Landing page — what it consists of and what actually decides its outcome

A landing page is one enquiry and one action. No menu, no second offer, no way out other than the one you want. That constraint is the entire definition, and nearly every mistake in this work is a violation of it.

What it is, and what it is not

It is a page built for one campaign and one intent, that a visitor reaches from an advertisement already knowing roughly what they came for. Its job is to confirm they are in the right place and to make one action easy.

It is not a new offer page with a nicer header. An offer page carries the site's full navigation, lives for years and answers many questions; a landing page gives the visitor one way forward and dies with the campaign. Calling the first one a landing page does not make it behave like one.

It is not a one-pager either. A one-pager is an entire website compressed onto one screen, with a menu that scrolls between sections. It is measured like an ordinary site, because that is what it is.

When it makes sense, and when it does harm

A landing page makes sense when you are paying for the traffic and you know what the visitor typed to get there. Then narrowing the page to that one intent is what the money buys.

It does harm in two situations. First, when it replaces a page that was doing several jobs — you have swapped a page answering five questions for one answering one. Second, when there is no campaign: a landing page with no traffic source is a page with no menu that nobody can find.

If you are not buying traffic, what you almost certainly want is an offer page inside the site, not a landing page beside it.

There is a third situation, less obvious and more common than either: a landing page built for a campaign that does not yet exist. It happens when the page is commissioned first and the advertising is planned afterwards, usually because the page felt like the concrete part of the job. What follows is a page nobody can reach, written against an intent nobody has defined, and a campaign later shaped to fit a page rather than the other way round. The order that works is the reverse — decide what you are buying, then write the page that answers it.

A useful rule of thumb for the decision: if you cannot say in one sentence what the visitor typed before they arrived, you are not ready to build the page. That sentence is the brief. Everything on the page is either an answer to it or a distraction from it.

Anatomy: what has to be visible without scrolling

The headline

It names what the visitor gets, in their words rather than yours. The test is mechanical: if a competitor could put the same sentence on their page without changing anything, it says nothing. "Comprehensive solutions for business" passes that test on every page in the industry, which is why it fails as a headline.

The second test is harder and more useful: the headline should repeat, nearly word for word, the promise in the advertisement that brought the visitor here. Not because repetition is elegant, but because the half-second in which someone decides whether they are in the right place is spent looking for the thing they just clicked. A headline that paraphrases the advertisement creatively costs you that half-second, and the reader spends it leaving.

The qualifying line

Who this is for and what it covers. One sentence. Its job is to let the wrong visitor leave quickly, which is a feature — a page that keeps everyone keeps the people who will never buy.

Proof

A price, a number, a photograph of actual work, a named client. Anything a sceptical reader can check. This is the element most often replaced by adjectives, and adjectives are the cheapest thing on the internet.

One call to action

One action. A second option is not a choice, it is a deferral — and a deferral on a page you paid to have visited is a loss. If you genuinely have two audiences, that is two landing pages, not one page with two buttons.

The form

The form is where the page either collects or leaks. Three rules, in order of how much they cost when broken.

As few fields as the next conversation needs. Name, a way to reply, and one sentence of context is usually enough. Every additional field is one more reason to postpone.

No field whose only purpose is to qualify the lead. Budget ranges and drop-downs of how they heard about you are questions you can ask in the reply. Asked before the reply, they filter out people who would have answered them happily afterwards.

Say what happens next. "We reply within one working day" belongs beside the button. Its absence is a silence the visitor fills with their own worst guess.

One further thing about forms that gets discovered late: what happens after submission is part of the page. A confirmation that says only "thank you" wastes the one moment when attention is guaranteed. That screen is where you say what happens next, when, and from whom — and, if there is anything worth reading while they wait, where it is. It is also the only reliable place to count a conversion, which is the second reason not to leave it as an afterthought.

What not to put on a landing page

A menu leading off the page. A second, competing offer. A carousel of rotating slogans — nobody waits for slide three. And anything with its own "more" button, because every one of them is an exit you paid to create.

There is one more, subtler than the rest: social proof that proves something else. Logos of clients from a different sector, a testimonial about a service you are not advertising here. Proof that does not match the intent reads as decoration, and decoration is what the reader was already suspicious of.

Three constructions for three situations

The short page — one screen and a form. For a single, well-understood offer where the visitor already knows what they want. The cheapest to build and the fastest to test.

The long page — the offer, the objections, the proof, the action repeated. For a considered purchase, where the visitor arrives interested but not convinced. It is longer because the objections are real, not because length helps.

The page with a calculator or a configurator — where the answer depends on the visitor's own numbers. The most expensive of the three and the only one that earns its cost by giving the visitor something before asking for anything.

