How Far Down a Long Page Does Attention Actually Go?

How far down a long page do people actually read?

Attention concentrates near the top and falls away sharply. Nielsen Norman Group's eyetracking research found that users spent about 57 percent of their page viewing time above the fold, and 74 percent of viewing time in the first two screenfuls. Everything below that shares the remaining quarter.

I like this research because it gives designers a number instead of a feeling. Arguments about page length usually run on instinct, and instinct varies by whoever is loudest in the room.

The number does not mean long pages fail. It means the order of what you put on them matters far more than the length.

What did the research actually measure?

It measured viewing time, not comprehension or conversion. Eyetracking records where eyes rest and for how long. That tells you which parts of a page get looked at, which is the foundation of every other outcome but not the same as any of them.

The distinction matters when someone quotes the figure at you. A section that gets 5 percent of viewing time is not worthless. It may be the section a serious buyer studies carefully before signing, and serious buyers are a small fraction of visitors by definition.

So use the research to decide sequence, not to decide what deserves to exist. Those are two different questions and people routinely collapse them into one.

Why did the number change between 2010 and 2018?

Because people got used to scrolling. Nielsen Norman Group's comparison is explicit: in 2010, 80 percent of viewing time was spent above the fold, and by the later study that number was 57 percent. Behaviour shifted as long scrolling pages became normal everywhere.

What did not shift is the shape. The research notes that the pattern of a sharp decrease in attention following the fold remained the same in the later study as in the earlier one. People scroll more than they used to, and they still pay dramatically less attention as they go.

That combination is the actual insight. You can build long pages, and you cannot expect the bottom of one to carry the same weight as the top. Designers who learned the first half of that lesson and not the second produce beautiful pages that bury the point.

What does this mean for the top of your page?

It means the first screenful has to do the whole job on its own. Name what you are, say who it is for, and give one clear next step. If a visitor left immediately after the first screen, they should still be able to describe your business correctly.

The common failure is using the top for atmosphere. A large image, an abstract headline, and a scroll prompt spend the most valuable real estate on the page telling the reader nothing. That space is not a mood board. It is the only part you can count on.

I am not arguing for cramming. A dense hero is as bad as an empty one. One idea, stated plainly, with room around it, consistently outperforms both. My rules for the top of an article specifically are in designing the area above the fold.

What belongs in the second screenful?

The proof. With 74 percent of viewing time landing in the first two screenfuls, the second screen is where a skeptical reader decides whether to keep going. Give them the specific evidence that makes your first screen credible rather than a list of features.

Specific means a real outcome in a real situation, not adjectives. A sentence naming what changed for a customer does more work than six icons. If you cannot be specific for confidentiality reasons, describe the situation and the shape of the result honestly.

Order the proof by how skeptical your reader is, not by how proud you are. The strongest evidence goes first even if it is the least glamorous. A plain sentence about a measurable change beats a beautifully set quote that says the team was a pleasure to work with.

What does not belong there is a second call to action competing with the first. Stacking calls to action near the top splits attention at the exact moment you most want it focused. I argued that case in how many calls to action a page should carry.

How do you keep attention past the third screen?

Give the reader a reason to continue that is visible before they commit to scrolling. Descriptive section headings do this better than any animation, because a skimmer can see what is coming next and choose to arrive there deliberately rather than being carried there by momentum.

Structure beats stimulation. Scroll animations and parallax effects attract the eye and rarely hold it, because they signal decoration rather than information. A clear heading that names a question the reader has is a stronger pull than any movement.

Break long pages into sections that could each stand alone. Readers who arrive from a search result, an AI answer, or a shared link often land mid page. A section that only makes sense after reading the three above it will lose those readers entirely.

And write for the person who is scanning rather than reading, because most people are. The behaviours behind that are in designing for the skim reader.

One practical test I use on long pages is the heading walk. Read only the headings, top to bottom, and see whether they form a coherent argument on their own. If they do, a skimmer gets the whole story for free and a reader gets it twice. If they read like a table of contents for a document nobody wrote, the page has structure without meaning, which is the most common failure on long B2B pages.

Does this apply to AI answer engines reading the page?

Not in the same way, and the difference is useful. A retrieval system does not lose interest partway down. It processes the whole page and pulls whichever passage best answers a question, which means the bottom of a long page can be cited even though few humans read it.

That creates a genuine design tension. Human attention rewards putting everything important at the top. Machine retrieval rewards complete, self contained answers wherever they sit. Fortunately the resolution is the same in both cases, which is clear headings with direct answers underneath them.

What this does not justify is padding a page for machines. A page nobody reads does not earn citations either, because the systems doing the citing are increasingly sensitive to whether a page is actually used. Write for the human first and the structure will serve both.

What do people get wrong when they hear this research?

They conclude that pages should be short. That is the wrong lesson. The research says attention declines with depth, not that depth is bad. Cutting a page in half does not move attention downward, it just removes information the committed reader wanted.

The second mistake is treating one number as universal. These are averages across many pages and many users. A pricing page read by someone ready to buy behaves differently from a homepage read by someone who arrived by accident.

The third is optimising the page while ignoring the entry point. If half your visitors arrive on a blog post rather than the homepage, the hero everyone argued about is not the first thing most people see. Check your analytics before you redesign the top of a page.

There is a fourth mistake that shows up in redesign meetings. Teams treat the fold as a fixed line and design to it, which produces pages that look deliberately truncated on a laptop and absurd on a large monitor. The fold is different on every device, so the useful version of this research is about order and density, not about a pixel measurement you can design against.

What should you do next?

Open your most important long page and ask what a visitor would know if they read only the first two screenfuls. If the answer is thin, move your strongest material up and your atmosphere down. That single reordering usually beats any redesign.

Then check the same page against your entry data. Redesign where people actually land, not where you assume they start. The two are different on almost every site I audit.

If you want a second opinion on whether your page order matches how people read it, send me the link and I will tell you what I would move.