UX Audit: How a Design Studio Finds What Your Site Is Losing

What does a UX audit actually tell you?

A UX audit tells you where real visitors struggle on your site, why they struggle, and which problems cost you the most. It turns a vague sense that "something is off" into a ranked list of evidence-backed fixes.

The output is not a mood board or a list of opinions. Every finding should point to a specific screen, explain the principle it breaks, and carry a severity rating.

Three ways to look at a website

Audits combine three kinds of evidence. Each one sees something the others miss.

• Expert inspection. A trained reviewer walks through key journeys and checks them against known usability principles.
• Behavioral data. Analytics, heatmaps and session recordings show where people actually stop, hesitate or leave.
• User testing. Real participants try real tasks while you watch. It is the most revealing method, and the most expensive.

The Nielsen Norman Group frames expert reviews as a broader version of a heuristic evaluation. The reviewer draws on guidelines, cognitive psychology and hands-on research experience. Their advice is to combine expert reviews with usability testing, because each method catches different issues.

What a good finding looks like

Compare two findings. The first: "The homepage feels cluttered." The second: "On mobile, the main call to action sits below three autoplaying sections, so most visitors never reach it." The first is a reaction. The second is something a designer and a developer can act on this week.

Severity is what makes a list usable. Jakob Nielsen's classic severity scale weighs three factors: how often a problem occurs, how hard it is to overcome, and whether it keeps coming back. Problems are then rated from cosmetic to catastrophic. Without that ranking, a report treats a typo and a broken checkout as equals.

The numbers behind invisible friction

Most friction never reaches your inbox. Visitors who get confused rarely complain. They simply leave, and your analytics records a bounce with no explanation attached.

Frustration is the norm, not the exception

Large-scale session data shows how common this is. According to Contentsquare (2025 Digital Experience Benchmarks), 40% of online visits in 2024 involved user frustration. The same report links slow-loading content to 53% of users leaving after a single page. The analysis covered 90 billion sessions across 6,000 websites.

That figure reframes the question. You are not asking whether your site frustrates people. You are asking where it does, and how much that costs.

Forms are where good intentions go to die

Forms are the most measurable friction point, and checkout data makes the case clearly. According to Baymard Institute (Checkout Optimization: 5 Ways to Minimize Form Fields in Checkout, June 2024), the average checkout in 2024 had 11.3 form fields. Most sites only need 8. The research also found that the number of fields matters more than the number of steps.

You may not sell anything online. The lesson still applies to every contact form, booking flow and application page. Each extra field is a small tax on intent, and an audit is how you find out which ones to cut.

Accessibility issues hide in plain sight

Accessibility is where audits find the most issues, and where brand sites are often most exposed. According to WebAIM (The WebAIM Million, 2026 report), 95.9% of the top one million home pages had detectable WCAG failures. Low-contrast text alone appeared on 83.9% of them.

The standard behind those checks is public. The W3C's WCAG 2.2 is the reference, in its current version dated December 12, 2024. It sets a minimum contrast ratio of 4.5:1 for normal text. It also asks for touch targets of at least 24 by 24 CSS pixels at level AA. Light gray captions are the classic contrast failure, and tiny icon buttons put the target size rule at risk. Our guide to website accessibility compliance covers compliance in more depth.

Who should run your UX audit?

Someone who did not design the site, and who understands how it was built. Those two conditions rarely meet in the same person, which is why the choice of auditor shapes the result as much as the method.

Why fresh eyes matter

The Nielsen Norman Group makes a simple point about expert reviews. The best reviewer is someone who was not involved in creating the design. Teams that built a site see what they intended, not what is on the screen. It is the same reason writers miss their own typos.

This does not mean your team has nothing to add. Your sales and support people hear the real objections every week. A good auditor interviews them before opening a single page.

In-house, specialist or studio

Each option has a clear trade-off.

• An in-house review is fast and cheap. It suffers from familiarity, and it rarely challenges internal assumptions.
• A research specialist brings rigor and benchmarks. The recommendations may ignore your codebase, your budget and your brand.
• A studio that designs and builds can judge what is feasible and what it costs. Its risk is bias toward rebuilding, so ask it to justify every major recommendation.

Questions to ask before you hire an auditor

A short call is enough to separate a real audit from a repackaged scan. Ask how many reviewers will inspect the site independently. Ask which journeys and devices are in scope, and why. Ask what data they need from you, and what they will do if you have none. Finally, ask to see an anonymized finding from a past report. If it reads like an opinion rather than evidence, you already know what the full report will look like.

