011 min
When a visual anchor is seen but not read
In 2007, Nielsen Norman Group ran an eye-tracking test on the U.S. Census Bureau homepage. The task was simple: find the current population of the United States. The answer was already on the homepage, in large red numbers under the label "Population Clock".
Most people still missed it. In that test, 86% of users did not find the population, even though the homepage showed it in large red numbers. Only 14% of the participants used the Population Clock to answer the question.
The eye-tracking heatmap, a map of where people's eyes rested and for how long, shows what went wrong. Most people did look at the Population Clock. But they looked only at the first third of the number and moved on. They never read the rest of the number, so they never noticed that it was the answer.
That is the job a visual anchor is supposed to do, and here it failed halfway. A visual anchor is the place on a page where the eye lands first. From there, the reader decides what the page is about and where to look next. The Census number got the first look. It did not get read. The rest of this article follows that homepage to show what makes an anchor work.
023 min
How a visual anchor gets the first look
The Census number got that first look for the reasons designers usually rely on. Two kinds of signals decide which element becomes the anchor.
What stands out on its own
The first signal comes from the page itself. An element that differs from everything around it gets noticed before the reader decides to look for it. These differences are called preattentive attributes: properties the eye picks up in a fraction of a second, before focused attention starts. The ones that matter most for anchors are:
- Size: the largest text or image on the page.
- Contrast: a colour or a brightness that differs from the rest.
- Type weight: bold or heavy type among regular type.
- Isolation: an element with empty space around it.
- Motion: anything that moves on a still page.
What the reader expects
The second signal comes from the reader. People have learned where useful things usually sit. The logo is in the top left, the navigation runs across the top, and the main content is in the middle. Readers also arrive with a task, so their eyes go first to the places where past pages put the answer. When the two signals agree, the anchor is strong. When they disagree, what people expect can win over what stands out.
Anchors tell the eye where to look next
A good anchor does more than get the first look. It also tells people where related things are. Research on real rooms shows this clearly. Vision scientists describe a room as groups of objects, each built around one large anchor object. In a bathroom, the sink is an anchor, and the soap, mirror and small cabinet sit in fixed places around it.
In 2018, Sage Boettcher, Dejan Draschkow, Eric Dienhart and Melissa Võ tested what anchors do during search. Participants looked through photos of rooms for a named object while an eye tracker recorded their gaze. In half of the scenes, the anchor that belongs next to that object was swapped for a different object that still fit the room. With the right anchor in place, people found the target faster, searched less of the scene and moved their eyes from the anchor to the target sooner.
Anchors tell people what a place is for
A 2025 study from Võ's lab added a second job. Participants judged what action a room was for, such as brushing teeth in a bathroom. When the anchor linked to that action was removed, people understood the scene more slowly. An anchor helps people find things, and it also tells them what the place is for.
Pages work the same way. A heading tells the reader what the paragraph under it covers. A product photo tells them that the price and the buy button are close by. An anchor in the wrong place, or with the wrong label, sends the eye to the wrong part of the page.
031 min
Where the idea of a visual anchor comes from
Nobody coined "visual anchor" in one paper. It is practitioner convention, not a research finding. Painters, poster designers and newspaper editors built their work around one main element that the eye meets first. They called it the focal point or the entry point, and web designers took the habit with them.
The first attempt to turn the habit into a rule for web pages came from Peter Faraday, at a conference on human factors and the web. In the list that Djamasbi and colleagues later summarised from Faraday's 2000 paper, motion comes first: animated elements draw attention before any other element. Size, images, colour, text style and position follow it. For years, designers used this list to predict which element would become the anchor. Later eye-tracking work tested the list directly, and it did not hold up well.
