The "What Is This?" Problem Most Homepages Never Solve

There’s a question every first-time visitor silently asks a homepage, and it’s the same question you’d ask if someone handed you an unlabeled object: what is this? Not what’s your mission, not what do you believe about the future of your industry. What is this thing. It’s the most predictable question a page will ever face, which makes it remarkable how many pages never answer it. But what interests me is not that they fail. It’s how they fail, because the failures fall into patterns, and each pattern has a different cause.

The first pattern is the scattered answer. The information is technically all there — a fragment in the headline, a fragment in a subhead, the crucial noun finally appearing in a feature card halfway down. A determined reader could assemble it, the way you could assemble a chair from parts shipped in separate boxes. But visitors aren’t determined readers. They’re triaging. Asking a visitor to assemble your explanation from fragments is asking them to do the writer’s job, and they respond the way anyone responds to being handed someone else’s job: they decline. Scattered answers usually come from pages written by committee, where each section had an author and no one owned the whole.

The second pattern is the implied answer. The page assumes you arrived knowing. It speaks in the definite article — “the platform,” “your workflow,” “your team’s data” — as if resuming a conversation you don’t remember starting. This is common in companies that grew through word of mouth or sales teams. For years, everyone who hit the homepage had already heard the pitch from a person, so the page never needed to introduce anything, and nobody noticed it couldn’t. Then the company starts buying ads or ranking in search, cold strangers arrive for the first time in volume, and the homepage — which was really a lobby for people who’d already been sold — quietly fails them all.

The third pattern is the deferred answer. The page knows it owes you an explanation and keeps promising to pay: “Discover a new way to work.” “See how it works.” “Watch the demo.” Every element points to some future moment of understanding — a video, a tour, a sales call — and the present moment stays empty. I think this comes from a belief that the product is too rich to summarize, that any one sentence would undersell it. But refusing to summarize doesn’t preserve the richness; it just moves the summarizing job onto the visitor, who does it with far less information than you have, and often gets it wrong. If someone will only carry one sentence out of the visit, and they will, the only question is whether you wrote that sentence or they improvised it.

The fourth pattern, subtlest of all, is the answer to the wrong question. The page crisply answers “why are we exceptional?” — fastest, smartest, most trusted — before answering “what are we?” Superlatives without a noun. “The most powerful platform for modern teams” is grammatically an answer, but to a stranger it identifies nothing, because every kind of software could hide behind it. It’s as if you asked what’s in the box and were told “the best one in its category.” You still don’t know the category. Comparatives and superlatives are only meaningful after the noun is established, and the noun is exactly what these pages withhold.

Notice that none of these is really a copywriting problem, in the sense of needing better sentences. They’re structural problems: the answer is in the wrong place, or assumed, or postponed, or answering a different question. That’s why hiring someone to “punch up the copy” so often fails to fix them. Punchier fragments are still fragments. What’s needed is more like editing than writing: someone has to stand where the stranger stands, ask “what is this?” at the top of the page, and refuse to move until the page answers in one place, in the present tense, in words that name the thing.

The reason the people who own the page can’t do this themselves is that they are the one audience on earth incapable of asking the question sincerely. You cannot un-know what your product is. When you look at your homepage, the answer is superimposed on it like a watermark only you can see. This is why every version of the fix involves borrowing eyes — a friend from another industry, a hallway stranger, or a machine. It’s a big part of why I built GazeSite the way I did: it looks at a page the way a first-time visitor does, with no memory of the sales deck and no watermark, and reports whether the “what is this?” question actually gets answered on the first screen, and if not, which of these failure modes it found. A machine has no politeness and no context, which for this particular job are both features.

But the diagnosis matters more than the tool. Once you know the four patterns — scattered, implied, deferred, wrong question — you start seeing them everywhere, including, uncomfortably, on your own homepage. Seeing the pattern is most of the cure. The rest is one honest sentence, placed where the question is asked.