Most small business owners know something is wrong with their website. Traffic is decent, but the leads do not come. The problem is rarely one big broken thing. It is usually a dozen small friction points spread across navigation, page speed, forms and mobile screens.
A UX audit of your website is how you find them. And contrary to what most agencies suggest, you do not need a designer, a research budget or a six-week engagement to run a first pass. With a laptop, a free analytics stack and about four focused hours, you can produce a prioritised list of fixes ranked by revenue impact.
This guide gives you the exact process we use at Magnetik, stripped down to what a non-designer can execute alone, including the specific questions to ask at each step.
What is a UX audit (and what it is not)
A UX audit is a structured evaluation of your website against known usability principles and your own analytics data, designed to identify the points where visitors get confused, slowed down or discouraged before they convert.
Two things it is not:
- It is not a design opinion. “I do not like the blue” is not an audit finding. “63% of mobile users never scroll past the hero, so the pricing block is invisible” is.
- It is not a technical SEO crawl. There is overlap (speed, mobile, internal linking), but a UX audit measures human friction, not crawl budget.
The output of a good audit is a single document: a list of problems, evidence for each, and an estimated euro or dollar value for fixing it. The piece Everything I know about running UX Audits makes a good next read.
The free toolkit you need before you start
Set these up first. If you are missing GA4 or a heatmap tool, install them and give yourself 7 to 14 days of data collection before auditing. Everything below has a genuinely usable free tier.
| Tool | What you use it for in the audit |
|---|---|
| Google Analytics 4 | Landing page performance, device split, exit points, conversion rates per page, funnel drop-off |
| Google Search Console | Search intent behind your top pages, click-through rates, Core Web Vitals and mobile issues at scale |
| Microsoft Clarity | Free unlimited heatmaps, session recordings, rage clicks, dead clicks, scroll depth |
| PageSpeed Insights | Real-world Core Web Vitals plus lab diagnostics per template |
| Chrome DevTools | Device emulation, throttled network testing, tap target inspection |
| Your own phone | The single most underrated audit tool. Real thumbs, real network, real distractions |
Optional additions: a screen recorder to capture findings, and a spreadsheet for your findings log. Create four columns now: Page, Problem, Evidence, Estimated impact.
The 10-step UX audit process
Step 1: Define the one conversion that matters
An audit without a goal turns into a list of cosmetic complaints. Pick the single action that generates revenue: a quote request, a booking, a checkout, a phone call.
Questions to ask:
- What is the one action that, if it increased 20%, would change my month?
- What are the top three paths visitors take to that action? (for example: Google to service page to contact form)
- What is my current baseline conversion rate for that action, on desktop and on mobile separately?
Write the baseline down. Everything you find later gets judged against it. dovetail.com makes the same point with more data.
Step 2: Pull the quantitative picture from GA4
Spend 30 minutes here before you look at a single design element. You are looking for where attention and money already are.
In GA4, open Reports > Engagement > Landing page and add your key conversion event as a metric. Then open Reports > Tech > Overview for the device split.
Questions to ask:
- Which 10 pages receive 80% of my entrances?
- Which of those pages have a conversion rate below site average despite high traffic? (These are your priority pages.)
- What percentage of sessions are mobile, and is the mobile conversion rate more than 30% lower than desktop?
- Where do sessions end? Which page is the most common last page before an exit without conversion?
- Do returning visitors convert far better than new ones? If yes, your first-visit clarity is weak.
Shortlist five pages. The rest of the audit focuses on them. Auditing 60 pages is why most audits never get finished.
Step 3: Audit navigation and information architecture
Navigation failures are silent. Nobody emails you to say your menu confused them, they just leave.
Do the five second test on your homepage and your top service page: show it to someone unfamiliar for five seconds, close the laptop, and ask what the company does and what they should do next.
Questions to ask:
- Does the main menu use the words my customers use, or internal jargon? (“Solutions” and “Our approach” are jargon.)
- Are there more than seven top-level items? Can any be merged?
- Can a visitor reach the conversion page from any page in one click?
- Does the logo link to the homepage? Does the site have a visible, working search if it has more than 50 pages?
- On a deep page, can the visitor tell where they are (breadcrumbs, active menu state)?
- Are there dead ends: pages with no next step, no internal links, no call to action?
In Clarity, check the heatmap of your header. If a menu item gets almost no clicks but sits in prime position, it is stealing real estate from something that converts.
Step 4: Audit page speed and Core Web Vitals
Speed is a UX problem before it is an SEO problem. Run PageSpeed Insights on one URL per template: homepage, service page, blog article, product page, checkout or contact page.
