off annoyance is a 1; a frequent, blocking, repeated failure is a 4.

A worked example of a single issue log row, the kind of artefact your report should contain one of per finding:

To layer business value on top of raw severity, score each finding on an impact-effort matrix, or weight it: severity multiplied by (commonality plus frequency) multiplied by business importance. A severity-3 problem on your highest-traffic page outranks a severity-4 problem on a page twelve people see a month. Plain rule of thumb: severities 3 and 4 are blockers you act on now, 1 and 2 go to the backlog. This prioritisation logic is the same muscle used in conversion rate optimisation, where findings become a ranked list of tests rather than a wish list.

From report to tracked backlog to re-measurement

A report that sits in a shared drive changes nothing. The handoff that works:

1. Every severity 3 and 4 finding becomes a ticket in your normal tracker, with its issue log row, screenshot and recommended fix attached. Do not invent a separate “UX backlog” nobody looks at.
2. The severity 1 and 2 findings go into the same backlog at lower priority, so they surface during quieter sprints rather than being lost.
3. Pin the baseline metrics from step three to the relevant tickets. When the fix ships, you re-measure those exact metrics, not vaguely similar ones.

Re-measuring the same baseline is the difference between ux optimisation and redecorating. If task success on the checkout was 71 percent before and 84 percent after, you have evidence. If you changed the metric definition halfway through, you have an argument.

SEO and user experience: where they overlap

SEO and user experience are not separate disciplines at the technical layer, because Google measures parts of the experience directly. Core Web Vitals are field-data UX signals and remain a ranking input in 2026, working as a tie-breaker in competitive niches rather than the dominant factor: good content with average vitals beats perfect vitals with weak content. The three current metrics and Google’s “good” thresholds are worth checking on every audited page.

One freshness check for any audit you commission: INP officially replaced First Input Delay as a Core Web Vital on 12 March 2024, and FID was removed from Search Console that day. If a report still cites FID, it is out of date. Google’s own Core Web Vitals documentation is the authoritative reference. The user impact is blunt: pages loading under two seconds see bounce around 9 percent against roughly 38 percent for pages over five seconds, and with more than 60 percent of searches now on mobile, the mobile experience is the primary ranking signal, not an afterthought.

Accessibility belongs in every UK audit

Most US-written guides skip this entirely, which is a problem if you operate in Britain. Accessibility is not an optional extra in a UK UX audit; for many organisations it is a legal duty.

Public-sector bodies must meet WCAG 2.2 levels A and AA and publish an accessibility statement, under the Public Sector Bodies (Websites and Mobile Apps) Accessibility Regulations 2018. The EHRC enforces this through investigations, unlawful act notices and court action. The rules are set out on GOV.UK’s accessibility requirements guidance.

Private-sector sites are covered by the Equality Act 2010, which requires “reasonable adjustments” for disabled users. It does not name a technical standard, but it underpins the duty. The newer pressure is the European Accessibility Act, which became applicable on 28 June 2025 and extends accessibility duties (a WCAG 2.1 AA baseline) to private-sector digital services such as ecommerce and banking. The EAA does not apply directly to the UK, but a UK business selling to EU consumers must comply for that audience, and enforcement has teeth: one German retailer was fined 150,000 euros over checkout alt-text failures. Run an accessibility checker over every audited page as standard.

End user experience monitoring: keeping it audited

An audit is a point-in-time snapshot. The continuous layer that tells you whether your fixes held up is end user experience monitoring, usually called Real User Monitoring (RUM).

RUM is passive: it collects performance and experience data from the browsers of your actual visitors in production, which is field data, as opposed to synthetic or lab tests run on a clean machine. It captures real page load, INP and responsiveness, JavaScript errors, network requests, and session interactions across the real devices, networks and locations your users actually have, with alerting when something degrades. By 2026 RUM has shifted from pure load-speed measurement toward broader experience insight. Vendor explainers from Dynatrace and New Relic describe the category in more depth.

The editorial point: the audit finds the problems, you fix and re-measure them once, and RUM watches the live product so the next regression shows up in a dashboard rather than in a quarter of lost revenue. One catches systemic design faults; the other catches the day the new payment script started failing on a particular Android browser.

A note on cart abandonment, the audit’s favourite target

Checkout is where audits earn their fee, and the numbers explain why. Baymard Institute’s research across 50 studies puts documented average cart abandonment at 70.22 percent. Among shoppers who were not merely browsing, the stated reasons are: extra costs too high (39 percent), delivery too slow (21 percent), not trusting the site with card details (19 percent), forced account creation (19 percent), and a checkout that is too long or complicated (18 percent). The structural clue for your heuristic walk: the average checkout has 23.48 form elements against an optimal 12 to 14. If your audit finds a checkout with thirty fields and a mandatory account, you have found severity-4 work before you have watched a single recording.

Frequently asked questions

What is a UX audit and what does it include? It is an expert-led review of a live product against usability heuristics, your analytics and your business goals. It includes a scoped flow, a heuristic evaluation, an accessibility check, a severity-ranked list of findings, and recommendations. The deliverable is a prioritised report you can turn into tickets, not an opinion piece.

What is the difference between a UX audit and usability testing? An audit infers problems top-down by measuring the product against fixed standards, and suits products with a history and data. Usability testing infers problems bottom-up by watching real users attempt tasks, and suits new features. Use the audit to find issues and testing to validate the fixes.

How much does a UX audit cost in the UK and how long does it take? Indicative agency pricing puts a comprehensive audit at roughly one to five thousand pounds, with day rates around one hundred and twenty to two hundred and fifty pounds an hour and about 20 to 40 hours of work. These are vendor figures, not official ones, and scope is the main driver.

Can I run a UX audit myself or do I need an agency? You can run a basic audit with the heuristic checklist, your analytics and a free accessibility checker. Teams hire out anyway because an independent evaluator catches the conventions in-house people invented and stopped questioning, and a small panel of evaluators finds more than one person working alone.

How do I prioritise the issues an audit finds? Score each finding on Nielsen’s 0 to 4 severity scale, blending frequency, impact and persistence, then weight by business value with an impact-effort matrix. Severities 3 and 4 are blockers you fix now; 1 and 2 go to the backlog for later sprints.

How often should I run a UX audit? Run a full audit after any major redesign or roughly once a year, and a scoped audit whenever a key metric drops. Between audits, lean on Real User Monitoring so regressions surface continuously instead of waiting for the next scheduled review.