What Eye-Tracking Studies Reveal About How We Read Online
Eye-tracking research shows one thing clearly: people scan before they read. Their eyes look for headings, numbers, bold text, links, and short sections that seem to answer the question they arrived with. Careful reading usually starts only after they decide the page holds something worth their attention.
That behaviour is not evidence of people losing the ability to concentrate. There is a task behind the visit: a price, a definition, a date, a set of instructions, a comparison, or a quick answer. Some aspects of page design make that search easier and some make it harder. Well-built headings and paragraphs direct attention; large undifferentiated blocks of text force readers to build their own route.
What eye tracking actually measures
An eye tracker registers where a person looks while visiting a website. It records fixations, the points where the eyes come to rest, and the fast movements between them. Researchers present the results as heat maps, gaze plots, and similar visualisations.
A five-year literature review by the Nielsen Norman Group about online reading covered five studies completed over 13 years. In total, more than 500 people and more than 750 hours of eye-tracking data were analysed. The authors concluded that general reading habits stayed constant despite changes in websites, monitors, and portable devices.
This research was a desk study based on existing findings about eye tracking and usability of websites written for this article by EssayPro. Special attention should be paid to students' reading habits. Adam Jason, an educationalist at an essay writing service, EssayPro, is particularly interested in it.
For example, a student scans dates, names, headings and evidence, before making a decision whether or not this source will be read carefully. In times of intense studying, a student may say, “Write my essay for me” to manage the writing process and save time for reading and analyzing sources. Specific goals play an important role in such reading processes.
Most people skim before they start reading
Studies from the early days of web usability found that 79% of test participants skimmed a new web page, compared with only 16% who read word by word. That work was done at the very beginning of the internet, but the eye-tracking studies that followed confirmed the same habit: readers look for useful cues first.
Further Nielsen Norman Group research estimated how many words a visitor could actually read during an average visit. On a page of 593 words, a visitor would have time for about 28% of them. Once you account for everything else a visit involves, such as navigation and looking at images and links, a more realistic estimate is around 20%.
These percentages are a general rule rather than a law. A student preparing for an exam will read almost everything. An online shopper checking a delivery cost will read one sentence.
The four main online reading patterns
Heatmaps do not always form the same shape. Researchers have identified several common paths, each tied to a different goal or page design.
| Reading pattern | What the eyes do | When it commonly appears | What content creators should do |
|---|---|---|---|
| F-pattern | Read across the top, move down, read a shorter line, and scan the left side. | Long pages with few headings, bullets, or visual guides. | Put keywords near the start of headings and paragraphs. |
| Spotted pattern | Jump between numbers, links, bold terms, names, and unusual words. | Readers are searching for one fact or answer. | Make important details easy to notice. |
| Layer-cake pattern | Move through headings and subheadings before reading selected sections. | Well-organized articles and guides. | Use clear headings that describe the section below. |
| Commitment pattern | Read most words in order and return to difficult passages. | High motivation, trusted sources, tests, contracts, or important instructions. | Keep the text clear, even when readers are likely to study it closely. |
The famous F-pattern is a warning
The F-pattern is the best-known finding from web eye-tracking work. In left-to-right languages, people usually start near the upper left, scan a few lines of text, move down, scan fewer lines, and continue down the left-hand side.
It does not follow that every page should be designed in the shape of an F. The pattern turns up when people face walls of text with no orientation cues. Attention goes to the first lines and the first words, and text on the right and toward the bottom can go unseen entirely.
The Nielsen Norman Group identified the F-pattern during an updated eye-movement study covering both desktop and mobile screens. That research also found the pattern appearing in right-to-left languages. According to the researchers, F-scanning is what happens when readers try to get information with the least possible effort.
Better design gives the eyes clear stop points: headings, meaningful highlighted text, short passages, and lists.
Headings act as road signs
A heading should tell the reader what to expect beneath it. Catchy headings make an impression, but vague wording leaves the reader guessing.
Compare these two:
- "Confidently Moving Forward"
- "Five Tips for Enhancing Your Reading Comprehension Skills Online"
The first could sit above anything. The second states exactly what follows, so someone looking to improve their reading recognises its relevance immediately.
What this means for the way you write a page
Conclusion
According to eye-movement studies, the online reader is an active decision-maker looking for cues about what is relevant, credible, and worth the effort. People scan because choice on the web is unlimited and most visits have a purpose.
Good pages respect that. They offer clear headings, useful cues, well-defined paragraphs and supportable claims, and they give a motivated reader enough substance to move from scanning to understanding.
People will keep skipping parts of a page. Good writing decides which parts.
Write paragraphs built for scanning
Generate a draft, then tighten it into short sections with headings that say what is underneath.