Above-the-Fold UX Audit: 7 Friction Points Killing Your Landing Page Conversion Rate
Visitors decide in about five seconds whether a page is relevant. Seven specific things determine that decision, and most of them are fixable in an afternoon.
Conversion optimisation gets sold as a testing programme — run experiments, accumulate learnings, iterate. That is the right answer when a page is fundamentally sound. It is the wrong answer when the page has obvious structural problems, and most underperforming landing pages have three or four of them.
Find and fix those first. They are cheaper than a test programme and they move the numbers more.
1. The visitor cannot tell what this is in five seconds
The test: show the page to someone unfamiliar with the business for five seconds, hide it, and ask what the company does and who it is for. If they cannot answer, no amount of button colour testing will help.
The usual failure is a headline written for the company rather than the visitor — "Transforming digital experiences through innovation" says nothing. The replacement names the outcome and the audience: "Audit any landing page in 60 seconds" is unglamorous and immediately understood.
2. The primary action competes with four others
A header with Login, Sign up, Pricing, Docs, Contact and Start free trial gives the visitor six decisions. Each one splits attention, and the important one loses.
On a campaign landing page, strip navigation to the logo and one action. On a homepage you need navigation, but only one element should carry primary visual weight. If two buttons have the same colour and size, you have two primary actions and therefore none.
- Exactly one solid, high-contrast button above the fold.
- Secondary actions as text links or outlined buttons — visibly subordinate.
- No competing offer in the hero. A "Book a demo" next to a "Start free" halves both.
- Remove the campaign page from the main navigation, and remove the main navigation from the campaign page.
3. The CTA does not say what happens next
"Submit", "Get started" and "Continue" describe your system. "Get my free audit", "See pricing", "Book a 20-minute call" describe the visitor's next moment. The second kind consistently performs better because it reduces the uncertainty that causes hesitation.
Pair it with a risk-reducing line directly under the button — "No credit card", "Takes 60 seconds", "No signup needed to see your score". This line does more work than the button text itself, and it is a ten-minute change.
4. There is no reason to believe you
Trust signals belong in the immediate viewport, not in a section halfway down that most visitors never reach. What works, in rough order of strength:
| Signal | Strength | Note |
|---|---|---|
| Named customer logos | High | Only with permission, only real customers |
| Specific numbers | High | "108 checks across 12 modules" beats "comprehensive" |
| A named testimonial with role and company | High | Anonymous quotes read as invented |
| Security or compliance badges | Medium | Only if genuinely held |
| Review scores from a platform | Medium | Must be verifiable |
| "Trusted by thousands" | None | Unverifiable claims reduce trust |
5. The form asks for more than the offer is worth
Every field is a small negotiation. Nine fields to download a PDF is a bad trade and visitors recognise it. Match the ask to the value: an email for a checklist, name and email for a newsletter, full qualification only when someone is requesting a quote or a call.
Where you genuinely need the data, get it later. A short form followed by a qualification step on the thank-you page captures the lead first and the detail second, so a drop-off costs you the detail rather than the lead.
- Cut every field that does not change what happens next.
- Ask for a phone number only when you will actually call.
- Never ask for information you can derive — company size from the email domain, city from the PIN code.
- One column. Side-by-side fields break badly on phones.
6. Microcopy contradicts the offer
A headline promising "Free forever" above a button saying "Start trial" creates a question the visitor resolves by leaving. So does "No credit card required" next to a form asking for billing details, or "Instant results" above a "We will get back to you within 48 hours" confirmation.
Read the hero out loud as a single paragraph — headline, subhead, button, supporting line, form labels. Contradictions are obvious when spoken and invisible when scanned in a design tool.
7. The page moves while they are reading it
This is a UX problem that presents as a performance metric. Images without reserved dimensions, late-loading banners, fonts swapping at different metrics — each one shifts the layout, and on a phone that means the button is not where the thumb is going.
It is measured as Cumulative Layout Shift and it is worth treating as a conversion bug rather than a technical one. The Core Web Vitals article covers the specific fixes, which are mostly setting width and height on images and reserving space for anything injected.
The five-minute audit
On a phone, on mobile data, in a private window:
- Five-second test — can a stranger say what this is and who it is for?
- Count the actions above the fold. More than two is too many.
- Read the button. Does it describe the visitor's next moment?
- Is there one specific, verifiable trust signal in the first screen?
- Count form fields. Is each one worth the offer?
- Read the hero aloud. Any contradictions?
- Reload. Does anything jump after the first paint?
Seven questions, and they will usually surface more addressable conversion loss than a month of A/B tests on a page with structural problems. Run the tests after these are fixed, when you are optimising rather than repairing.
The Conversion & UX module checks CTA clarity, heading structure, trust-signal presence, form field count and layout stability on any live URL.
Audit my landing pageThen, and only then, test
Once the structural issues are gone, testing earns its place — and it needs enough traffic to be meaningful. Below roughly a few hundred conversions a month, most A/B tests will not reach significance before the business context changes underneath them, and you will make decisions from noise.
At that volume, qualitative work pays better: five user sessions recorded on a phone, or five people asked the five-second question. Cheaper, faster, and it tells you why rather than which.