All posts
Web & Email

Best startup websites: what the great ones do above the fold.

What do the best startup websites put in the first screen? We took apart nine Y Combinator companies' homepages, from Stripe to 2025 startups: what each one opens with, the next step it offers and the proof it shows.

Jessica Wells·10 min read

The best startup websites do one thing well before anyone scrolls: they answer what the product is, who it is for, why to believe it and what to do next. On September 13, 2026, we looked at the first screen of nine Y Combinator companies' homepages, from Stripe to startups that went through YC in 2024 and 2025. These are observations of public pages, not conversion data, so treat them as patterns to test, not proof of what works.

Why the first screen still matters

“Above the fold” is whatever shows before a visitor scrolls. People do scroll, but they still look hardest at the top. In Nielsen Norman Group's 2018 eyetracking study of 120 people, 57% of page-viewing time went to the content above the fold, and 74% to the first two screenfuls. The first screen does not need to hold everything. It needs to make the visitor want the second one.

One first screen, taken apart: Supabase

Supabase, a database platform for developers that went through YC in Summer 2020, fits every part of a working first screen into a desktop window 800 pixels tall. Here is what we saw, and the visitor question each part answers.

  • The headline is a two-line promise about how fast you can build and how far it scales. It answers “why should I care?” in seven words.
  • A short paragraph beside it names the core, a Postgres database, and lists what you can add to it: sign-in, APIs, functions, real-time data, storage and vector search. It answers “what exactly is it?” for a technical reader who wants specifics.
  • Two buttons split the visitors: start a project for people ready to try it, request a demo for teams that want to talk first. That answers “what do I do next?” for both kinds of buyer.
  • Proof sits in the navigation: the company's GitHub star count, a signal its developer audience recognizes.
  • The product starts at the bottom of the screen, with the first feature cards and small interface previews, so the scroll begins with the product rather than more promises.
  • One extra: a banner across the top for the company's upcoming event. Useful for existing users, and one more thing for a first-time visitor to read.

A developer can tell what Supabase is, whether it fits and how to try it without scrolling. That is the job, and it takes fewer words than most founders expect.

//Free website review sheet

Write down what the page needs to explain.

Use the worksheet to compare the customer, product, proof and next action without copying another company's site.

Free. We email it to you, usually within a minute.

Already signed up on another device?

Eight more first screens

The same four questions, answered eight different ways. Each row is what the first screen showed on a desktop window 1280 by 800 pixels.

What each first screen showed on September 13, 2026 (our observations, not conversion data)
CompanyOpens withNext step offeredProof in the first screen
Stripe
YC Summer 2009
A one-line promise, a live usage counter and a sentence naming its main productsGet started, or sign up with Google; Contact sales in the headerA strip of well-known customer logos under the buttons
Airbnb
YC Winter 2009
No pitch at all: a search bar, then listings with photos and pricesSearchGuest ratings on the listings
Gusto
YC Winter 2012
The jobs a small business owner needs done, in plain wordsCreate a free account, or see how it works; a demo link in the headerNone in the first screen
Vanta
YC Winter 2018
The feeling its buyer is after, then the category and how many frameworks it coversAn email field and a demo requestA banner featuring two customers, and customer logos at the bottom edge
PostHog
YC Winter 2020
A playful, desktop-style interface and a promise to teams building AI agentsGet started, or install it with AI, beside a short list of reasons it is low risk to tryThe number of teams using it
Resend
YC Winter 2023
Its category and audience in three words, then the problem it solvesGet started, or read the documentationNone in the first screen
Mem0
YC Summer 2024
A Backed by Y Combinator badge, then what the product doesGet started, or set up an agent; tabs show how to connect itThe YC badge and its GitHub star count
Browser Use
YC Winter 2025
Its category in plain words, with one sentence for each of its two productsOne button per product, with free credits for eligible sign-upsIts GitHub star count, with more proof starting at the bottom edge

The patterns worth borrowing (not the words)

