Guides

A product launch page that converts, beat by beat

How to plan, write and structure a product launch page around one action: positioning, the order of beats, headlines, calls to action, honest proof, waitlists and what to measure.

By The Laarpi teamUpdated 6 min read

A launch page has one job: get the right people to take one action. Join the waitlist, pre-order, start a trial, buy. Everything on the page should make that action more likely for someone who'd genuinely want the product, and nothing should get in their way.

That sounds obvious, and most launch pages still miss it. They open with a slogan that could belong to any product, list features nobody asked about, borrow trust they haven't earned and scatter four different buttons across the page. The launch page best practices in this guide cover planning, structuring and writing a page that converts, beat by beat, with examples from showcase pages made by the Laarpi team with Laarpi's library and process.

Decide the one action first

Before a word is written, decide what the page is for. Pick one:

StageThe one actionWhat the page must do
Before launchJoin the waitlistMake the idea irresistible; ask almost nothing
LaunchPre-order or buyAnswer every serious objection; make paying easy
Software launchStart a trial or sign upShow the product working; remove risk
After launchBuy, with confidenceAdd real proof as it arrives

Then measure that action and only that action. Clicks, scroll depth and time on page are diagnostics, not results.

Write the positioning before the page

If you can't say what the product is in one sentence, the page can't either. Fill this in honestly:

For [who], who [need or frustration], [product] is [what it is] that [the one thing it does], unlike [the alternative they use today].

You won't publish that sentence, but every headline, section and button should be consistent with it. Then list the objections: the reasons an interested person wouldn't buy. Price, trust, compatibility, effort, timing. Get them from support emails, sales calls, community threads, or by asking ten people in your audience. The page's job is to answer them in order.

The beats, in order

A launch page is read top to bottom, so plan it as a sequence of beats, each answering the next question in the visitor's head.

BeatThe visitor's questionWhat worksCommon failure
HookWhat is this, and is it for me?Name, one plain sentence, the actionA slogan with no information
ShowWhat does it look like?The real product, big, or workingA mock-up with invented data
DifferenceWhy this and not the other one?One idea, explained properlyTwelve features of equal weight
How it worksWill it work for me?Steps, with the real interface or objectAbstract icons
SpecificsWhat exactly do I get?Specs, contents, compatibility, limitsHiding the details people need
ProofCan I trust it?Real proof: demo, specs, makers, processInvented testimonials and logos
PriceWhat does it cost, when do I get it?Price, shipping date, what's included"Contact us" for a simple product
ObjectionsWhat if…?An FAQ answering the real objectionsAn FAQ of softballs
ActionHow do I get it?The same primary action, againA new, different button
PATCH-01 by Ferro Instruments: A Eurorack voice you can play in the page: turn it, patch it, take it apart, hear it.
PATCH-01 by Ferro InstrumentsA Eurorack voice you can play in the page: turn it, patch it, take it apart, hear it. Made by the Laarpi team with Laarpi’s library and process.

PATCH-01, a showcase launch page made by the Laarpi team with Laarpi's library and process for an imagined synthesizer module, follows this order. The hook names it and lets you play it: "One oscillator, a wavefolder, a resonant filter, a VCA and an LFO, in sixteen HP. Digital where it helps. Analog where it shows." The show and difference beats are the module turning and separating into its layers. The specifics beat is a paper datasheet with drawings and a spec table. The action beat returns to the module with a finish switch and the price.

Headlines that can't be swapped

The test for a launch headline: could a competitor paste it onto their page unchanged? If yes, it's not doing its job.

SwappableSpecific
"The future of sound""A Eurorack voice you can play in the page"
"Your workflow, reimagined""Turn a support inbox into a searchable FAQ overnight"
"Built for creators""Colour grading presets for film photographers, made from real stock scans"

Other rules that hold up:

  • Say what it is before what it means. People need the noun before they care about the benefit.
  • Use the customer's words. The ones from those support emails, not the ones from your pitch deck.
  • Specifics persuade. Numbers, materials, names, dates, as long as they're true.
  • No hype words. They read as filler, and visitors skip them.

Calls to action

  • One primary action, in the first screen and again at the end. On long pages, repeat it after the beats that answer big objections.
  • Specific labels. "Pre-order PATCH-01" beats "Get started"; "Join the waitlist" beats "Submit".
  • Say what happens next next to the button: the price, when it ships, whether a card is needed, how many emails to expect. Only what's true.
  • A secondary action, if any, should be visibly secondary and lower commitment.
BASALT Tapes: Obsidian Hours: An LP as a slab of obsidian: light it, turn it, break it into its eight hours.
BASALT Tapes: Obsidian HoursAn LP as a slab of obsidian: light it, turn it, break it into its eight hours. Made by the Laarpi team with Laarpi’s library and process.

