Landing pages that convert: structure, order, common mistakes
A landing page has exactly one job: to make a decision possible. Anything that does not prepare that decision makes it less likely — however well made it is on its own terms. That is why cutting almost always helps.
In short
- Three things belong at the very top: what it is, who it is for, and what happens next. All visible at a glance, without scrolling.
- The order of sections follows the order of objections — not the logic of your offering.
- A landing page has one call to action, repeated. Two different goals halve both.
- The five most expensive mistakes are almost always the same: too many ways out, too many form fields, benefit stated from the vendor's view, no evidence, and no answer to "what does it cost".
The most common reason a landing page does not work is not poor craft. It is that it was built like a home page — as an overview of everything the company can do.
A home page answers "who are you". A landing page answers "should I do this". Those are different structures.
What belongs at the very top
The area visible without scrolling decides the rest. It needs three pieces of information — and no fourth.
1. What it is
In one sentence, no jargon, no wordplay. Anyone who does not know what this is about after that sentence will not read the second.
Test: would someone from your audience who does not know your company understand it in four seconds?
2. Who it is for
The exclusion matters more than the invitation. "For trades businesses with 5 to 50 staff" works better than "for companies of any size" — because the reader recognises themselves.
Test: is it apparent who is not meant?
3. What happens next
Not "get started", but what actually follows: "fill in the form, call back within one working day". Uncertainty about the next step is one of the biggest barriers there is.
Test: does the reader know what happens in the next 24 hours?
The order follows the objections
This is where most pages make the decisive mistake: they are built along the logic of the offering — features first, then benefits, then price, then contact. But the reader does not think in that order.
They think in objections, and always in the same sequence:
| Order | The objection | The section that clears it |
|---|---|---|
| 1 | "Am I in the right place?" | the header with the three pieces of information |
| 2 | "Does this solve my problem?" | the problem in their words, then the solution |
| 3 | "How does it actually work?" | three steps, not a feature list |
| 4 | "Is what you claim true?" | evidence: figures, a named reference, an example |
| 5 | "What does it cost?" | a price or an honest range |
| 6 | "What if it does not fit?" | risk reduction: cancellation, trial, data export |
| 7 | "And if I have another question?" | a questions section with the real objections |
Worth knowing
The questions section at the end of a landing page is rarely the weakest part — often it is the most read. It is the only place written in the language of doubt rather than the language of selling.
That is why it pays to answer the uncomfortable questions there: "who is this not suitable for?", "what happens to my data if I cancel?". One honestly answered uncomfortable question convinces more than three arguments.
The call to action
Three rules that hold up in nearly every test.
One goal, not two. "Book a demo" and "subscribe to the newsletter" on the same page means both perform worse. Someone who cannot decide decides on nothing.
Repeat it, worded identically. After the header, after the evidence, at the end. Different wording for the same goal reads as different goals.
Name the result, not the action. "Request a free assessment" is more concrete than "submit" and more honest than "start your journey".
The most reliable single improvement is shortening the form. Every additional required field costs completions — disproportionately so for the fields that feel like an assessment: revenue, headcount, budget.
The usual justification is that those details are needed for qualification. In most cases that is not true: what is actually needed is a name, a company and a way to get in touch. Everything else can be clarified in the first conversation — where, in any case, people answer it more willingly than in a form.
The five most expensive mistakes
1. The full navigation in the header. Every menu item is an exit. On a landing page, at most the logo belongs there — and even that unlinked if the page is reached from an ad.
2. Too many form fields. Three to four is nearly always enough. Every further one has to justify itself, not the other way round.
3. Benefit from the vendor's point of view. "Modular", "scalable", "holistic" describe the product. The reader is looking for what will be different for them.
4. No evidence. Claims without a figure, without a name, without an example. One reference with company and role weighs more than five anonymous quotes.
5. No answer to the price question. If you cannot or will not state a price, at least state the range or the method. A page with no indication at all produces exactly the assumption you were trying to avoid.
Testing with little traffic
Conventional A/B tests need numbers small companies do not have. At 200 visitors a month, a meaningful test takes longer than the offer exists.
Three approaches that work even with little traffic:
- Watch five people. From your audience, without explanation, with a task. Wherever they hesitate is the fault. Five people surface most of the coarse problems.
- The four-second question. Show the page, close it after four seconds, ask: what was it about, and who is it for? If they cannot answer, they did not understand the header — and the header is then wrong.
- Measure where scrolling stops. The section before that point is either superfluous, or an objection is unanswered there.
Review the text of my landing page. Be strict, praise nothing. Audience: [who exactly] Goal of the page: [one single action] Page text: [paste text, section by section] Tasks: 1. From the audience's point of view, answer after the first four seconds: What is this? Who is it for? What happens next? If an answer is missing, say so plainly. 2. Assign each section to one of these seven objections: am I in the right place / does it solve my problem / how does it work / is it true / what does it cost / what if it does not fit / open questions. Name sections that fit none, and objections that go unanswered. 3. Mark every phrase that describes the product rather than the result for the customer. Rewrite them. 4. Name every claim without evidence. 5. Name the three changes with the greatest effect, sorted by effort. Do not invent customer quotes or figures.
In closing
A landing page rarely improves by adding something. It improves when the order follows the objections and everything that belongs to no objection falls away.
The three pieces of information in the header, seven sections in the order of the doubts, one goal, three to four form fields. Start there and you have gained more than from any design round.
Common questions
What belongs at the very top of a landing page?
Three pieces of information, visible without scrolling: what the offering is, who it is for, and what happens next if you act. The third is the most frequently forgotten — uncertainty about the next step puts more people off than a high price does.
How long should a landing page be?
As long as it takes to answer all of your audience's objections — and not one paragraph longer. For simple, low-priced offerings that is often three sections; for ones needing explanation, seven. The length follows from the objections, not from a rule of thumb.
How many form fields make sense?
Three to four: name, company, a way to get in touch, and at most one more. Every additional required field costs completions, especially details that feel like an assessment — revenue, headcount, budget. Those are easier to clarify in the first conversation.
Should the navigation appear on a landing page?
No. Every menu item is an exit from a page meant to prepare exactly one decision. At most the logo makes sense — and on pages reached from a paid ad, even that without a link.
How do you test a landing page with few visitors?
Not with A/B tests — there are not enough numbers. Three approaches work: watch five people from your audience using it, ask the four-second question about content and audience, and measure at which point scrolling stops.
Marketing that sets itself up
The Studio Engine beta is live. Claim your spot and help shape it from the start.
Join the beta →