Teardown: what can a small company copy from Ramp’s homepage?

Contents 7 sections

Ramp sells corporate cards and finance software. Its homepage is built around money from the first word, which makes it a useful page to take apart if you are about to brief your own. I opened ramp.com on 11 November 2026, read it top to bottom on a laptop and at phone width, and measured what it loads. Homepages change every few weeks, so read this as a record of one day.

What it does well comes first. Then what I would change, and which half a ten-person company can use.

Fig.

What does the first screen say?

The headline is five words: “Time is money. Save both.” It doesn’t name the product. The line under it does: cards, expenses, bill payments and banking. So the buyer reads the outcome first and the category second.

That works for Ramp because its buyer, a finance lead, already knows what a corporate card is. The headline only has to say why this one. If your buyer has never heard of your category, the line under the headline has to do more of the work, and I’d let it run a little longer.

Above the headline, a counter ticks: the share of US corporate payments that go through Ramp, to seven decimal places. At the foot of the first screen a strip counts “agents at work today”: expenses reviewed, invoices processed, accounting fields coded. Proof that moves is memorable. I couldn’t find a note on the page saying how those numbers are counted, and a CFO will want one.

Where can a visitor act?

In the first screen, in one place: an email field and a button that says “Get started for free”. The header adds “See a demo”. The same email field comes back twice more, in the closing block and in the footer, and three more buttons lower down say “View Demo”. There are other links, to reports and product pages, but the two main actions never change.

One action, repeated, is easier to act on than five different ones. If you sell a service, the email field becomes your booking link, and it can sit in the same three places.

One snag. At phone width, the cookie consent panel covered the button until I answered it. That is the law doing its job, not a design fault, but it is worth opening your own homepage on a phone with a clean browser and seeing what the first screen really shows.

How does it prove the big number?

The first proof block reads: “Join 85,000 of the world’s most ambitious companies growing 3.2x faster than the average American business.” Under it, a link: “Read the report”.

The report is good, and I’d copy the habit of linking it. It says where the number comes from: revenue data that businesses shared with Ramp, tracked from six months before signing up to twelve months after, against a US baseline from the Federal Reserve Bank of Atlanta’s Survey of Business Uncertainty.¹1Published 16 April 2026. Ramp-approved companies that moved a large share of their spend to Ramp grew 19.1%. Ramp, Ramp customers grow revenue 3.2x faster than the average American business, read 11 November 2026. It also says, in its own words:

This analysis does not prove that Ramp caused faster growth.

Ramp, Ramp customers grow revenue 3.2x faster than the average American business

The report has three groups, and the homepage uses one of them.

The 3.2 is 15.9 divided by 5, near enough. The fairer comparison is the second group: companies that passed Ramp’s approval and then didn’t use it. Against them, users grew about 1.5 times as fast. That is still a strong number, and it is the one a CFO will work out anyway. I would put it on the homepage and let the honesty do the selling.

Lower down, every quote carries a person’s name, job title and company. That costs nothing and it is the cheapest proof there is.

What does the quiz do?

Halfway down, a block asks “What would you like to spend less time on?” with six answers: chasing receipts, approving spend, paying bills, closing the books, managing travel and tracking subscriptions. You type your company’s website, pick one and press “Get my recommendations”. A line under it says Ramp uses AI to suggest the content.

Ramp is testing this block. On 12 November the version I was sent had no list of chores, only one box for a company website or a few words about the business.

Each answer is a chore in the buyer’s words, not a feature name. A controller doesn’t wake up wanting accounts payable automation. They want to stop chasing receipts. The quiz is also a second way in for someone who isn’t ready to sign up. The first thing it asks for is a website, not an email, and Ramp learns which chore brought them.

A small company can build this without the AI. Three questions, each in the buyer’s words, and an answer that ends at the booking page.

Which worry does it answer?

A company that buys Ramp already has a way of paying for things, and has to move off it. Ramp answers that worry on a button in the first section after the proof: “Switch in days, not months”. Five words, linked to a page that walks through the switch in four steps.

An answer in an FAQ at the bottom reaches only the people who scroll that far. An answer on a button reaches everyone who reads the button. If you know the question every first call starts with, this is where its answer goes.

What does it cost to load?

This is where I’d be careful copying. I loaded the page in Chrome, empty cache, without accepting cookies, and counted what arrived in the first eight seconds without scrolling.

5.21 MBVideo2.25 MBScripts1.5 MBData0.9 MBThe rest

Fig.

Source

My measurement, headless Chrome, 11 November 2026

That is 9.9 megabytes, more than half of it three video files, from 31 different hosts, most of them analytics and advertising services. The median home page on a phone weighed 2.56 MB in 2025.²22.56 MB on mobile and 2.86 MB on desktop, July 2025. HTTP Archive, Web Almanac 2025, Page weight, read 11 November 2026. Ramp’s page is almost four times that.

Ramp can probably afford it. My guess is that its buyers read the page at a desk, and the brand carries a lot of the weight. If your traffic is paid clicks on phones, it can’t, and the video is the part I’d leave out.

What Ramp gets right underneath: the headline, the quiz and the buttons are in the HTML the server sends, before any script runs. A slow phone shows the words first, and AI crawlers that don’t run JavaScript can read them. That part costs nothing to copy.

What can a ten-person company copy?

The left column is cheap. None of it needs a big team, only someone who decides what the page is for and says it plainly.

If your homepage needs the left column, that is a website project.

Notes and sources

1. Published 16 April 2026. Ramp-approved companies that moved a large share of their spend to Ramp grew 19.1%. Ramp, Ramp customers grow revenue 3.2x faster than the average American business, read 11 November 2026. ↩︎

2. 2.56 MB on mobile and 2.86 MB on desktop, July 2025. HTTP Archive, Web Almanac 2025, Page weight, read 11 November 2026. ↩︎

Written by Dion Zeneli. AI helped with research and editing. Sources are linked where they are used. Read this note as markdown.