An audit is only worth what gets shipped

Even the best auditors face an implementation gap. According to Baymard Institute (UX Audit Service page, accessed October 2026), its audits deliver 40 research-backed improvements prioritized by conversion impact. Clients implement more than 21 of them on average, which is barely more than half. And Baymard says its research is used by 71% of Fortune 500 ecommerce companies.

That gap is the real argument for involving whoever will build the fixes. A recommendation that ignores the codebase is a recommendation that waits.

Inside a studio-led audit

A studio audit adds one question most checklists never ask: does this experience match the level of the brand? Here is how that changes the work.

Three lenses on every screen

For each key screen, we suggest reading the page through three lenses rather than one.

• Use. Can a first-time visitor understand what to do and do it without effort?
• Positioning. Does what they see match the price point and the ambition of the offer?
• Buildability. Can the fix ship in the existing code, at what cost, and with what risk?

A finding that improves use but damages positioning needs a better solution, not a shrug. A finding that scores high on both but cannot be built this quarter goes on the roadmap with a date, not into a drawer.

How a finding gets scored

Here is a worked example of the grid in action. Take a homepage hero where the headline is set in pale gray over a looping video.

• Use: high impact. On mobile, the offer is hard to read in the first seconds, and the contrast falls short of WCAG.
• Positioning: high risk if handled badly. The video and the typeface carry the brand. Deleting them would solve the contrast and lose the identity.
• Buildability: low effort. A darker overlay behind the text and a slightly heavier weight are a few lines of CSS.

The recommendation writes itself: keep the video, fix the overlay, ship it this week. A checklist alone would have suggested removing the video. A design-only review might have missed the contrast failure entirely. Scoring all three lenses together is what produces the fix that is both cheap and on-brand.

When best practices erase personality

Here is a pattern we often see with brands moving upmarket. An audit flags the long editorial homepage as "too much scrolling." It recommends collapsing the story into a grid of identical cards with short teasers. On paper, it is cleaner. In practice, the brand that sold itself through narrative now looks like every competitor in its category.

The fix is not to ignore usability. It is to solve the real problem, which was usually orientation, not length. A clear progress cue or anchored sections can keep the story intact. Motion can help too. Thoughtful web microinteractions confirm an action and reinforce the brand at the same time.

Techunt: when the audience is split in two

The Techunt project shows why the positioning lens matters. Techunt is a recruitment firm specializing in AI, Web3 and blockchain talent. In a crowded market, its challenge was to stop being seen as a simple middleman. It needed to come across as a credible player for the most specialized candidates and companies.

That makes it a two-sided site. Hiring companies want proof of sector expertise. Scarce engineers want to feel understood before they share a résumé. The project page shows the answer in its structure: a dedicated page for companies, a candidate page and a separate form.

A standard audit would check that the form works. A studio audit asks whether each audience finds its own proof in the first few seconds. On a site like this, a smooth journey that sounds like a generic staffing agency still fails.

[VISUEL ORIGINAL À FOURNIR : annotated screenshot of a Metabole project screen or an excerpt from an audit prioritization grid. It should show one finding, its severity, the developer's effort estimate and the recommended fix.]

The developer's view

[AVIS À VALIDER PAR MATTEO]

Matteo Courquin, co-founder and lead developer, on audits: "An audit written without looking at the code is a wish list. Every finding should come back with an effort estimate from the person who will actually build it. And a usability rule on its own is never a reason to delete what makes a brand recognizable. You look for another solution first."

Red flags in a UX audit proposal

Most disappointing audits were predictable from the proposal. Three warning signs come up again and again.

Red flag 1: a fixed number of findings promised upfront

"You will receive 150 recommendations." A promise like that rewards volume, not insight. Clients often accept it because it sounds thorough. Then the team spends a week reading and still does not know what to fix first. Ask instead how findings will be ranked, and how many will be flagged as top priority.

Red flag 2: no access to your data requested

Here is a story that comes up with first-time audits. The auditor spends days polishing the homepage and the About page. Nobody checked the analytics, which would have shown that most visitors land directly on blog articles and service pages. The team trusted the homepage because it is the page they look at most. An auditor who never asks for your data will audit the site you see, not the one your visitors use.

Red flag 3: recommendations with no owner

We see this when the audit and the fixes are handled by different vendors. The report recommends a sticky booking bar on every page. The developers then find it clashes with the scroll-driven animations that define the site. Everyone agrees the goal was right, and the ticket sits untouched for months. A proposal should state who estimates the effort of each fix, and when.

