Website UX Audit: The 5-Step Data Method

Table of Contents

• Before you start: what you need for a reliable audit
• Step 1, Map the journeys that matter for your business
• Step 2, Measure real behavior with GA4
• Step 3, Assess technical performance with Lighthouse
• Step 4, Observe real usage with heatmaps and recordings
• Step 5, Prioritize fixes with an impact/effort matrix
• Mistakes to avoid when auditing UX yourself
• Frequently asked questions

Before you start: what you need for a reliable audit

A reliable UX audit rests on three access points and a choice of pages, not a fifty-line spec sheet. Get these in place first, or you'll be working blind.

The classic trap I run into regularly: launching the analysis on dirty data. A GA4 account installed three weeks ago tells you nothing reliable. You need at least 30 days of history, ideally 90, to smooth out spikes and read real trends.

What data and tools do you need?

Three tools are enough to get started, all free. You need admin access to Google Analytics 4 (which records your visitors' behavior), Google Search Console (which shows how you appear in Google search), and a heatmap tool like Microsoft Clarity, completely free with no traffic limit.

Add Google Lighthouse, built directly into Chrome's developer tab, and PageSpeed Insights for field measurements. No license to pay for at this stage. Above all, check that conversion tracking is properly set up in GA4: without a purchase or form-submission event, the analysis loses half its value.

How do you define your priority journeys and pages?

Focus on the pages that drive revenue, not the ones with the most visits. A heavily-viewed "legal notice" page has zero impact on your sales.

List your three to five key pages: homepage, flagship category, best-selling product page, cart page, contact form. For a services site, that's more likely the services page and the quote request form. That's where your conversions happen, so that's where every second of friction costs you money.

Step 1, Map the journeys that matter for your business

Draw the actual path a visitor follows to purchase or contact you, step by step, before measuring anything. Without this map, your numbers float in a vacuum.

Grab a sheet of paper or a spreadsheet. Note each step: site entry, product page, add to cart, cart page, payment form, confirmation. For a services site, it would be: homepage, service page, pricing page, quote request form. Every transition between two steps is a place where you lose visitors. Your job is to know how many, and why.

The top 10 Google results on this topic list tools without ever connecting the map to the conversion funnel. That's the mistake. A journey analysis disconnected from revenue produces nice diagrams and zero decisions.

What does website UX actually mean for conversion?

UX, on the conversion side, is the sum of the friction between the moment a visitor arrives and the moment they buy or contact you. Think of a supermarket: if the checkouts are tucked away down a poorly-marked aisle, people abandon their cart. On a website, it's the same, except you don't see people leave.

It shows up in the numbers: a high bounce rate on a product page, a 70% cart abandonment rate, a form nobody finishes. The three pillars of UX remain usefulness (the content answers the need), usability (people find their way without effort), and desirability (the whole thing inspires trust). To understand where these leaks happen in the funnel, our analysis of the e-commerce conversion funnel breaks down where the losses concentrate, often at the cart and checkout stage.

Step 2, Measure real behavior with GA4

Google Analytics 4 reveals where your visitors drop off, as long as you look at the right reports. The engagement rate and funnel paths are worth ten surveys.

When I open a GA4 account for a diagnostic, I never look at the default dashboard. It buries what matters under the secondary stuff. I go straight to Explorations, the space where you build your own custom reports, documented in Google's official GA4 Explorations help.

What's the best analytics tool for a website?

For behavior, Google Analytics 4 remains the most complete free reference for a small or medium business. Matomo offers a privacy-friendly alternative that can be self-hosted in France, but it requires more setup.

The "best" tool depends on your need. GA4 for behavior and conversions. Lighthouse for performance. Microsoft Clarity or Hotjar for visual observation. No single tool covers all three angles, which is exactly why you need to combine them. To go further on reading reports, our method for actually making sense of Google Analytics breaks down the six reports that really matter.

Which GA4 reports reveal friction points?

Three GA4 reports expose friction: the funnel exploration, exit pages, and engagement rate by page. They show where, when, and how many visitors drop off.

Start by building a funnel exploration that mirrors the steps on your map. GA4 then shows the percentage drop between each step. A step that loses 60% of traffic is your immediate priority. Next, look at engagement rate by page: below 50%, the page struggles to hold attention. Cross-reference with exit pages to spot where visitors leave the site right before converting.

On a ready-to-wear store I worked with last year, the GA4 funnel showed a 48% drop between cart and checkout. The culprit: a "promo code" field that pushed people to go look for a code elsewhere, and they never came back. Two lines of code to hide that field, and the drop-off was cut in half.

Step 3, Assess technical performance with Lighthouse

Lighthouse measures the perceived speed of your pages and scores your Core Web Vitals, the three metrics Google uses to judge experience. A slow site drives people away before the content even loads.

Run Lighthouse from Chrome's developer tab, with mobile mode on. The report gives you a score out of 100 and the breakdown of the three key signals. Except the lab score doesn't replace the field data from PageSpeed Insights, which measures the experience of your actual visitors.

What Core Web Vitals thresholds should you aim for?

Aim for an LCP under 2.5 seconds, an INP under 200 milliseconds, and a CLS under 0.1. These thresholds define a good experience according to Google.

