Testing with 5 users uncovers 85% of usability problems.
Jakob Nielsen's foundational research still holds. We combine heuristic evaluation, session recording analysis, and task-based user testing into a severity-scored findings report — before a redesign starts, not after users complain.
What does a UX audit actually include?
A UX audit includes heuristic evaluation against Nielsen's 10 usability principles, session recording analysis, an accessibility review against WCAG 2.2, and task-based testing with real users — delivered as a severity-scored findings report rather than a design mockup.
The number that justifies auditing before building anything new: fixing a usability problem after launch costs 10 to 100 times more than fixing the same problem before launch, per Baymard Institute research. A redesign or new feature built on top of an unaudited interface just multiplies whatever's already broken.
- Panel, not one reviewer
- 3–5 independent evaluators, since one alone catches ~35%.
- Confirmed with real users
- Heuristics flag; task testing proves what is genuinely broken.
- WCAG 2.2 AA checked
- Automated scanning plus manual review of what tools miss.
- Ranked by impact
- Every finding severity-scored into quick wins and deeper fixes.
UX audit, A/B testing, or CRO?
These sit in sequence. The audit finds the problems; testing proves the fix.
| Service | Answers | Best fit |
|---|---|---|
| UX Audit | What's broken, and how severe is it. | The diagnostic step before a redesign, testing program, or CRO roadmap. |
| A/B Testing | Does a specific fix actually improve the metric. | Statistically proving a change works before it ships permanently. |
| Conversion Rate Optimization | What to prioritize and in what order. | The overall roadmap connecting audit findings to a testing plan. |
What a single reviewer misses, a panel catches.
And why heuristics alone are not enough to act on.
One evaluator catches ~35%; five approach 75%
A single expert catches roughly 35% of usability problems working alone. Every additional independent evaluator surfaces issues the others missed — up to a point. 3 to 5 evaluators is the practical ceiling before added cost outweighs new findings. Five-user task testing, separately, reaches about 85%.
Heuristics flag issues; users confirm which are real
Heuristic evaluation alone catches only about 36% of the usability issues real user testing finds, and roughly 43% of heuristic-flagged “issues” turn out not to be real errors once tested with people. Pairing both is what makes a finding worth acting on.
Two categories account for most failures: navigation is the single most common failure category across digital products, and forms are second, regardless of industry. Both get dedicated attention rather than a single general pass.
UX audit work businesses bring us.
Five stages — skip one, and you're guessing at priority.
Heuristic evaluation
Independent expert review of navigation, forms, checkout, and core flows against Nielsen's 10 principles.
Session recording review
20 to 40 recorded sessions analyzed for rage clicks, dead clicks, and drop-off patterns.
Task-based user testing
5 to 8 real participants completing your core tasks, watched, recorded, and debriefed.
Accessibility review
WCAG 2.2 AA compliance check, automated scanning plus manual review of what tools alone miss.
Analytics funnel analysis
Bounce rates, drop-off points, and conversion funnels reviewed to confirm heuristic findings with data.
Severity-scored findings report
Every issue scored for severity and business impact, ranked into quick wins versus deeper fixes.
A clear path from scope to a ranked fix list.
Four stages, with scope and baseline data agreed before evaluation starts.
Scope & baseline data
We agree the flows in scope and pull baseline analytics — bounce rate, funnel drop-off, and session data.
1 week · ScopingHeuristic evaluation
3 to 5 independent evaluators review every in-scope flow against Nielsen's principles.
1–2 weeks · EvaluationUser testing
We recruit and run task-based sessions with 5 to 8 real users to confirm and expand findings.
1–2 weeks · TestingReport & readout
We deliver the severity-scored report and walk your team through the prioritized fix list.
1 week · DeliveryThe tools we use for evaluation and testing.
Behavioural analytics, accessibility scanning, and user testing infrastructure.
Hotjar
Microsoft Clarity
Lighthouse
Maze
FigmaExplore more Conversion Optimization services.
UX audits are one of three services we offer under Conversion Rate Optimization.
Conversion Rate Optimization
The roadmap findings feed into.
ExploreA/B Testing
Prove the fixes an audit surfaces.
ExploreLanding Page Optimization
Where most audit findings get implemented.
ExploreUI/UX Design
Rebuilding the flows an audit finds broken.
ExploreWebsite Redesign
A full rebuild, once the audit says it is warranted.
ExploreCopywriting
The clarity fixes behind many findings.
ExploreTechnical SEO Audit
The crawl-side equivalent of this diagnostic.
ExploreWebsite Security Audit
The security-side audit for the same product.
ExploreUX audit questions
The things clients ask us most before starting a UX audit.
A UX audit includes heuristic evaluation against Nielsen's 10 usability principles, session recording analysis, an accessibility review against WCAG 2.2, and task-based testing with real users — delivered as a severity-scored findings report rather than a design mockup.
Testing with roughly 5 users uncovers about 85% of usability problems, per Jakob Nielsen's foundational research, which is why most audits run 5 to 8 participants per round instead of dozens. A single expert evaluator catches roughly 35% of issues alone, while a panel of about 5 evaluators working independently approaches 75%.
No. Heuristic evaluation is one component of a UX audit, not the whole audit. It's an expert review against fixed usability principles — narrow in scope and fast to run — while a full UX audit combines that heuristic review with analytics data, session recordings, accessibility checks, and real user testing, over days or weeks rather than hours.
Not on their own. Heuristic evaluations uncover only about 36% of the usability issues that real user testing finds, and roughly 43% of issues flagged through heuristic principles turn out not to be actual errors once tested with real users. Pairing heuristic review with task-based testing catches what either method misses alone.
10 to 100 times more than fixing the same problem before launch, according to Baymard Institute research, since a post-launch fix requires re-engineering shipped code, retesting, and redeployment, while a pre-launch fix is a design change. This gap is the core financial argument for auditing before a redesign, not after one.
A UX audit is diagnostic — it finds and prioritizes usability problems using expert evaluation and user testing. A/B testing is the statistical proof layer that comes after, validating whether a specific fix actually improves the metric it targets. Most audit findings should feed directly into a testing roadmap, not get shipped on faith.
$3,000 to $25,000, depending on scope. A focused audit of a single user flow, like checkout, runs $3,000 to $6,000. A full product audit with heuristic review, session analysis, and real user testing runs $8,000 to $18,000, and enterprise-scale audits with multiple user roles run $18,000 to $25,000 or more.
Navigation and forms are the two most common failure categories across digital products, regardless of industry, per Baymard Institute research — which is why both get dedicated attention in every audit we run rather than a single general pass.