The Homepage Elevator Pitch, Written Down

Here is an experiment that takes ten seconds. Read your homepage headline out loud, as if you were answering a person who just asked what your company does. Not performing it. Answering with it. If you feel a flicker of embarrassment — if some part of you wants to add “well, what that actually means is…” — your headline is broken, and you just proved it to yourself.

I’ve watched this happen many times, and the strange thing is that the same founder who published the broken headline can usually deliver a perfectly good pitch in conversation. Ask him at a dinner what he’s building and he’ll say something like “we make scheduling software for dental offices.” Clear, concrete, done. But go to his homepage and it says “Reimagining practice workflows for the modern care journey.” Same person. Same product. One sentence works and one doesn’t, and he wrote both.

Why does this happen? I think it’s because speech has a built-in clarity check that writing lacks: the listener’s face. When you pitch in person and the other person’s eyes glaze, you feel it instantly and you fix it mid-sentence. You simplify, you give an example, you say “you know how dentists always double-book? We fix that.” Conversation punishes fog in real time. A homepage gets no such feedback. You publish the foggy sentence and nobody’s face falls where you can see it. The glazed eyes are real, but they’re distributed across thousands of browser tabs, and each one closes silently.

So the fix I recommend is almost insultingly simple: write down what you say, instead of saying what you wrote. Next time someone asks what you do, notice the exact words that come out of your mouth. Those words have been debugged. They’ve survived dozens of live conversations, been simplified every time a listener frowned, and converged on something that works on cold audiences. That sentence is your headline. It’s already written. You just haven’t put it on the website yet.

There’s a predictable objection here, which is that the spoken version sounds too plain. “We make scheduling software for dentists” doesn’t feel like it belongs on a homepage. It doesn’t soar. And that instinct — that a homepage should soar — is exactly the bug. Somewhere along the way we all absorbed the idea that websites are supposed to speak a different dialect than humans do, a dialect of empowering and seamless and end-to-end. Nobody talks this way at dinner, because at dinner it would get you a puzzled look. The dialect exists because writing for an invisible audience makes people nervous, and nervous writers reach for impressive words the way nervous speakers reach for a podium. But a visitor is not impressed by fog. A visitor is a person at a dinner party who asked a simple question, except she’s holding a mouse, and the puzzled look becomes a click on the back button.

The elevator pitch framing also fixes a subtler problem: order. When you pitch out loud, you instinctively lead with the thing the listener needs first — what it is — and only then get to why it’s better. Nobody answers “what do you do?” with their differentiation. Yet homepages do this constantly. The headline announces a benefit (“Cut costs by half”) or a philosophy (“We believe software should be human”) before the visitor knows what category of thing she’s even looking at. A benefit is meaningless until it’s attached to a noun. Cut costs of what? In speech you’d never make this mistake, because the listener would interrupt you. On a page, she can’t interrupt. She can only leave.

I built GazeSite partly around this observation. When it audits a homepage for clarity, the core question it asks is the dinner-party question: can a stranger, cold, in a few seconds, say what this is and who it’s for? The pages that fail are almost never failing because the founder can’t explain the product. They fail because the founder explained it to a text editor instead of a person, and the text editor didn’t frown.

If you want a slightly more rigorous version of the ten-second experiment, try this. Call someone — actually call, voice, not text — and pitch them the product. Record yourself, with permission, or just jot down what you said immediately after. Then put that transcript next to your homepage and compare, line by line. The transcript will be shorter. It will use smaller words. It will contain at least one concrete example that the homepage lacks. And it will almost certainly answer “what is this?” in the first sentence, where the homepage takes three screens to get around to it. Every difference between the two is an edit the homepage needs.

The deep point is that your homepage is an elevator pitch you don’t get to deliver in person. You won’t be there when the visitor arrives. You can’t read her face, can’t clarify, can’t add the example that saves the pitch. The page has to do the whole job with the words that are on it. Which means the words have to be the ones that already work out loud — because out loud, in front of a live human, is the only place they’ve ever been tested.

You already know your pitch. Say it, write it down, and publish that.