Speed and stability first (checks 1–5)
Nothing else in UX matters if the page hasn't loaded. Test: mobile load time on a real phone over 4G, not office wifi (1); Core Web Vitals in PageSpeed Insights — LCP under 2.5s, CLS under 0.1 (2); whether anything shifts around as the page loads (3); image weight — a single 4MB hero photo can double load time (4); and whether tapping anything during load does what the visitor expected (5).
These five checks explain more abandoned visits than everything else combined. Google's data has been consistent for years: every extra second of mobile load time sharply increases bounce probability. Speed isn't a technical nicety — it's the first UX feature.
Mobile reality check (checks 6–10)
Over 60% of your visitors are on a phone, so audit there first. Check: tap targets big enough for thumbs, with space between them (6); text readable without zooming — 16px minimum body copy (7); no horizontal scroll or content cut off at the edges (8); sticky elements (cookie bars, chat widgets, banners) not burying the content (9); and the primary action — call, quote, buy — reachable within one thumb-scroll of landing (10).
The classic mobile failure is a desktop site squeezed smaller: menus that need surgeon-level precision, forms that trigger the wrong keyboard, phone numbers that aren't tap-to-call. Each is a small tax; together they're why mobile conversion rates lag desktop on badly built sites.
Navigation and message (checks 11–15)
Five-second test: show a stranger the homepage for five seconds — can they say what you do, who it's for, and what to do next (11)? Then: is the primary call-to-action visible without scrolling (12); can a visitor reach any key page in two clicks (13); do menu labels use customer words, not internal jargon (14); and does every page have an obvious next step, or do dead-end pages strand people (15)?
Message clarity beats cleverness everywhere it's tested. 'Hand-coded websites from £299, live in 14 days' outperforms 'Digital experiences, reimagined' because visitors decide in seconds whether they're in the right place — and a confused visitor doesn't ask for clarification, they leave.
Forms and trust (checks 16–21)
Forms are where conversions die. Count the fields — every field beyond four measurably cuts completion (16); check error messages say what's wrong and how to fix it (17); confirm the submit button says what happens next ('Get my quote', not 'Submit') (18); and test the whole thing on a phone (19). Then trust: are real reviews, credentials or live work visible near the ask (20), and are contact details — a phone number, a real address or registration number — findable (21)?
Run all 21 and score honestly: pass, partial, fail. Most sites fail 6–10 of them, and fixing those is usually worth more than any redesign. If you'd rather have professional eyes on it, a structured audit with prioritised fixes costs far less than the leads the leaks are costing you.
Get a professional UX design audit from £399 →
Frequently asked questions
How is a UX audit different from a redesign?expand_more
An audit diagnoses; a redesign rebuilds. The audit tests your site against evidence-based checks, then ranks the problems by impact. Often the top fixes are small — a form, a headline, image compression — and you may not need the redesign at all. Audit first, spend second.
Can I do a UX audit myself?expand_more
This checklist gets you a long way, and free tools (PageSpeed Insights, Search Console) supply the data. What a professional adds is prioritisation — knowing which of the 12 failures is costing you leads and which is cosmetic — plus patterns from having seen hundreds of sites.
How often should I audit my website?expand_more
A full audit yearly, or after any major change (new pages, new branding, replatforming). A light quarterly check on speed and forms catches regressions — sites decay as plugins update, images pile up and content gets bolted on.
What does a professional UX audit cost?expand_more
In the UK market, £300–£1,500 depending on site size and depth. Ours starts at £399 and produces a prioritised fix list you can action with any developer — the deliverable is the roadmap, not a sales pitch for a rebuild.