  • Open with the product, the buyer or the outcome. Resend and Browser Use name their category; Gusto and Vanta name what the buyer wants; Stripe does both in one line. Whichever you open with, the next sentence supplies the other half.
  • Keep the next step obvious. Every site put its main action in the first screen. Seven of the nine added a second button, usually for a different kind of visitor or a faster route: a demo (Supabase), the documentation (Resend), a walkthrough (Gusto) or sign-up with Google (Stripe). Vanta, which sells through demos, offered one.
  • Choose proof your buyer trusts. Customer logos for business buyers, GitHub stars for developers, an investor badge for a young company, ratings for a marketplace.
  • Show the product early. Supabase's feature cards, Mem0's setup tabs and Airbnb's listings all start in the first screen.
  • Lower the risk beside the button. A free account, free credits or a short list of reasons it is safe to try all sit right next to the action.
  • Give second audiences a path, not the headline. Gusto sends accountants to their own section of the navigation, and Stripe puts Contact sales in the header, so each headline stays on its main buyer.
  • Watch what else claims the first screen. Announcement banners (Supabase, Vanta, Mem0), chat windows (Vanta) and cookie notices (PostHog, Mem0) all appeared above the fold. Each one competes with the headline, so each should earn its place.
  • Check the phone. On a phone screen 390 by 844 pixels, Stripe kept its headline, both buttons and the start of its logo strip, but not the longer sentence below the headline. Gusto pinned its main button to the header. Your phone view is a separate first screen, and it holds less.

What these examples can and cannot tell you

Everything above is what nine public pages showed on one day. We have no access to these companies' analytics, so nothing here says which layout converts best, and the pages will change. Large companies can also do things a new startup cannot: Airbnb can skip the explanation because most of its visitors already know what it does. A startup nobody has heard of has to explain itself in the first screen.

Use the patterns as hypotheses, then test them on your own page. Our guide to what makes a landing page convert covers how to run that test.

How to build a startup website

  • Write the four answers first: who it is for, what it does, why to believe it and what to do next. The worksheet's website review tab is built for this.
  • Draft the first screen in words before any design: a headline, one supporting sentence, one or two actions and one piece of proof you have permission to show.
  • Start with the pages buyers need: a homepage, a page for each product or service, pricing or how to buy, about and contact. Add more when there is a reason.
  • Choose how to build it. A website builder is often enough at the start. A custom build makes sense when the site needs features a builder cannot handle or is part of the product itself.
  • Make it readable to search engines from the first day. Our guide to SEO-friendly web design covers the build choices that help or hurt.
  • Give each page one goal, such as sign-ups or demo requests, measure it, and change one thing at a time.

If you would rather build it with a team, our website design page explains how we work.

Frequently asked questions

What are the best startup websites?

The ones that answer four questions in the first screen: what the product is, who it is for, why to believe it and what to do next. In our September 2026 look at nine Y Combinator companies, Supabase, Stripe and Resend did this in very few words. These are observations, not conversion data.

What should be above the fold on a startup website?

A headline that says what the product is or what it does for the buyer, one supporting sentence, one or two clear next steps and one piece of proof. In Nielsen Norman Group's 2018 eyetracking study, 57% of page-viewing time went to the content above the fold.

How do you build a startup website?

Decide who it is for, what it does, why to believe it and what to do next. Draft the first screen in words, build the few pages buyers need, make it readable to search engines and give each page one goal you can measure.

Do startups need a custom website?

Not usually at the start. A website builder is often enough to launch and test a message. A custom build makes sense when the site needs features a builder cannot handle, or when the site is part of the product itself.

How do you know if a startup website is working?

Give each page one goal, such as sign-ups or demo requests, and measure it in your analytics. Change one thing at a time so you can tell what made the difference. Visits alone do not show whether a page is doing its job.

Sources

Checked September 13, 2026. Homepages change often. Each company's own site is the final word on what it shows today, and nothing here is conversion data.

Put the website question to our founders.

Tell us who the site has to convince and what it should get them to do. We will discuss what to fix first and the work that would make sense to start with.

Talk to our founders

//For businesses investing at least $3,000 a month in marketing.

Get a plan built around the business.

Our founders will review the goal and current marketing, then put the priorities in writing. You keep the plan. Any paid work is scoped separately.

No cost. No obligation. Starts with a conversation with our founders.