Because most people argue about sections when the real question is jobs. A homepage has three jobs. Tell a stranger what you do, give them a reason to believe you, and point them at the next step. Every section either serves one of those or it is decoration.
Homepage meetings go badly for a predictable reason. Everyone in the room already knows what the company does, so nobody is qualified to judge whether the page explains it. The founder wants the story. Sales wants the offer. Design wants the space. The visitor wants to know if they are in the right place.
We build a lot of homepages, and the ones that work are almost always shorter and blunter than the first draft everyone agreed on.
It is a routing page for people who do not yet know what they need. Most visitors arriving on a homepage are either early in their research or checking whether you are real. The page has to orient them quickly and then send them somewhere more specific. It is not where deals close.
This is worth internalising because it changes what success looks like. A homepage that keeps everyone on it is not doing well. A homepage that sends the right people to the right deeper page quickly is doing exactly its job.
It also explains why homepages get bloated. Every team wants their thing represented, and because the homepage belongs to nobody in particular, everything gets added. The page ends up describing the company org chart instead of answering the visitor's question.
The discipline is to treat homepage space as the most expensive space you own. Something only earns a place if it helps a stranger decide, quickly, whether to go deeper.
Around ten seconds. Nielsen Norman Group's research on how long users stay on pages is direct about this, stating that "the first 10 seconds of the page visit are critical" and that "to gain several minutes of user attention, you must clearly communicate your value proposition within 10 seconds."
The underlying data is substantial rather than anecdotal. Nielsen Norman Group cites a Microsoft Research study by Chao Liu and colleagues, published in 2010 in the proceedings of the 33rd international ACM SIGIR conference, which analysed 205,873 web pages with more than 10,000 visits each, covering over two billion dwell times.
Nielsen Norman Group summarises the outcome simply, noting that "the average page visit lasts a little less than a minute." That is the whole budget for most visitors, and the first ten seconds decide whether you get the remaining fifty.
So the honest test for any homepage is this. Show it to someone outside your industry for ten seconds, take it away, and ask what the company does. If they cannot answer, no amount of scrolling further down will save it.
What you do, who it is for, and what makes you different, in language a stranger uses. Not your tagline. Not your mission. A plain sentence that would make sense read aloud. Then one obvious next step beneath it.
The most common failure we see is a headline written for people who already understand the company. Phrases like "reimagining workflow intelligence" survive internal review because everyone in the room fills in the meaning from memory. A first-time visitor cannot.
Specific beats clever every time. "We rebuild slow websites in Webflow" tells a stranger more in six words than a paragraph of aspiration. If your headline could belong to a competitor, it is not doing any work.
The visual half matters as much as the words. A hero image that shows the actual product or the actual work beats abstract illustration, because it answers the "what is this" question without the visitor reading anything. We go deeper on that in our guide to designing a hero section that converts.
They scan in patterns, not lines. Nielsen Norman Group's eyetracking research identified an F-shaped pattern, where users read horizontally across the top, make a shorter horizontal pass lower down, then scan vertically down the left edge. Content outside those paths is frequently never seen.
That study observed "232 users looked at thousands of Web pages" and the finding has held up. Nielsen Norman Group notes that an eleven year follow-up study confirmed the original results, which is unusual longevity for web research and a good reason to design around it.
The practical instruction from that research is blunt. Nielsen Norman Group advises that "the first two paragraphs must state the most important information" and to "start subheads, paragraphs, and bullet points with information-carrying words that users will notice when scanning down the left side."
Applied to a homepage, this means section headings carry the message on their own. If a visitor read only your headings and nothing else, they should still understand your offer. We design for that reading and treat the body copy as support, not as the primary channel.
Five, usually. A hero that explains the offer, a section that names the problem you solve, proof that you have solved it, a short explanation of how you work, and a clear close. Everything else is optional and should have to argue for its place.
The problem section is the one most often missing. Before a visitor cares what you sell, they need to recognise their own situation described accurately. A single paragraph naming the frustration in their words does more persuasion than three paragraphs of feature description.
The how-you-work section exists to reduce risk, not to explain process in detail. A stranger deciding whether to enquire wants to know roughly what happens next and roughly how long it takes. Three sentences beat a flowchart.
The close should repeat the offer and the next step. Visitors who scroll to the bottom are the most interested people on the page, and asking them to scroll back up to find a button is a genuinely costly mistake.
Early, then repeated. Put one credibility signal near the top, within the first screen if you can, because that is where the stranger decides whether to keep reading. Then place fuller proof next to the claims it supports rather than gathering it into one block.
A logo strip near the top works because it is fast to process. It answers "have real companies trusted these people" without requiring any reading. It is weak proof, but it arrives at exactly the moment weak proof is enough.
Specific proof belongs deeper and works hardest when it sits beside a claim. A results figure next to the promise it backs up is far more persuasive than the same figure in a testimonial carousel three screens away. Proximity is most of the power. We break down what makes proof believable in our piece on designing social proof people actually believe.
Be careful with numbers you cannot support. A visitor who doubts one statistic starts doubting the rest of the page, and the recovery cost is much higher than whatever the number was worth.
One primary action, repeated. You can offer a secondary path for people who are not ready, but there should never be two equally weighted options competing for attention. Two equal buttons make the visitor choose between you and you, which usually means choosing neither.
Repeat the primary action as the page gets longer. Someone who scrolls past four sections has demonstrated interest, and they should never have to hunt for the next step. Repetition is not clutter when it is the same action.
Make the secondary path genuinely secondary. A quiet link to case studies or pricing gives the unready visitor somewhere to go without stealing weight from the main action. Styled as a full button, it competes; styled as a text link, it supports.
The wording matters more than the colour. "See how we work" sets a clear expectation. "Learn more" sets none, which is why it consistently underperforms. We cover this in more depth in our guide to whether above the fold still matters.
Fast enough that the ten second window is spent reading rather than waiting. Google's web.dev documentation sets the good thresholds at 2.5 seconds for Largest Contentful Paint, 200 milliseconds for Interaction to Next Paint, and 0.1 for Cumulative Layout Shift, measured at the 75th percentile of page loads.
Homepages are usually the slowest page on a site, which is backwards. They carry the biggest hero image, the most animation, the video background, and every tracking script the marketing team has ever added. The page with the least patience from visitors gets the heaviest build.
Layout shift deserves special attention here. A homepage that reflows while loading pushes the headline out from under the visitor's eye during the exact seconds when they are deciding whether to stay. Reserving space for images and fonts costs nothing and prevents it.
Our approach is to set a weight budget for the homepage before design starts, then hold to it. It is far easier to decide the video is too expensive at the wireframe stage than to remove it after everyone has fallen in love with it.
Run the ten second test this week. Show your homepage to three people outside your industry for ten seconds each, then ask what your company does and who it is for. Their answers, not your opinion, tell you whether the top of the page works. Fix that before you touch anything below it.
Then go through every remaining section and ask which of the three jobs it serves. Anything that explains, proves, or directs stays. Anything that exists because someone wanted it represented is a candidate for a deeper page instead.
If your homepage is doing a lot of work and none of it seems to land, we are happy to look at it with you and say plainly what we would cut. Reach out through phoenix.studio and send us the link.
Tell us where you want to go. We'll tell you how we'd get you there.