How to write a landing page for an MVP

Most visitors will not read your MVP landing page; they will scan it, so write the page for the scan. In his research on how users read on the web, Jakob Nielsen of Nielsen Norman Group found that "79 percent of our test users always scanned any new page they came across; only 16 percent read word-by-word." A later Nielsen study on how little users read estimated that on an average page, users have time to read at most 28 percent of the words during a visit.

Those numbers come from usability research on general web pages, not from MVP landing pages specifically, and the first study dates from 1997. They still describe the reader you are writing for: someone who arrived from a link, gives the page a few seconds, and leaves unless the first words tell them this is about their problem. One page, one promise, one ask is the structure that survives that kind of reading.

What is the one promise an MVP landing page should make?

The promise is the single problem you solve, stated in the words your future users already use for it. An MVP exists to test one assumption, and the landing page is where that assumption meets strangers, so a page that promises three things tests nothing: if it fails, you cannot tell which promise failed.

The 37signals book Getting Real lists nine things a promotional site can include in its chapter on a powerful promo site, from a tour and a manifesto to case studies, a forum and a blog. Several of those items, such as case studies, customer testimonials and a forum, only work once people already use the product. For an MVP, cut it to one promise, the proof behind it and one ask, and add the rest only once the promise has held up.

Find the promise before you write a word of copy. Read your interview notes, support emails or forum threads and copy out the sentences where people describe the problem in their own words. If you have none, the page is premature; validate the idea first, because a landing page written from your own description of the problem will test your vocabulary, not demand.

What is the outline of an MVP landing page?

The outline has five elements, and each has one rule. The table is the template; the filled example comes after it.

Element	Rule	Fill in
Headline	Leads with the pain, eight words or fewer, no product name, and no words like AI or fast	The moment the problem hurts, in the user's words
Subhead	One sentence on what the product does and for whom	[Product] [does what] for [who] so they can [outcome]
Three bullets	Each is an outcome in the user's work, not a feature or a time saving	Something that happens in their week that would not happen without you
Proof	Only what you actually have	Your own experience, a shared interview line, a screenshot or short video of the real thing
One ask	A single call to action, sized to the trust you have earned	Email, a 15-minute call, or a presale

The headline rule does most of the work. Kara Pernice's article on the F-shaped reading pattern at Nielsen Norman Group advises starting headings with the words that carry the most information, because "if users see only the first 2 words, they should still get the gist of the following section." An eight-word limit with the pain at the front means the first two words are about the reader, not about you.

The bullet rule is the one most founders break. Export to PDF is a feature. Saves 3 hours a week is a time saving the reader cannot check. An outcome is a thing that happens in their work: an invoice gets paid before the next one is due, a client says yes to the first draft. If a bullet would read the same on a competitor's page, it is not an outcome yet.

What does a feature-led page look like before and after the rewrite?

A feature-led page describes the product; a pain-led page describes the reader's week. The example below is hypothetical, written for this guide: an invoice-reminder tool for freelance translators, with no real company or data behind it.

Element	Before (feature-led)	After (pain-led)
Headline	AI-powered invoicing that saves you hours	Stop chasing clients for late invoice payments
Subhead	Smart automation for modern freelancers	Polite reminders go out for you, in your client's language, until the invoice is paid
Bullet 1	Automated reminder sequences	You stop writing the awkward third follow-up email yourself
Bullet 2	Multi-language templates	Clients in Germany and Brazil get a reminder they can read without a translator
Bullet 3	Real-time dashboard	You see on Monday which clients have not paid, before rent is due
Proof	Logos of tools it integrates with	A two-line note from the founder, a freelance translator who chased payments for years, and a 30-second screen recording
Ask	Start free trial, Book a demo, Join newsletter	Join the first 20 translators testing it

The before column is not wrong, it is just about the product. Every line in the after column could only be written by someone who has listened to translators describe the problem. The ask also changes from three choices to one, and the one it picks matches a cold visitor's trust: an email address, not a credit card.

To practise the same rewrite on a page with real-looking numbers attached, work through the landing page conversion diagnosis challenge, which asks you to diagnose a page with 847 visitors and 31 signups, then rewrite its headline under an eight-word limit.

What proof can you show when you have no users yet?

