How to write a landing page that converts

A landing page is not a small website. It is a single argument that ends in a single action, and almost every failing page fails because it forgot that: it offers a demo and a newsletter and a pricing comparison and a blog link, and each option added reduces the number of people who complete any of them.

Decide the one action first

Before any copy, write down the single thing you want the visitor to do. Start a trial. Book a call. Download the guide. Buy. Everything on the page either moves someone toward that action or is removed — including the navigation bar, which on a paid-traffic landing page is mostly an exit.

A second action is allowed only when it is a lower-commitment version of the first (start free instead of buy now), never when it is a different destination.

The first screen: three answers in five seconds

Above the fold you owe the visitor exactly three things:

1. What this is, in concrete language. Not "the operating system for growth" — what does it do, literally?
2. Who it is for. Naming the audience ("for Shopify stores under 500 products") loses the wrong visitors deliberately, which raises conversion rather than lowering it.
3. What happens if they click. "Start free — no card" is a different promise from "Request a demo", and hiding which one it is costs you the click.

The headline should be the promise, not the category. The subhead does the qualifying. The button says what happens, in the first person if it reads better ("Start my free trial"), and the line under it removes the last hesitation ("No card needed. Cancel any time.").

Then: the objections, in order

The body of a landing page is a sequence of answers to reasons not to act. Write the reasons down before writing the copy — from sales calls, support tickets, cancellation reasons, and the questions people ask before buying. A typical list:

• "I do not believe it does that." → A screenshot, a short demo, a specific example with real inputs and outputs.
• "It will not work for my situation." → Name the situations. Segments, integrations, sizes, industries.
• "It is too much work to switch." → Show the actual first steps and how long they take.
• "I have been burned before." → Refund terms, cancellation terms, and what happens to their data if they leave.
• "It is too expensive." → Price the alternative, including the hours. Never argue value in adjectives.
• "Why you and not the well-known one?" → Answer honestly, including where the other one is the better choice. It is more persuasive than pretending they do not exist.

Proof, in descending order of power

1. A number from a real customer, with the context that makes it credible.
2. A screenshot or short clip of the product doing the thing.
3. A named testimonial with a face, a role and a company — anonymous praise is worth almost nothing.
4. Volume signals (customers, sites, items generated), if the number is genuinely impressive.
5. Logos. The weakest form of proof and the most overused, and worthless if the visitor does not recognise them.

Place proof next to the claim it supports, not in a testimonial strip at the bottom. A claim and its evidence separated by three sections is a claim with no evidence.

Words that cost you money

• Adjectives doing a fact's job: powerful, seamless, cutting-edge, revolutionary. Delete them and the sentence usually improves.
• "We" copy. The visitor is not interested in your journey; rewrite each "we built" as "you get".
• Feature names as headings. "Smart Sync Engine" means nothing; "Your stock updates itself" means something.
• Vague quantities. "Save time" is unfalsifiable and therefore unbelievable. "Cuts a two-hour job to ten minutes" is a claim.
• Jargon the visitor does not use. Match the words in your support inbox, not the words in your roadmap.

Match the page to the traffic

The same page cannot serve an ad click and a search visitor well, because they arrive with different amounts of context.

• Paid traffic: the page must continue the ad — same promise, same words, same offer. Trim navigation. If the ad said "free trial", the button says "start free trial", not "learn more".
• Search traffic: the visitor arrived from a question, so answer it directly near the top and keep the surrounding site links. This page needs to be structured for search too, with a real h1 and headings that match the query.
• Email traffic: the audience knows you. Skip the introduction, go straight to the offer.

The form is part of the copy

Every field is a reason to leave. Ask only for what you need to deliver the next step — usually an email, occasionally a name. Company size, role and phone number can be asked after the account exists, when the person has already committed.

And say what happens next. "Check your inbox for the guide" or "You will get a calendar link, not a sales call" removes the fear that submitting starts a phone campaign.

What to change first when it is not converting

1. The match between the traffic and the promise. Most "bad landing page" problems are actually wrong-traffic problems.
2. The first screen. If people leave in under ten seconds, nothing below the fold is being read.
3. The objection you are not answering. Read the last twenty support or sales conversations and find the one the page ignores.
4. The form length. Removing two fields is the cheapest test in marketing.
5. The button label. "Get started" against "Start free — no card" is a real difference, not a cosmetic one.

Change one at a time, and give each change enough traffic to mean something. A page with 200 visitors a week cannot resolve a 10% difference in under a month, and pretending otherwise is how teams talk themselves into rewrites that changed nothing.

Frequently asked questions