Which of the three to build is decided by one question: how much does the visitor already know when they arrive? Traffic from a branded search or a retargeting campaign knows a great deal and needs the short page. Traffic from a broad phrase knows little and will not act until the objections are answered, which is the long page. And traffic that cannot act because the price genuinely depends on their situation needs the calculator, not more persuasion.

Getting this wrong is expensive in a way that does not show up in the page: a short page under broad traffic looks like a conversion problem, and the usual response is to rewrite the button.

A one-pager is a different thing

The three are confused constantly, and the confusion is expensive because it changes what you measure. A landing page is measured by one conversion. An offer page is measured by traffic and contact over time. A one-pager is measured like a website, because it is one.

The commonest error is calling a new offer page a landing page. This is not a naming question — a page carrying the full menu stops being a landing page, because it gives the visitor ten ways out instead of one.

Speed — and the one place where it enters the auction

Load time matters on a landing page for a reason it does not matter elsewhere: in paid search, the quality of the landing experience feeds into what you pay per click. That makes speed a budget question rather than a comfort question.

But it is worth being precise about how much of the problem it usually is, because "make it faster" is the advice most often given and least often the answer.

Two things about landing pages make speed easier to fix here than elsewhere. The page is small, so there is little to optimise. And it is usually new, so it has not accumulated the layers of scripts that slow an older site. Which means that if a landing page is slow, the cause is normally one identifiable thing — an oversized hero image, a chat widget, a tag manager loading everything at once — rather than a general condition requiring a project.

It also means that when speed is already good and the results are still poor, no further engineering will help. That is the situation we measured on our own pages.

Our own campaign landing pages — and the conclusion that surprised us

We measured our own. Every page has its own brief and its own intent — no shared clones. Of those built and being served, the ones with field data collected on mobile all show load time in the good band.

And the landing page experience across the whole ad account still rates below average.

The conclusion is uncomfortable and worth publishing: speed was not the cause. This corrects a note in our own earlier report, which pointed at mobile load time as a measurable cause — the measurement does not support it. What is left as the likely cause is the match between what the advertisement promises and what the page says, which is a writing problem rather than an engineering one.

What the work on a landing page consists of

Four parts, and only one of them is building.

The intent. What the visitor typed and what they expect. This comes from the campaign, not from the page, and getting it wrong makes everything downstream irrelevant.

The text. The headline, the qualifying line, the objections. This is most of the work and most of the value.

The proof. Gathering it is usually the slowest part, because it means asking clients for permission to use their names.

The build and the measurement. Including deciding, before launch, what counts as a conversion. A page with no defined conversion cannot be judged, only admired.

The proportions surprise people who have not commissioned one before: the build is typically the smallest part. A landing page is a day or two of construction sitting on top of a week of deciding what it should say — and the week is not compressible, because it is where the intent, the objections and the proof are settled. A quote that is mostly build time is a quote for a template into which somebody else's words will be poured.

How to tell whether it works

One conversion, counted. Not visits, not time on page. A landing page has one job and it is countable.

Compare it against the page the campaign used before. Absolute conversion rates vary too much between offers to mean much on their own; the comparison is what carries information.

Read the enquiries, not just the count. Ten enquiries from people who misunderstood the offer is a worse outcome than four from people who did not, and only reading them shows the difference.

Our wider treatment of conversion measurement sits in conversion rate.

What a landing page will not fix

It will not fix an offer people were not buying. It will not fix a campaign aimed at the wrong intent — a perfect page under the wrong advertisement converts worse than a mediocre page under the right one. And it will not build anything durable: when the campaign stops, the page stops, because it has no traffic of its own and, by design, nothing linking to it.

That last point is the honest limit of the whole format. A landing page rents attention. If you want attention that accumulates, the answer is in marketing and SEO, and it is slower.

Five mistakes we see most often

• The page and the advertisement describe different things. The single most common cause of a landing page underperforming, and the one least often investigated, because both halves look fine on their own.
• Two buttons for two audiences. This is two pages. A choice on a landing page is a deferral, and a deferral is a loss on traffic you paid for.
• Proof from a different sector. Client logos that do not match the intent read as decoration, and decoration is what a sceptical reader was already expecting.
• No defined conversion before launch. A page you cannot judge will be judged on impression instead, and impression always favours the prettier version.
• The page outliving the campaign. A landing page left up with no traffic source is a page with no menu that nobody can find — and, if search does find it, a thin page competing with your own offer pages.

The pattern behind four of those five is the same: each comes from treating the page as the deliverable rather than as one half of a pair. The advertisement and the page are a single object that happens to be built by two different people, and most of the failure lives in the seam between them.

What this article deliberately leaves out

• Which type of site you need — that decision comes first.
• The smallest sensible site — brochure website.
• Whether a site pays for itself — the arithmetic.
• Running the campaign itself — a different discipline and a different budget.