What the first ten seconds of a startup homepage have to do

The first ten seconds on a startup homepage above the fold have to answer two questions: what this is, and how to talk to a person. Nielsen Norman Group's review of page-visit research puts a clock on that moment. People often leave a web page within 10 to 20 seconds, and the first 10 seconds are when they decide whether to stay. Jakob Nielsen's summary is the practical rule: to earn several minutes of attention, the page has to communicate its value inside those 10 seconds.

A short brand line can sit above that answer. It cannot replace it. The next sentence has to name the product in words a stranger already uses, and a visible action has to show how a conversation starts. If either answer is missing, the rest of the page is material the visitor is unlikely to reach. In the same 2011 Nielsen Norman Group article, the average page visit lasts a little less than a minute, and people have time to read only about a quarter of the text on pages they do open. The first screen is the page, for most arrivals.

Two questions, and only two

Judge the opening with two questions.

What is this? How do I talk to a person?

"What is this" means a category the visitor already understands, tied to the specific thing you sell. "Scheduling for independent clinics" is an answer. "The future of care" is a mood. "How do I talk to a person" means a way to book, write, or call, with a label that says what happens next. "Book a 30-minute intro" is an answer. "Learn more" still needs a destination.

Everything else in that first view is support. A product picture, a customer type, one true proof point. If an element does not help one of the two questions, it can move down the page. A mission paragraph, a logo wall, and a carousel of features can be useful later. They are expensive in the first glance, because they spend the only seconds you are likely to get.

Where a brand line still belongs

A poetic line is a fair choice when the following sentence does the explaining. The pattern is short. Line one can be the feeling you want associated with the company. Line two names the buyer and the thing you make. The action sits where a thumb can reach it.

Hanabi keeps a brand line on its own site, then says what the studio designs and builds. That pairing is the standard to copy, in structure if not in wording. The feeling and the fact are allowed to sit together. The feeling is not allowed to stand alone.

Read the pair out loud, without the layout. A friend who does not know the company should be able to repeat what you do and how to reach you. If they can only repeat the mood, the brand line is carrying a job it was not given.

Teams stall here for a human reason. They have spent months on the name and the tone, and a sentence like "we build scheduling software for independent clinics" feels plain. Plain is the point of that second line. The brand line can carry the care. The next sentence carries the fact a stranger can act on.

What the first screen should hold

Keep the set small enough to read without scrolling on a phone.

• A line that can be poetic, if you want one.
• A sentence that says who it is for and what it does.
• One action. If the page is long, you can repeat it at the end. In the first screen, once is enough.
• One picture of the product, the place, or the work. A concrete thing is easier to trust than a gradient.
• One proof point, if you have one that is true. A named kind of customer. A public project. A constraint you are known for handling. Skip a wall of logos you cannot stand behind.

The picture has to arrive with the sentence. Google's Core Web Vitals guidance treats Largest Contentful Paint as the loading measure, and asks that it occur within 2.5 seconds of the page starting to load. When the headline and the product image are the largest things in view, that threshold is what protects them. A frame that shows up late has missed the judgment Nielsen describes.

Type size matters more than decoration. On a phone, the explanatory sentence should be readable without pinching. The action should be a real control with visible words, not an icon the visitor has to interpret.

What can wait below

Below the first view, you can teach.

A short section on how the product works. Two or three examples. A note on who it is not for. The shape of pricing, if you are ready to say it. A few questions you already hear on calls, answered in full sentences.

None of that repairs an opening that failed the two questions. People who leave in the first 10 seconds do not scroll to a FAQ. Write the lower page for the person who stayed. Nielsen's account is that once someone has stayed for about half a minute, there is a fair chance they will stay much longer, often two minutes or more. Earn that half minute. Then spend it on facts.

Each lower section should add something the hero did not say: a step, a limit, a piece of proof, a question with an answer. Repeating the hero in softer words does not count. The visitor who stayed is willing to read, and they can tell when a section is a second copy of the first.

How to test the ten seconds

You can test the opening in an afternoon. You do not need a research panel.

1. Open the page on a phone, on a normal connection. Use the live page or a real build, not a still frame in a design file.
2. Show it to someone who has not worked on it. Give them about ten seconds, then hide it.
3. Ask what the company does, and what they would do if they wanted to talk to someone.
4. If either answer is wrong or empty, change the sentence or the action. Do not add a paragraph under the fold to compensate.

Three people are enough if you are listening for the same miss. If two of them say "some kind of platform," the sentence is still abstract. If they describe the product and then ask where they would click, the action is too quiet.

If you feel the urge to explain before they answer, the page is relying on you. A homepage cannot bring the founder into every visit. Run the test again after visual polish. Decoration has a way of shrinking the sentence that does the work.

When the opening is still wrong

A few patterns show up once the test is honest.

The product is hard to explain, and the team reached for a category word. The fix is a concrete object and a one-sentence job, which is a different piece of writing from a vague brand line. The page is attractive and still hides the conversation. The fix is one action with a plain label, placed where the first screen can hold it. The page lists everything the company has ever done. The fix is to choose the offer this site is for, and let the other offers wait on a later page or a later year.

If the first screen is doing its job and you are deciding what else the site should contain, the next choice is which pages a young company actually needs. Read the pages a seed-stage site actually needs before you add a blog, a careers page, or a second product line.

When you want that first screen shaped with the rest of the site, look at Hanabi's design and development services. The work is a homepage that answers quickly, then the pages that support the answer.