Good homepage design must show who you are, what you offer, who it is for, and what to do next within the first few seconds. Visitors form a first impression before reading a single word, and decide within ten to twenty seconds whether to stay.

We summarize research on first impressions and online reading, five questions for the top of the page, common business website mistakes, and how to test your site with five people in one afternoon.

50 milliseconds and 10 seconds: what research says about first impressions

In a 2006 study by Gitte Lindgaard and colleagues, people rated website appeal the same after 50 milliseconds as after a longer look. Good looks open the door, but without a clear offer they keep no one. We cover this further in our article on how emotions and the subconscious mind drive our decisions.

In 2010, Chao Liu, Ryen White, and Susan Dumais at Microsoft Research analyzed 205,873 pages. In a 2011 article, Jakob Nielsen of Nielsen Norman Group concluded that people often leave within 10 to 20 seconds—yet a page with clear value can hold them for several minutes.

Your homepage works like a shop window on a busy street: passersby stop for a clear selection, not for decorations.

Clarity, not effects

A website needs clarity, not effects: a clean layout, a readable headline, and a plain offer. Animations and large photos must not hide the message or slow down loading.

Five questions for the top of your homepage

The top of your homepage—what visitors see without scrolling—must answer five questions. This aligns with the 2024 five homepage design principles from Nielsen Norman Group: clearly show who you are, what you do, and what to do next.

1. Who you are. Logo and name at the top left. For a smaller business, a photo of your team or premises helps.
2. What you offer. A specific main headline. Not "Quality and tradition," but "Repairs and servicing for passenger cars of all makes."
3. Who it is for. A subheadline specifying who and where: "For businesses and private drivers in Kolín and nearby."
4. Why choose you. Proof instead of promises: ratings, years in business, testimonials, or photos of your work.
5. What to do next. A prominent button ("Book service," "Schedule a consultation") and, at most, one quieter alternative.

How does this look in practice? A made-up auto repair shop puts "Car service and repairs in Kolín" at the top, adds a subheadline about servicing all makes and providing a loaner car, shows customer ratings, and includes a "Book service online" button. A made-up accounting firm writes "Accounting and taxes for freelancers and small businesses," mentions remote collaboration, and backs it up with years of practice and client industries. For lawyers, see our 5 website tips for a law firm.

Neither example needs words like "welcome," "quality," or "professional"—it gives only concrete facts for making a decision.

How people read online

People do not read websites word by word; they scan. Nielsen Norman Group has tracked eye movements since 2006, when it described the F-shaped pattern, and Kara Pernice confirmed in a 2017 summary that it holds on both desktop and mobile. Readers read the first lines, shorten the next ones, and finally skim down the left edge, missing line ends and the bottom.

In the layer-cake pattern, readers scan only headings and stop at relevant sections. On mobile, marking dominates: the eyes stay fixed while a finger scrolls the text, so the main point must be visible before scrolling.

Rules for web copy:

• Lead with the point. Put the key message in the first two paragraphs and the opening words of each.
• Write clear headings. They should state what the section covers on their own.
• Use short paragraphs and lists. Number steps, bullet lists, and keep paragraphs to two to four sentences.
• Bold sparingly. Highlight key terms, not whole sentences.
• Cut the text. Move details to subpages.

Common mistakes in homepage design

The most common mistakes on business websites waste prime space on things that tell visitors nothing. You can fix most of them just by rewriting the copy at the top of the page.

Sliders and pop-ups

Rotating sliders stay popular even though research speaks against them. In 2013, Jakob Nielsen described a test where a user missed a washing machine discount because the panel auto-advanced before she could read it; the deal was visible only about 20% of the time. Motion looks like an ad, which people ignore, and hurts accessibility. The next slide should appear only when the user asks for it.

Pop-ups right on arrival do similar harm. Nielsen Norman Group advises skipping them unless required by law. Google advises against covering page content; legally required notices are exempt, yet even those need not cover half a phone screen. We cover the link to accessibility in web accessibility and the law.

When a visitor clicks an ad

When someone arrives from an ad, the page must pick up right where the ad left off. This is called message match: the headline and offer match the ad, so visitors know right away they are in the right place.

If an ad for "Winter tires fitted within 24 hours" leads to a generic auto shop homepage, visitors have to hunt for the offer and often leave. Take them straight to winter tires with pricing and a booking button, and they continue without hesitation.

Google evaluates message match too. According to Google Ads help on Quality Score, landing page experience is one of three components of this score, revealing weak spots in your ads. We explain full campaign setup in how to run a successful online PPC campaign.

Check four things for message match:

• Headline: repeats the promise from the ad.
• Offer: price, discount, and terms match the ad.
• Visuals: colors and images align with the ad.
• Button: leads straight to the promised action, not a generic contact page.

When to build a dedicated landing page

A dedicated landing page pays off for a specific offer your general homepage does not cover. It has one goal, one main button, and no menu distractions. If you promote three distinct services, point each ad to its own landing page, not a homepage hub.

Test your homepage in one afternoon

Usability testing with people unfamiliar with your site is the cheapest way to see whether your business website works. A five-second test and a task test with five people are enough—you can run both in one afternoon without paid tools.

The five-second test

The five-second test checks whether the top of your page answers the five questions above:

1. Find five people similar to your customers who do not know the site.
2. Show them the top of the homepage for five seconds.
3. Hide the page and ask: What does the business do? Who is it for? What would you do next? What do you remember?
4. Write down their answers word for word.
5. If most cannot tell what you do, rewrite the top section.

Task test with five people

A task test shows whether people find what they need. Prepare three to five tasks (for example, "Find out how much an oil change costs" or "Book an appointment"). The participant completes the tasks while thinking aloud. Do not give hints; note every hesitation.

In his 2000 article on why you only need to test with five users, Jakob Nielsen notes that five people uncover roughly 85% of usability problems. Three rounds of five users with fixes in between work better than one big test. If you serve distinct groups (businesses and households), test three to four people from each.

Why a small website cannot rely on A/B tests

An A/B test compares two page versions and requires thousands of visitors. To lift conversion from 2% to 2.5% at 95% confidence and 80% statistical power, a site needs roughly 14,000 visits per version (about 28,000 total). At 3,000 monthly visits, that takes over nine months. For a small site, user tests are faster and cheaper.

Before and after: rewriting the top of a page

Here is how rewriting the top of a homepage looks for a made-up joinery workshop.

The original version is too vague. The rewrite answers all five questions:

• Who: a joinery workshop.
• What: custom kitchens and built-in wardrobes.
• Who and where: apartment and house owners in Brno and nearby.
• Why them: photos of finished work and customer reviews.
• What next: book a measurement.

Follow the same pattern: a standalone headline, a clarifying subheadline, one strong piece of proof, and an action button. Verify the result with a five-second test.

Check mobile too. The headline, subheadline, and button must all fit on the first screen without scrolling—uncovered by a huge photo, cookie bar, or chat widget. For online stores, this clarity matters all the way to checkout; read more in our article on why shoppers abandon carts.

We start every new website with research and design, and we build brochure websites from CZK 30,000. You can get a ballpark estimate in our project configurator.

Frequently asked questions