Proof on an MVP page is whatever a sceptical stranger could check, and the honest list is shorter than most templates suggest. You probably have your own history with the problem, a few interview lines people agreed to let you share, and a product, or part of one, that works. Show those. A short screen recording of the real flow, even an unpolished one, is stronger proof than a polished mockup, because it shows the thing exists.

What you do not have, you leave off. No invented testimonials, no trusted-by logos of companies that tried it once, no user counts rounded up. A visitor who spots one fake signal discounts everything else on the page, and the people most likely to spot it are the early adopters you are trying to reach.

Your name and face count as proof. A page with a named founder, a sentence on why they built it and a real email address answers the question every visitor to an unknown product has: is anyone behind this?

How do you choose the one call to action?

Choose the ask by how much trust the page has earned with the person reading it, and then show only that ask. Getting Real's chapter on soliciting early makes the lowest-trust ask the default for a brand new product: "Get some sort of site up and start collecting emails as soon as possible." An email address is the right ask for cold traffic from a community post or a launch.

Move up the ladder as trust grows. People you have already interviewed can be asked for a 15-minute call. People who have told you the problem costs them money can be asked to pay before the product is finished, and how to presell a product covers that page and the trust levers it needs. Whatever you choose, one page gets one ask; a page with a trial button, a demo button and a newsletter form splits a small audience three ways and teaches you nothing about any of them.

Make the click lead somewhere honest. If the button promises early access, the next screen should say what happens now and when you will be in touch, and you should then write to every signup yourself. Paul Graham's essay Do Things That Don't Scale is blunt about why: "You can't wait for users to come to you."

Which three checks should you run before publishing?

Run three checks, in this order, and fix the page between each.

• The scan check. Show the page to someone who fits your user for five seconds, then close it and ask what the product does and who it is for. If they describe a feature or cannot answer, the headline and the first words of each bullet are not carrying the message.
• The swap check. Replace your product name with a well-known competitor's in every line. Any line that still reads as true is generic; rewrite it until it only fits you. AI for positioning and messaging walks through this test and how to generate variants that pass it.
• The ask check. On your phone, click the call to action as a stranger would. Confirm it works, that the next screen matches the promise, and that you have a written plan for replying to each signup within a day.

Getting Real's chapter on copywriting as interface design sums up the standard these checks enforce: "Keep it short and sweet. Say what you need to and no more."

How do you read the result of an MVP landing page?

Read the result against a number you wrote down before the page went live, and read each stage separately. Before launch, write three things: where the traffic will come from, how many of the right people you expect to reach, and how many of them need to take the ask in a fixed window, such as two weeks, for you to keep going.

Then split what happened into stages, because each stage failing means something different:

Stage	If it fails, it usually means	What to change
Visitors who left without acting	The headline or audience is wrong	The promise, or the channel
Signups who never replied to you	The page promised something the follow-up did not deliver	Your first email and the expectation the ask set
People who replied but never used it	The product does not yet solve the problem the page named	The product, not the copy

Small numbers need care. A page with 40 visitors cannot give you a meaningful conversion rate, but it can give you five conversations, and five conversations about why someone signed up tell you more than a percentage. Count only the right people: signups from friends and colleagues measure goodwill, not demand.

The honest limit of a landing page is that it measures interest in a promise, not use of a product. A strong result is permission to build the smallest version that keeps the promise, and the activation metric guide is the next test, because it tells you whether the people who signed up got the value the page described.

Where does the landing page fit in a launch plan?

The landing page is the first artefact a launch plan produces, and it has to exist before any distribution move, because every channel sends people to it. A Product Hunt launch, a community post or an outbound message to 30 past contacts all end on the same page, so the page decides what each of those channels can teach you.

Builders Camp's From Idea to Launch with AI bootcamp covers this sequence in 1 week, with 2 live sessions and 8 self-paced microlessons. Its curriculum includes launch narrative and positioning (a story, messaging and a landing page outline), MVP scoping, distribution options for first users without an audience, and traction metrics for after the first launch.

See the From Idea to Launch with AI bootcamp

Before you write the headline, open your interview notes and highlight every sentence where someone describes the problem without mentioning a solution. The headline you need is usually in one of those sentences, already in the words your users will recognise.