A homepage is the only page on your site that doesn't know who it's talking to. A landing page knows the ad that sent you. A pricing page knows you're evaluating. A homepage gets a stranger, a returning customer, a job applicant, a journalist, and a competitor, all through the same door.
Most homepages fail because they try to serve everyone equally, and equal service means nobody is served well.
The five-second question
A first-time visitor is answering one question: what is this, and is it for me?
They will answer it in a few seconds, from the first screen, without scrolling. If the answer requires scrolling, the answer is "no."
So read your first screen as a stranger. Does it say what the thing is — in the words the visitor would use, not your internal name for it? Does it say who it's for? Most underperforming homepages lead with a slogan that would fit any company in any industry. "Empower your growth" is a sentence about nothing.
Homepages don't convert; they route
Here's the reframe that fixes more homepages than any redesign: your homepage's job is usually not to close. It's to identify the visitor and send them to the page that closes.
A homepage that tries to be a sales page for every audience becomes a very long page that persuades none of them. A homepage that quickly sorts people — new here / comparing / ready / already a customer — and hands each group the right next page, converts through those pages.
This is why homepage conversion rate is a misleading metric on its own. What matters is whether visitors leave the homepage in the right direction.
Where homepages leak
A single primary action that doesn't fit most visitors. If "Book a demo" is your only button and most visitors are two weeks from buying, you've offered the majority nothing to do.
Competing actions with equal weight. Two buttons, no explanation of the difference. This is the most common friction on the web and it's near-free to fix: one primary, one visibly secondary, one line saying who each is for.
Proof buried below the fold. Trust signals belong where the decision happens, not in a testimonial carousel three screens down.
Navigation as a filing cabinet. Menus organized by your org chart ("Solutions," "Platform," "Resources") rather than by what visitors want. If a visitor has to guess which menu contains the price, the menu has failed.
Diagnosing it
High bounce, short scroll → the first screen didn't answer the five-second question.
Deep scroll, no clicks → the page explained itself and offered no fitting next step. Check whether your only action suits the majority.
Clicks spread evenly across many links → visitors are hunting. Your routing isn't routing.
Returning customers bouncing → you've buried the login, or the homepage doesn't acknowledge they've been here before.
The limit of reading your own homepage
You can audit clarity, dominance, and proof placement by looking. What you can't see is which visitor segment your homepage is losing — whether strangers bounce or researchers stall — because they look identical in a page-level number.
Defrixa's free scan scores the structural layer of the first screen and names its biggest friction. Which visitors leave, and from where, is what the snippet measures.
Common questions
One dominant action, plus clearly secondary routes for people who aren't ready. Several co-equal actions is the failure case.
Sliders reliably underperform because they move the promise before it's read. If the first slide is your best message, the others are costing you.
Usually not — it's the expected shape. Landing pages get pre-qualified traffic. Judge a homepage on whether it routes people onward, not on closing.