Let's translate each metric simply. LCP is the time it takes to render the page's main element, usually the image or the headline: it should stay under 2.5s. INP measures the delay between your click and the page's visible reaction: under 200ms. CLS measures page stability, in other words the absence of those content jumps that make you click the wrong spot: under 0.1. These values come from Google's official Core Web Vitals documentation on web.dev.

To fix these signals step by step, our guide to optimizing Core Web Vitals for small businesses walks through the process, from LCP to visual shifts.

Why should mobile come first?

Mobile comes first because it often carries more than half of your traffic and concentrates the worst friction. A site that's fast on desktop can fall apart on a smartphone.

Google's docs highlight the overall score. In practice, on the accounts we diagnose, I always start with mobile: according to Arcep's digital usage barometer, smartphone internet use in France now exceeds desktop use. A product page that loads in 2s on fiber can take 6s on 4G. That's where you're losing sales, not on the developer's desktop machine.

This is exactly the kind of gap, invisible in a one-off audit, that Lysible monitors continuously between your mobile and desktop visitors.

Step 4, Observe real usage with heatmaps and recordings

Heatmaps (click and scroll heat maps) and session recordings show what the numbers will never tell you: why people drop off. It's the missing link in the analysis.

GA4 tells you a page converts poorly. The heatmap tells you why. Microsoft Clarity offers both for free, with no traffic cap. Hotjar offers a free plan up to 35 sessions a day, enough for a small business just starting out.

Hotjar and heatmaps: what should you look at first?

Start with the scroll map and click map of your conversion pages. They reveal whether your visitors see your buttons and where they click for nothing.

The scroll map shows how far down people go. If 70% stop before reaching your buy button, the button is too low: it's that simple. The click map reveals "dead clicks," those spots where people click thinking it's interactive, with no reaction at all. A non-clickable visual that looks like a button creates frustration and silent departures.

What does UX analysis mean beyond the numbers?

Beyond the numbers, UX analysis means observing real behavior to understand the intent and frustration behind every statistic. The Nielsen Norman Group, a global authority in the field, points out that combining quantitative and qualitative methods gives a far more reliable read than a single source.

Here's a concrete case. On an accessories store we tracked in early 2025, GA4 showed a product page with a decent engagement rate but few cart additions. The heatmap explained everything: the "Add to cart" button blended into the background color on mobile. Nobody saw it. A color change, stronger contrast, and cart additions rose about 22% in three weeks. No single number alone would have pinpointed that friction.

Step 5, Prioritize fixes with an impact/effort matrix

Prioritize your fixes with a four-quadrant matrix crossing revenue impact and implementation effort. You tackle high impact, low effort first.

This is the step everyone skips, and it's the most profitable one. After your four audit steps, you have a list of friction points. Without prioritization, you'll either paralyze the team or launch an unnecessary full redesign. In most audits I run, two or three friction points in the funnel account for most of the possible gains.

Which friction points should you fix first?

Fix high-impact, low-effort friction points first: an invisible button, an unnecessary form field, an oversized image. These fixes take a few days and pay off fast.

Sort each friction point into one of the four quadrants below. The "high impact, low effort" quadrant is your list for the next two weeks. The rest waits or gets scheduled based on your resources.

This priority-based approach ties into our method for conversion rate optimization for small businesses, which is precisely about tackling the leaks with the highest return on effort first.

Mistakes to avoid when auditing UX yourself

The biggest mistake in a DIY UX audit is looking at a single data source and drawing final conclusions from it. Always cross-reference behavior, performance, and observation.

Three traps keep coming up. First: relying solely on the owner's gut feel, since they know their site by heart and no longer see the friction a new visitor hits. Second: judging from too short a data window (a week of traffic is worth nothing). Third: wanting to fix everything at once, which dilutes effort and stalls decisions.

Two other common mistakes deserve a mention. Analyzing only the desktop version when mobile carries most of the traffic means you're looking at the wrong storefront. And confusing high traffic with an important page: a heavily-visited page with no role in conversion doesn't deserve your priority attention.

Do you need an agency to analyze your site's UX?

No, a small or medium business can capture about 80% of UX gains in-house with GA4, Lighthouse, and a free heatmap tool. An agency becomes useful for heavy redesigns or advanced user testing.

The market often sells UX audits as a service worth several thousand euros. The reality is more nuanced. To spot and fix the obvious friction points in your funnel, you don't need anyone outside your team. According to France Num, the French government's program supporting small business digitization, mastering these basic tools is well within reach of small organizations. Delegate when you're after complex A/B tests, user interviews, or a full site architecture redesign. Not to read a GA4 funnel. If you're unsure, our comparison on whether a web performance audit is worth it lays out the right benchmarks.

Scaling your UX analysis with Lysible

A one-off audit ages fast. Friction comes back, your pages change, your campaigns shift the traffic mix. The real challenge is tracking the experience over time without reopening five tools every week. That's exactly the gap we see on the accounts we work with: the data exists, but it sits dormant in separate tabs. Lysible centralizes GA4 behavior, Lighthouse performance, and UX signals in a single view, translated into clear decisions. You spot a friction point when it appears, not three months later. For a small or medium business, that's the difference between putting up with your site and actually running it.