Focus on the field data section (real users) rather than the lab score. The three thresholds to hit:
| Metric | Good | Usual small business culprit |
|---|---|---|
| LCP (largest contentful paint) | Under 2.5s | Huge uncompressed hero image, slider, video background |
| INP (interaction to next paint) | Under 200ms | Chat widgets, tag manager overload, heavy page builders |
| CLS (cumulative layout shift) | Under 0.1 | Images without dimensions, late-loading banners, cookie bars |
Questions to ask:
- How many third-party scripts load on my key conversion page, and can I name a business reason for each one?
- Is my hero image over 200 KB? Is it served in WebP or AVIF?
- Does the layout jump while loading on a 4G connection? (Test with DevTools network throttling set to Slow 4G.)
- Does the page become usable before it is fully loaded, or does everything appear at once after three seconds?
Step 5: Audit mobile usability with your thumbs
Do not emulate this step. Take your phone, use mobile data instead of wifi, and complete your own conversion path from a Google search.
Questions to ask:
- Can I complete the whole journey with one hand, without zooming?
- Are tap targets at least 44 by 44 pixels, with enough space between them?
- Is the primary call to action visible without scrolling, or reachable via a sticky bar?
- Does the cookie banner or a popup cover the content or the button?
- Do phone numbers trigger a call on tap? Do addresses open maps?
- Does the keyboard show the right type on each form field (numeric for phone, email keyboard for email)?
- Are tables, pricing grids and PDFs readable, or do they force horizontal scrolling?
Cross-check with Search Console’s page experience and mobile reports to see if the issue is sitewide or isolated to one template.
Step 6: Audit above-the-fold clarity and message match
Half of all UX friction is actually copy friction. Visitors are not confused by your buttons, they are unconvinced by your value proposition.
Open Search Console, find the top queries bringing traffic to each shortlisted page, then reread the page as someone who typed that query.
Questions to ask:
- Does the H1 contain the same words as the query that brought people here?
- Within five seconds, is it clear what you sell, who it is for and what to do next?
- Is the first call to action a low-commitment step, or does it jump straight to “Book a 60 minute call”?
- Are the benefits specific (numbers, timeframes, guarantees) or generic (“quality”, “passion”, “tailor-made”)?
- Is pricing, or at least a price range, mentioned anywhere? Missing price information is one of the most common reasons B2B and service visitors bounce.
Step 7: Read the heatmaps and watch 15 session recordings
This is the step non-designers find most revealing. In Microsoft Clarity, filter recordings by your five priority pages, then by sessions that included rage clicks or dead clicks.
Watch 10 to 15 recordings. Do not skip. Take notes with timestamps.
Questions to ask:
- Where do people click that is not clickable? (Dead clicks reveal missing links, expectations of expandable sections, images that look like buttons.)
- Where do they click repeatedly in frustration? (Rage clicks usually mean slow response, broken elements or unclear feedback.)
- At what scroll depth does 50% of the audience stop? Is the main call to action above or below that line?
- Do users scroll rapidly up and down? That is a search behaviour: they cannot find something you have hidden.
- Do they abandon at a specific form field? Note which one.
- Do mobile and desktop behaviours differ so much that the pages need different layouts?
Step 8: Audit forms and checkout field by field
Forms are where the money leaks. Audit each form as a standalone product.
Questions to ask:
- How many fields are there, and can I delete any of them? (Every field you remove typically lifts completion. Ask yourself: will anyone actually use this data this week?)
- Are required fields clearly marked, and is validation shown inline rather than only after submission?
- Do error messages explain how to fix the problem, in plain language, without erasing what was already typed?
- Does the button say what happens next (“Get my quote”) instead of “Submit”?
- Is there a reassurance line near the button: response time, no spam promise, no obligation?
- Is there a visible confirmation after submission, and does it set expectations (“We reply within one business day”)?
- For e-commerce: is guest checkout available? Are shipping costs shown before the final step? Are payment logos visible?
Test the form yourself on mobile, then check your inbox and spam folder. In our experience, roughly one audit in five uncovers a form that has silently stopped delivering emails.
Step 9: Check trust signals and accessibility basics
Small businesses lose conversions not because they are untrustworthy, but because they never prove otherwise.
Questions to ask:
- Are reviews, testimonials or client logos placed next to the decision point, or buried on a separate page?
- Is there a real address, a real phone number and real photos of real people?
- Are legal pages, privacy policy and return conditions easy to find?
- Does text pass a basic contrast check, and can the page be navigated with the keyboard alone (tab through it)?
- Do images have alt text? Do videos avoid autoplay with sound?
- Is the base font at least 16 pixels on mobile?
Step 10: Prioritise every finding by revenue impact
You will finish with 30 to 50 findings. Shipping them in random order is how audits die. Score each one.