BASALT Tapes: Obsidian Hours, a showcase release page for an imagined LP, puts "Pre-order from €9" in the first screen beside three facts set like a specimen label: release date, length, and how many copies were pressed. At the end, three formats appear as three plain rows with three buttons. Atmosphere in the middle, clarity at both ends.

Both pages are live in the community; remix either one and the beats come with it, ready to be rewritten for your product.

Proof without borrowed trust

New products rarely have reviews, and inventing them is both dishonest and, in many countries, illegal. Proof that's available on day one:

  • The product itself, working: a demo, a recording, a configurator.
  • Specifics: a spec table, materials, dimensions, compatibility, what's in the box.
  • The makers: who's behind it, with real names and faces, and why they built it.
  • The process: how it's made or how it works, shown honestly.
  • Guarantees you actually offer: returns, refunds, trials.
  • Press and customers, only when real and with permission.

Design that serves the beats

The look of a launch page should come from the product, and it should make every beat easier to read.

Industrial Spec Sheet specimen: Tolerances, part numbers and safety orange. The site reads like an engineering datasheet for something you want.
Industrial Spec SheetTolerances, part numbers and safety orange.

A direction such as Industrial Spec Sheet, with its datasheet layout and safety orange, makes specifics the hero for tools and hardware. For consumer products, a warmer or more playful look may suit better. Typography does more of the work than people expect; see choosing fonts that aren't Inter.

If the product is physical, real-time 3D can let people turn it and see inside it, as long as the text loads first and the 3D explains the product rather than decorating it. See how to build a 3D website.

Waitlists done properly

A waitlist page is a launch page with the price beat removed and the bar for action lowered.

  • Ask for an email. At most one more question, if the answer genuinely helps you (which colour, which platform).
  • Confirm straight away, on the page and by email, and say what will happen next and when.
  • Own the list. Store sign-ups somewhere you control and can export.
  • Tell them first. People on the list should hear about the launch before anyone else, and often get something for waiting.

Launch day checklist

  • The page loads fast on a phone, with the action visible without scrolling
  • The share image and description look right when the link is pasted into messages and social posts
  • The checkout or sign-up works end to end, tested with a real payment or account
  • Analytics records the one action
  • The FAQ answers the questions you've already heard
  • Someone is watching the inbox

After launch: improve with evidence

Watch where people drop off, read every question that reaches you, and update the page to answer them. Add real proof as it arrives. If you test variants, test big differences and judge by the one action; small tests on small traffic produce noise.

Building a launch page with Laarpi

Laarpi is a coding agent that builds launch pages from a description. It asks the questions that change the page (is this a waitlist or a sale on day one; should people be able to use the product on the page), writes a plan with the beats in order that you can edit, and builds it in real code. Checkout for software and digital products runs through your own Polar account, wired to your buttons; waitlist sign-ups save to your own Supabase project. It checks its work at six widths and has a separate art director score it before revising at least twice on a Finish. See product launch pages and SaaS landing pages for how it works in practice.

Questions

Fair questions

What is a good conversion rate for a launch page?

It varies so much by product, price, audience and traffic source that any single benchmark misleads. Measure your own baseline from launch day, then improve against it. Traffic from your own mailing list will always convert better than cold traffic.

How long should a launch page be?

As long as it takes to answer the objections of someone who's interested, and no longer. A cheap, familiar product can be short. An expensive or unfamiliar one needs more: how it works, specifics, an FAQ. Put the action at the top and the end so nobody has to read it all.

Should a launch page have more than one call to action?

One primary action, repeated where it's needed. A secondary, lower-commitment action, such as 'see how it works' or 'join the list', is fine if it's clearly secondary. Several competing primary buttons split attention.

How do I write a launch page with no testimonials?

Use the product as proof: a live demo, real screenshots, a recording, specifications, who's behind it and how it's made. Never invent testimonials or logos. Add real ones with permission once customers have them.

Should I build a waitlist before launching?

Usually yes. A waitlist gives you a warm audience on launch day, which is when a page converts best. Ask for an email and at most one useful question, confirm the sign-up straight away, and tell people what will happen next.

Should I A/B test my launch page?

Only if you'll have enough traffic for the result to mean something. With small numbers, test big differences (a different headline or a different first screen), not button colours, and judge by the action that matters, not clicks.

Related

Start building

Describe the site in a sentence. It asks what matters, then designs and builds it from scratch.

One sentence is enough.