A landing page has one job. Everything on it either helps a visitor decide to do that one thing or takes attention away from it, and most pages that underperform are not badly written — they are doing four jobs at once and doing none of them clearly.
The single-action test is the most useful constraint in this kind of copywriting, because it is unforgiving. It forces every section to justify its presence, and it usually removes a third of the page without losing anything a visitor wanted.
The one-action test
Write down the one thing you want a visitor to do, as a verb and an object: start a trial, book a call, download the guide, add an item to a cart. If you cannot write it in those terms, the page is not ready to be written, because every other decision depends on that answer.
Then apply the test to every element. The navigation, the second call to action, the newsletter box, the case study link — each one either supports the action or competes with it. Secondary actions are not forbidden, but they belong below the point where the primary decision is made, and there should be at most one.
The exception worth keeping: a page can carry two actions for different readiness levels, such as "start a trial" and "see pricing." That works when the second is a step towards the first rather than an alternative to it. It stops working the moment the two lead somewhere unrelated.
What the top of the page has to do
Whatever is visible without scrolling has three jobs, in this order: say what this is, say who it is for, and give a reason to act. Most pages handle the first and skip the second, which is why they attract visitors who were never going to convert.
The audience statement is the most commonly missing element and the cheapest to add. Naming who the page is for does two things at once: it tells the right visitor they are in the right place, and it tells the wrong one to leave, which is a service rather than a loss. A page that tries to include everyone converts a small percentage of a large, mismatched audience; a page that states its audience converts a larger percentage of a smaller, correct one.
Keep the top of the page mostly text. A hero animation that takes four seconds to communicate what the headline says in two is a net loss for every visitor except the ones who were already convinced.
Writing the headline
The headline's job is to be understood, not to be admired. That rules out cleverness that needs a second reading, wordplay that requires the rest of the page, and anything that could plausibly describe a competitor.
Two structures carry most of the load. The outcome structure names the result and the audience: what the visitor gets, and for whom. The problem structure names the situation the visitor recognises: what is going wrong, and what changes. Outcome headlines read as confident; problem headlines read as attentive. Which one fits depends on whether your visitor is shopping or stuck.
Write five versions and pick by one criterion: which one a stranger could repeat correctly after reading it once. That is a harsh test, and the versions that survive it tend to be plainer than the ones you were initially pleased with.
The subhead does the work
The subhead is where the specifics go, because the headline cannot hold them. Three things belong there: what exactly the thing is, what it requires from the visitor, and what happens next.
That second item is the most frequently omitted. Stating what the visitor needs — a work email, an existing account, thirty minutes — filters out people who cannot complete the action, which prevents a click that would have become a bounce. It costs a few conversions in the short term and improves the ratio that matters, which is sign-ups that stick.
Keep the subhead to two or three sentences. It is a support structure, not a second headline competition; if it needs more than that, the offer itself is unclear and no amount of subhead will fix it.
Proof without invention
Proof is where landing pages most often become dishonest, usually without intending to. Invented percentages, unattributed quotes and vague claims about being trusted by thousands are all easy to write and all corrosive when a visitor checks.
Use the three kinds of proof you can actually produce. Specific numbers you own and can explain — how long something takes, how many items it handles, how many of your own customers do a thing. Named examples with a description of the situation rather than a testimonial. And demonstrations: a screenshot of the actual interface, a sample output, a short recording of the thing working.
The demonstration is the strongest and the most underused, because it is the only one the visitor can evaluate themselves. If you have a product, showing it doing the single thing the page promises does more than any sentence about it.
Handling objections in place
Objections should be answered close to where they arise, not collected into a FAQ at the bottom. The visitor who doubts the price is at the pricing section; a reassurance three screens later arrives after they have left.
Three objections cover most pages: it will not work for my case, it is too complicated, and I cannot justify the cost. Each has a short answer that belongs in the relevant section — a fit statement near the top, a description of the setup next to the feature list, and a clear statement of what the price includes next to the number.
Answer them plainly. A reassurance that hedges reads as a confirmation of the doubt, and visitors are better at detecting evasive language than most writers assume. If the honest answer is "it is not for you," say that; it is a better outcome than a sign-up that churns.
The call to action and its surroundings
The button text should name the action and, where possible, the result: not "Submit" or "Learn more" but "Start writing" or "Get the checklist." First person sometimes helps and sometimes does not; the deciding factor is whether the visitor is doing something for themselves or requesting something from you.
What sits immediately above and below the button matters more than its colour. Above it, a line removing the last friction — what it costs, that no card is needed, how long it takes. Below it, a line setting the expectation for what happens next, which reduces the anxiety that produces a second-guessing click away.
Repeat the same action once or twice down the page, with identical wording. Different wording for the same destination creates hesitation at exactly the wrong moment: the visitor who has decided to act now has to re-read the button to confirm it is the same offer.
Reviewing the page out loud
Read the page aloud, skipping everything except the headline, the subhead, the section headings and the button text. If that path does not tell a coherent story with an obvious action at the end, the page's structure is the problem, and no amount of rewriting individual sentences will fix it.
Then read it once more as someone who has never heard of you and has thirty seconds. Mark the first sentence where you would have stopped, and fix that sentence. Repeat until you reach the end. That exercise finds the real problems faster than any amount of discussion about tone, and it is the single habit that most improves a landing page.