The new site looks good. The type is right, the animation is restrained, the photography was expensive. Then you read the home page twice and still cannot say who it is for. That is not a design problem, and a redesign will produce a better-looking version of the same silence.
A home page makes one argument or it makes none.
Every other page on a startup’s site has an easy job. Pricing lists prices. About says who you are. The home page has to take a stranger who arrived from a link and give them a reason to keep going, in the time it takes them to decide whether this is worth their afternoon.
An argument has a shape: this is who it is for, this is what changes, this is what to do next. A summary of the company is not that shape, and neither is a tour.
Pages fail because founders try to make them true for everybody who might land there. The page ends up true and useless.
The page owes a stranger three answers in a screen.
Who this is for, in words that group would use about themselves. Not “modern teams” or “growing companies”. A revenue operations lead at a company with two sales reps and a spreadsheet knows which of those descriptions is about them.
What changes when they use it. Not the mechanism, not the feature list. What is different on Tuesday.
What to do next, once and unambiguously. One action, repeated where the page naturally ends. A page offering a demo, a trial, a newsletter and a report is asking the visitor to decide on the company’s behalf.
Of the three, the first is the one founders skip, because naming the reader feels like turning people away. It is the opposite. A stranger who recognises themselves in the first line reads the second one, and a stranger who does not is gone whether or not you were vague enough to include them. The headline is where that recognition happens, which is why it is worth more of the week than the rest of the page combined.
Everything below that screen is evidence for those three answers. Evidence earns its place by being checkable: a customer sentence with a name and a company attached is checkable, a wall of logos you cannot explain is not, and for a technical buyer a plain description of how the thing works often does more than either.
A list of what you do is not an argument.
The grid of capabilities with icons is common because it reads as complete and asks nothing of the founder. It also leaves the visitor to assemble the argument themselves, and they will not.
The other common version is a claim broad enough to belong to any company in the category. If a competitor could put their logo on your home page and nothing would look wrong, the page is not saying anything. That is a positioning problem arriving as a copy problem, and it gets fixed one step upstream in the sentence your startup must earn.
If you cannot say it out loud, no designer can rescue it.
The test takes a minute. Say the argument to somebody who does not work with you, in your own voice, without slides. Who it is for, what changes, what to do. If you stumble, the page stumbles in the same place, and layout work will not fix it, because the layout is not what is broken.
This is often where a founder finds out they have two arguments and have been alternating between them. Pick the one your best customers arrived through. The other is a second page, or a later year.
The page gets better by subtraction.
Take what you have and delete every section that does not support one of the three answers. What is left is shorter, and the argument stops competing with the material around it.
Then read the first screen aloud once more, on a phone.
This piece is only about the home page. The rest of the first build, and the order to do it in, is in how a founder launches a startup’s first website. The failures worth avoiding across the whole site are in five website mistakes startups make.
Making one page argue is most of what Website is.
Tell us what you’re building.