Use this simple formula for each finding:
Estimated monthly value = monthly sessions on the affected page x realistic conversion lift x average value per conversion
Then divide by effort. Here is a worked example from a typical service business:
| Finding | Monthly sessions | Assumed lift | Value per lead | Est. monthly value | Effort | Priority |
|---|---|---|---|---|---|---|
| Contact form has 11 fields, drops to 5 | 1 200 | +0.8 pt | 400 | 3 840 | Low | 1 |
| CTA below the fold on mobile service page | 2 500 | +0.3 pt | 400 | 3 000 | Low | 2 |
| LCP 4.8s on blog template | 6 000 | +0.1 pt | 400 | 2 400 | Medium | 3 |
| Menu relabelling and IA restructure | Sitewide | Unknown | 400 | Unclear | High | 7 |
Your estimates will be imperfect. That is fine. The purpose is not accounting accuracy, it is ordering. A rough number beats a gut feeling every time, and it stops the loudest opinion in the room from setting the roadmap.
Split the final list into three buckets:
- This week: low effort, clear value. Copy changes, field removal, button labels, image compression.
- This month: medium effort. Template speed work, mobile layout fixes, trust blocks, form flow redesign.
- Next quarter: high effort or uncertain. Navigation overhaul, checkout rebuild, anything that needs testing before you commit.
Your afternoon schedule
| Time | Task |
|---|---|
| 0:00 to 0:20 | Steps 1 and 2: goal, baseline, five priority pages |
| 0:20 to 1:00 | Steps 3 and 4: navigation review and speed tests |
| 1:00 to 1:40 | Steps 5 and 6: mobile walkthrough and message match |
| 1:40 to 2:40 | Step 7: heatmaps and session recordings |
| 2:40 to 3:20 | Steps 8 and 9: forms, checkout, trust and accessibility |
| 3:20 to 4:00 | Step 10: scoring, prioritisation, three buckets |
Five mistakes that ruin first UX audits
- Auditing the whole site. Five pages carry the revenue. Audit those.
- Confusing taste with evidence. Every finding needs a screenshot, a recording or a number attached.
- Fixing everything at once. If you ship 20 changes in one week, you will never know which one worked.
- Ignoring the mobile and desktop gap. Average numbers hide the fact that one device is carrying the other.
- Never re-auditing. Put the same process in your calendar for two quarters from now. It takes half as long the second time.
How to turn the audit into a repeatable habit
Save your findings spreadsheet as a template. Keep the columns, delete the rows, and re-run the ten steps every six months or after any significant redesign, template change or campaign launch. Track one number over time: conversion rate per key landing page, split by device. If that line goes up quarter after quarter, your audits are working.
Also keep a running “friction log” between audits. Every time a customer says “I could not find” or “the site would not let me”, write it down with the date. By the time you sit down for the next audit, half the work is already done.
When to bring in outside help
Run the first audit yourself. It costs nothing and it will teach you more about your own website than any report. Consider external support when:
- You have found the problems but lack the development resources to fix them.
- The fixes touch structural elements: navigation, templates, checkout logic.
- You need A/B testing to validate a high-stakes change before rolling it out.
- Traffic is significant enough that a 0.5 point conversion lift pays for the work several times over.
At Magnetik, we run this exact process with clients, then handle the implementation and measurement side. If you have completed your own audit and want a second pair of eyes on the priority list, get in touch and send us your findings spreadsheet.
FAQ
How long does a website UX audit take?
A focused self-audit on five priority pages takes about four hours, assuming your analytics and heatmap tools already have one to two weeks of data. A full professional audit covering multiple templates, user testing and accessibility compliance typically takes two to four weeks.
Can I do a UX audit without any design experience?
Yes. The most valuable findings come from data and observation, not design theory. Session recordings, form drop-off points and mobile walkthroughs require attention, not aesthetic training. Bring in a designer for the solutions, not for the diagnosis.
What is the difference between a UX audit and a CRO audit?
A UX audit looks at the whole experience, including pages that have no direct conversion goal. A CRO audit focuses narrowly on the conversion funnel and usually leads directly into A/B testing. In practice they overlap heavily, and prioritising UX findings by revenue impact is what makes an audit actionable.
Are free AI UX audit tools worth using?
They are useful as a starting checklist, and they catch obvious technical and heuristic issues quickly. What they cannot do is understand your business model, your average deal value or why your specific audience hesitates. Use them to accelerate step 3 and step 4, then rely on your own analytics and recordings for the findings that matter. The reasoning is set out in this piece.
How often should I audit my website’s UX?
Every six months for most small businesses, plus after any redesign, platform migration or major campaign launch. E-commerce sites with frequent catalogue and checkout changes benefit from a lighter quarterly review of the funnel alone.
What should a UX audit report contain?
At minimum: the business goal and baseline metrics, the pages audited, each finding with supporting evidence, an estimated impact and effort score, and a prioritised action list split into immediate, short-term and long-term work. A report without prioritisation is a document nobody will act on.