a landing page for one offer.

sending paid traffic to a homepage is the most common and most expensive landing page mistake. a homepage is a directory built for every visitor and every intention; someone arriving from a specific ad has one intention, and the homepage asks them to find it themselves.

a campaign page fixes that by removing choice. fewer options, one argument, one action — which is why it usually converts several times better than the site it lives on.

the economics make this worth the extra build. if a campaign page converts at twice the rate of the homepage, every click you buy is worth twice as much — which is the same effect as halving your cost per click, achieved once and permanently rather than negotiated with an auction every month.

— 01match the ad, word for word.

the first thing on the page should repeat the promise that earned the click, ideally in the same words. if the ad said "same-day repairs in jumeirah", the page headline says that — not "quality service you can trust".

this is not a stylistic preference. the visitor is checking whether they arrived in the right place, and that check takes under two seconds. a mismatch reads as a bait-and-switch even when the offer is identical, because the language changed.

run one page per offer rather than one page for a campaign with four messages. it is more work and it is what makes the matching possible. if you are testing three angles, you need three pages. templating makes this cheap — one page structure, swapped headline, proof and offer — so building three takes an afternoon rather than a project.

and carry the visual through as well. the creative that earned the click should appear on the page, so the transition feels continuous rather than like a jump to a different brand. the same applies to the offer itself — if the ad named a price or a timeframe, that exact number belongs above the fold rather than three sections down.

— 02remove the navigation.

a campaign page should have no menu, no footer links to other services, no blog, no social icons. every one of those is an exit, and you paid for the visit.

this feels wrong to people who think in terms of websites and is obvious to anyone who has measured it. the visitor does not need to explore your company; they need to decide about one thing. the one exception worth allowing is a discreet link to a privacy or terms page, because some ad platforms require it and it does not compete for attention.

keep the page short enough to hold one argument: the promise, what they get, the proof, the price or range, the objections, and the action. anything that does not serve those six is decoration with a cost attached.

length itself is not the enemy. a high-value or unfamiliar offer often needs a long page, because the objections are real and each one needs answering. what kills conversion is length spent on brand narrative rather than on the visitor's decision.

— 03one action, repeated.

one thing to do, appearing two or three times as the page scrolls, always the same action with the same wording. a page offering a call, a download, a whatsapp chat and a newsletter has divided its own conversion four ways.

make the action the smallest reasonable commitment for the intent you bought. cold traffic converts to a low-friction action — a question answered on whatsapp, a price sent — far better than to a booked meeting. warm traffic can take the meeting.

and test the destination on a phone, on mobile data, as a stranger. the number of campaign pages we audit where the button opens a broken form or a desktop-only booking tool is genuinely high, and it is invisible to whoever built it on a laptop. add the whatsapp option explicitly for gulf audiences — for many visitors it is the only action they will take, and a form-only page quietly loses them.

— 04the things worth testing, in order.

the offer first — a different proposition beats a different button colour by an order of magnitude. then the headline, then the form length, then the proof, then the layout. most teams test in exactly the reverse order.

run one test at a time for at least a fortnight, and keep a record of what you learned. a page tested properly three times a year outperforms one redesigned twice, because you can attribute the improvement to something. keep the losing versions too; knowing which offers failed is as useful next year as knowing which one won.

and remember the page is only half the equation. a campaign page converting poorly with the right offer is a page problem; converting poorly with the wrong audience is a targeting problem, and the fix is upstream. our own service pages are written the same way for the same reason: one argument, one action.