Most sites are judged before anyone reads a word. The failure is rarely the design; it is what the page asks a visitor to do in the first three seconds.
A site can be beautiful, fast, and still lose the sale. In review after review, the moment a project goes wrong is not the moment someone dislikes the layout; it is the moment they cannot tell, within a breath, what this company does, who it is for, and what happens next. Everything after that is a rescue operation.
A visitor arriving from an ad, a search result or a referral is running one test: is this for me? They answer it from the headline, the first image and the shape of the page, not from your about section. If the answer is unclear they do not scroll to find it. They leave, and the analytics record it as a bounce rather than as the positioning problem it actually is.
SEO earns you a position. GEO earns you a mention. Increasingly, the mention is the whole game.
The headline describes the company, not the customer. “A creative agency built on craft” tells a visitor nothing they can act on; “Brand and web for hospitality groups in the Gulf” tells them whether to keep reading.
The page opens with atmosphere instead of an offer. A full-screen video with no proposition costs you the visitors who were closest to buying: the ones who arrived with intent and needed one sentence of confirmation.
The primary action competes with four others. When a hero carries a book-a-call button, a newsletter field, a scroll cue and a menu that opens over everything, none of them gets pressed.
The fix is not more copy. It is deciding what the first screen owes the visitor and cutting everything that does not serve it. Four things, in this order:
A claim you can disagree with
Specific enough that a wrong-fit visitor self-selects out. Vagueness is not safety; it is a slower no.
Proof within reach
One client name, one number or one recognisable logo, close enough to the headline to be read in the same glance.
One action
A single primary button with a verb and an outcome. Everything else is secondary styling, not equal weight.
Evidence of a real company
A location, a team, a phone number. Trust signals cost nothing and remove the quiet objection nobody tells you about.
Performance work is real and worth doing; Core Web Vitals affect both ranking and conversion. But a page that loads in 800ms and says nothing loses to a page that loads in two seconds and says the right thing. Fix the message, then fix the milliseconds. Doing it the other way round produces a very fast page that still does not sell.
Show the first screen to five strangers for five seconds:
then ask what the company does and who it is for. If three cannot answer, the headline is the problem.
Count the actions above the fold:
more than one primary button means you have not chosen.
Read the headline out loud to a client:
if it would sound odd said in a meeting, it will read as noise on the page.
Check the entry pages, not the homepage:
most paid traffic lands deeper in the site, where the same three-second test applies and is usually failed harder.
Watch ten session recordings:
the moment people scroll fast and leave tells you exactly which line lost them.
None of this requires a redesign. It requires deciding what the page is for and being willing to say something specific enough to be wrong. That is usually a week of work, not a quarter.
We will review your homepage and top two landing pages and tell you what a visitor actually understands in three seconds; free, in one 30-minute call.
Book a free call ↗Search intent covered
This article covers “why my website isn’t converting”, “above the fold best practice” and “homepage headline examples”: the questions clients ask before commissioning a redesign.
Reader questions
Why a good-looking site can still fail to convert.
First-screen anatomy
The four things a hero has to carry, in order.
A usable audit
Five checks you can run on your own site this week.
Most websites lose visitors in the first three seconds, before design or speed matter. Say what you do and who it is for, show one piece of proof, offer one action; then optimise everything else.
