Here is a test you can run in the next two minutes. Open your best-performing ad, click it as a stranger would, and then close the tab after one second.
Write down what you saw. Then ask whether a person who had never heard of you would know, from that single second, what you sell, who it is for, and what to do next.
On most homepages the answer is no, and that is not a criticism of the homepage. It is a description of what a homepage is for.
The two jobs are not the same job
A homepage serves people who already arrived with intent. They typed your name, or clicked a link from an email, or came back for the third time. They need navigation: a route to the thing they were looking for. So a homepage offers everything — ranges, the about page, the blog, the contact details, the careers link, a newsletter prompt and a chat widget.
Paid traffic is the opposite. Someone saw one claim, in one ad, and clicked because that one claim interested them. They have no relationship with you, no patience, and a thumb hovering over the back button. Giving that person eleven options is not generosity, it is work, and work is what they leave to avoid.
The gap is usually visible in the data before anyone names it. Look at your ad platform's reported landing page views against link clicks: when those two numbers are far apart, people are leaving before the page even renders, which is a speed problem. Then look at session duration for paid traffic against organic. Paid traffic that bounces much faster is almost never a traffic quality problem. It is a page problem.
What a page for cold traffic actually needs
Not a long page. A page where every element is doing one job.
The same promise as the ad, in the same words. This is the single biggest lever and the one most often missed. If the ad said twenty-four-hour delivery, the page says twenty-four-hour delivery above the fold, in those words. Scent matching, as it is sometimes called, costs nothing and routinely moves conversion by a double-digit percentage. When the ad and the page disagree, the visitor assumes they clicked the wrong thing.
One action. One primary button, repeated down the page, with everything else subordinate. Cutting the global navigation from a paid landing page feels wrong the first time and usually improves conversion, because every nav link is an exit someone paid for.
Proof near the claim, not in a carousel at the bottom. A review, a named customer, a number, a photograph of the real thing. Proof placed next to the claim it supports does far more than the same proof collected into a testimonials section nobody scrolls to.
The objection answered on the page. Whatever your sales team hears most — price, delivery, contract length, whether it works with the thing they already own — belongs on the page. Every unanswered objection becomes a tab the visitor opens to check, and most of those tabs never come back.
A form that asks for what you need and nothing else. Each additional field costs you completions. If the extra information genuinely improves lead quality, keep the field and accept fewer leads. If it is there because a form plugin suggested it, remove it.
Speed, on a phone, on an average connection. A page that takes six seconds has lost a large share of the people you bought before they saw a word of this. The same gap shows up on store pages between desktop and mobile, and it is always worse than anyone expects.
When the homepage is genuinely fine
I am not arguing that every campaign needs a bespoke page. Three cases where the homepage is the right destination:
Brand search. Someone typing your company name already knows you. Send them to the homepage and let them navigate.
Single-product businesses. If you sell one thing and the homepage is already a landing page for that thing, building a second version mostly creates maintenance work.
Low spend and no capacity. If you are spending a few hundred a month and cannot maintain extra pages, a sharpened homepage beats a neglected landing page. Tighten the top of the homepage instead: lead with the specific promise, make one action obvious, move proof up.
The cases where a dedicated page nearly always pays are the opposite: several distinct audiences, several offers, a considered purchase, or a high cost per click. The more you pay for the click, the more the page is worth.
Test it properly or do not bother
Most landing page tests I am shown prove nothing, because they were run for four days on a hundred and fifty visitors and the winner was noise.
Change one meaningful thing at a time. Headline, offer framing, form length, page structure. Changing six things and winning tells you only that one of the six helped, and you will never know which.
Run until you have enough conversions to be confident, not enough visitors to be bored. A few hundred conversions per variant is a reasonable bar. If you cannot reach that in a month, you do not have the traffic for split testing — in that case, make the change you have good reason to believe in and judge it on a longer before-and-after, accepting that it is weaker evidence.
Keep the traffic source constant. A test contaminated by a new campaign launching mid-way is not a test.
And measure the right outcome. Form fills are easy to improve by lowering quality, and a page that doubles leads while halving close rate has cost you money. Measure through to the sale wherever your tracking allows — that is the same failure I described in why cheap Google Ads leads are often useless.
Where the words come from
The best landing page copy is not written in a copywriting session. It comes from your sales calls and your support inbox: the phrases customers use, the objection that comes up every time, the thing people are relieved to hear.
If the page sounds like your industry rather than like your best salesperson, it will read as interchangeable with the cheaper option — which is the mechanism behind a premium brand that cannot charge premium prices.
The short version
A homepage is a map. A landing page is an answer. Cold traffic needs an answer, in the same words the ad used, with one action, proof beside the claim, the main objection handled, and a page that loads on a phone.
If you are about to put real budget behind anything — a seasonal push, a new product, a new market — build the page first and send traffic through it before the stakes rise. The five weeks before Black Friday is exactly the window for that kind of work, and the worst possible time to attempt it is the night before.
We build these as part of our web development work, and the ad account audit covers the page alongside the campaigns, because judging one without the other tells you very little. If you would rather start smaller, the free growth audit will tell you in writing whether the page or the account is costing you more.