How much should a UX audit cost?

A light expert audit starts at around €2,500, and the price rises mainly with scope and with user testing. Expect a prioritized report, visual evidence and a live walkthrough, not just a PDF.

What your money should buy

Whatever the budget, these deliverables are non-negotiable.

• An executive summary that a founder can read in five minutes.
• A ranked list of findings, each tied to a screen, a principle or a data point, with a severity rating.
• Concrete recommendations, with examples of how other sites solve the same problem.
• A list of strengths, so a redesign does not destroy what already works. The Nielsen Norman Group recommends it for every expert review.
• A live debrief, so designers and developers can challenge findings before estimating them.

Where the price comes from

Budget around €2,500 for an entry-level audit: an expert review of a few key journeys and a ranked report. Prices are in euros, the currency the studio bills in.

That figure matches market data. According to La Fabrique du Net (UX/UI agency rates 2026), the cheapest 10% of design projects sit at €2,500. Short engagements show a median of €4,000. The French benchmark draws on 761 real design project budgets.

From there, three variables push the price up. The number of journeys and screens, the volume of data to analyze, and above all moderated user testing. As an indication, the same benchmark puts user research and testing at around €30,000 at the median.

You can also keep the bill down without weakening the result. Limit the audit to the two or three journeys that carry revenue. Share your analytics access on day one, so nobody bills hours to guess. And gather your known constraints upfront: tech stack, brand rules and budget for fixes. A tight brief usually buys more insight per euro than a bigger scope.

The logic is straightforward. An expert review finds the likely problems quickly. Testing proves which ones cost real money. For a first audit, an expert review is usually the best value, with testing reserved for the questions it cannot settle. Our breakdown of UI UX design pricing covers what the fixes themselves tend to cost.

Your most common questions about UX audits

How long does a UX audit take?

It depends on scope more than on page count. A focused review of one or two journeys moves much faster than a full-site audit with user testing. For reference, Baymard quotes roughly 15 to 25 business days for a full-site audit. Ask any provider to break the timeline into scoping, review and prioritization. If someone promises a complete audit of a complex site in three days, they are likely running an automated scan and calling it an audit.

How is a UX audit different from a UI audit?

A UI audit looks at the visual layer: spacing, type, color, component consistency and contrast. A UX audit looks at whether people can understand the offer and complete key tasks. The two overlap, and many reports cover both. The distinction matters when you buy. A UI-only review can make a site more polished without fixing why visitors leave. Ask whether user journeys are in scope before you sign.

Can AI tools run a UX audit for us?

AI tools are useful for the mechanical parts. They can flag contrast failures, summarize session recordings and spot unusual drop-off patterns quickly. What they cannot do yet is judge your business context. They do not know which audience matters most, which brand codes are non-negotiable, or what your developers can ship next quarter. Treat them as a fast first pass, then have a person decide what the findings mean.

Is a UX audit worth it for a small site?

Usually, yes, as long as the scope is small too. A brand site with five pages and one contact form does not need a month of research. It needs a focused review of the homepage, the main offer page and the conversion path. That kind of targeted audit often surfaces the two or three changes that matter most. The risk for small sites is overbuying, not underbuying.

Should we audit our competitors' sites as well?

It helps when you are repositioning or entering a crowded market. Reviewing two or three competitor sites against the same criteria shows what visitors in your category already expect. It also shows where you can stand out. Baymard, for example, offers confidential competitor audits as an option. Keep it focused, though. The goal is benchmarking your gaps, not copying someone else's layout.

Can a template-based site pass a UX audit?

It can, and many do. Templates often handle basic usability well because they follow common conventions. The weak points tend to be elsewhere. Think generic messaging, a layout that fits any brand, and little room to fix structure without fighting the theme. If an audit keeps hitting that ceiling, our comparison of custom websites versus templates will help you weigh the next step.

Fix the friction, keep the signature

A UX audit earns its fee when it changes what is live. It should show where your site loses people, prove why, and rank what to fix first. It should also protect the things that make your brand worth choosing.

Hold any audit to three standards. Evidence over opinion. Priorities set with the people who will build the fixes. And a positioning lens, so every change makes the site clearer without making it generic. That last point is where web art direction and usability stop pulling in opposite directions.

Alan Chevereau, SEO consultant. He works with Metabole Studio on its content strategy and organic visibility, in English and French.