Users Do Not Read Your Website — They Scan It

This article reflects professional analysis and industry research. Individual results vary.

One myth shapes how most websites are designed and written. It says that if the copy is good enough, users will read it. The do users read websites UX myth dies the moment you look at real eye-tracking data. Users do not read websites the way they read books. They scan, they skip, and they hunt for the thing they came for. They leave the moment they cannot find it.

The Myth

The belief that users will read good copy is easy to understand. It flatters both writers and clients. It makes room for long explainers, big feature lists, and hero sections full of paragraphs. It is also wrong. Great copy is needed, but it is not enough. How that copy is designed to be scanned decides how much of it users take in.

The Evidence Against It

The Nielsen Norman Group has run eye-tracking studies on web behavior since the late 1990s. Their findings show up again and again. That covers hundreds of studies and millions of page views. The core finding: users read about 20 to 28 percent of the text on any web page. That number drops even lower on dense pages with weak structure. It also drops on pages that are slow to show why they matter.

The NNG named two main scanning patterns. The F-pattern shows up on text-heavy pages. Users read the first few lines in full, then less and less of each line that follows. That draws a shape like the letter F. The layer cake pattern shows up on pages with clear visual structure. Users jump between headings and bold text, and mostly skip the body paragraphs between them.

Jakob Nielsen's landmark 1997 study "How Users Read on the Web" found that 79 percent of users scan new pages. Only 16 percent read word by word. A 2008 update confirmed that this behavior had not changed. A 2020 NNG report on reading patterns found that mobile users scroll more, yet read even less per page than desktop users. Scanning is not a quirk. It is the default state.

Research by Chartbeat looks at real user behavior across billions of page views. It found that 55 percent of visitors spend fewer than 15 seconds on a page. That is not a failure of copy quality. It shows how people spend attention when they have not committed to reading it all.

What Is Actually True

The insight is not that copy does not matter. It matters enormously. The insight is that copy must be built for scanners, not written for readers. That shift in approach is a big one.

Designing for scanners means clear, descriptive headings that carry meaning on their own. It means leading with the key point instead of building up to it. It means bold text for truly critical phrases, not for decoration. It means short paragraphs, because long ones look like work to a scanner and get skipped. And it means bullet lists for items that would otherwise get buried in prose.

The NNG research also named what it calls "signposts". These are elements that work like navigation inside a page. Strong headings, pull quotes, captions, and bold key phrases all act as signposts. Users who scan them can build an accurate mental model of the page. They do that without reading the full body text. That is the goal. Design the page so a user who reads only the signposts still gets the offer, the evidence, and the action to take.

How to Design for Scanners While Still Serving Readers

The good news is simple. Designing for scanners does not mean dumbing down your content. It means shaping it so both reading modes are served. The scanner gets the key messages from headings, bold phrases, and short topic sentences. The reader gets the full evidence and reasoning in the body text below.

The inverted pyramid comes from journalism. It is the model that works best here. Lead with the news that matters most. Then add the detail that backs it up. Close with background or context. Each paragraph should stand on its own as a statement. Each heading should make sense without the body under it.

For web pages, this means the hero section must state your core offer in a headline and one sentence, not a paragraph. The first screenful must answer the user's most urgent question: what is this, and is it for me? Below that, the page can open up into evidence. But the scanner will never get there if the first screenful fails.

FAQ

**Q: If users only read 20% of my page, should I just write shorter pages?**

A: Not always. Longer pages can still convert well when they are built for scanning. The goal is not fewer words. It is fewer walls of text, clearer signposts, and a structure that lets users find what they need at whatever depth they choose. SEO also benefits from thorough content, so the answer is structure, not cutting.

**Q: Does the F-pattern apply to all pages?**

A: No. NNG research shows the F-pattern turns up most on text-heavy pages with weak structure. Think search results, or articles with no visual order. Strong design changes that. Clear section breaks and images cut up the F-pattern. They spread attention more evenly. Good design changes where users look.

**Q: What is the most important element to optimize if users are scanning?**

A: Your headings. NNG eye-tracking shows that headings get more fixations than body text on almost every page type. If your headings are vague, generic, or clever without being clear, scanners will not grasp your page. Write headings that state the point, not headings that tease it.

---

Is your website losing users before they get to the conversion point? The problem is usually structure, not copy. A Growth Assessment from Through The Glass Creatives shows you where attention drops off. It also shows you what to do next. Book yours at ttgcreatives.com/growth-assessment.

---