042 min
Common kinds of visual anchors
If what people expect matters so much, the next question is which elements they expect to be worth a look. Most pages use a small set of anchors.
| Anchor | What it tells the reader |
|---|---|
| Large headline | What the page is about |
| Logo in the top left | Whose site this is, and the way back to the home page |
| Photo of the product or of real people | What is on offer, and who is behind it |
| Main action button | The one thing to do on this screen |
| Headings and subheadings | Where each topic starts |
| Links and bold words in text | Which words carry the meaning |
| Highlighted plan on a pricing page | Which option the company recommends |
Three of these behave differently from the rest, and eye-tracking studies show how.
Headings as anchors: the layer-cake pattern
Kara Pernice of Nielsen Norman Group described this pattern in 2019. People look mostly at headings and subheadings and skip the body text. When they reach a heading that matches what they want, they read the text under it. In a heatmap this looks like horizontal stripes with empty bands between them, which is why NN/g calls it the layer-cake pattern. It only works when headings look clearly different from body text and say what their section covers.
Links and bold words as anchors: the spotted pattern
In the spotted pattern, the eye jumps between single words spread across the page. People fix on words that are styled differently from the text around them, such as links, coloured words, bold words and items in bulleted lists. Each of these words becomes a small anchor. On the same pages, people often follow the F-pattern too, reading the first lines and then scanning down the left edge.
Photos: real people get looked at, stock photos do not
A photo is not an anchor just because it is large. In 2010, Jakob Nielsen reported that eye-tracking participants ignored big decorative images. They studied photos of products, and photos of real people who worked at the company. Stock photos of generic models got little attention. A photo becomes an anchor when it carries information the reader wants.
052 min
How visual anchors show up in real interfaces
Real products show anchors that work and anchors that fail, often for the same reasons.
One primary action per screen
Android builds an anchor into one of its standard components. Android's documentation describes the Floating Action Button, or FAB, as a high-emphasis button for the primary action in an app. It is a round button that floats above the content, usually in the bottom right corner. The Gmail app for Android uses it for Compose. The button has the strongest colour on the screen and sits apart from the list of messages. Since a screen has only one, it never competes with a second anchor.
Subheadings that do their job
In the layer-cake study, one participant read an article on TechCrunch by scanning its subheadings. Each subheading looked different from the body text and said what its part of the article argued. The participant stopped at the subheading that matched their interest and read only the text under it.
When the label is a made-up name
Back on the Census homepage, NN/g named the main cause of the failure. The page used invented names like "Population Clock" and "Population Finder" instead of plain words. A reader looking for the population of the United States glanced at the anchor and did not see their question in it. The anchor got the look, but its label gave people no reason to read further.
When an anchor looks like an ad
The second cause was the look of the anchor. Users tend to skip areas with heavy formatting because those areas look like ads, so about a third of users never saw the Population Clock at all. This is banner blindness, the habit of ignoring anything that looks like an advertisement. Pernice's 2018 study found that real content is ignored too when it has the traits of an ad, such as animation or a place in the top banner or the right column. People learn to skip those places in the same way they learn to look for the logo in the top left.
062 min
How to choose and build a visual anchor
So a working anchor needs two things: people must look at it, and it must answer what they came for. Say a designer is asked to fix that Census homepage. The page has one main question to answer for most visitors, and the population is one of the top answers. The designer can work through five moves.
Name the page's main question.
Write down what most visitors come to find. The anchor should answer it, or point straight to the answer.
Label it in the reader's words.
U.S. population today says what the number is. Population Clock does not.
Style it like content, not like a promotion.
Use the page's own typeface. A larger size and a heavier weight are enough to make it stand out. A bright box, animation and a decorative frame make it look like an ad.
Put it where people expect content.
Place it in the main column, near the page title. Keep it out of the top banner and the right column, where people expect ads.
Give the rest of the page smaller anchors.
Clear headings and link text let people keep scanning once the first look is over.
Put the meaning in the first words
Those smaller anchors work best when their first words carry the meaning. In Nielsen's 2006 eye-tracking study of 232 users, people scanned down the left side of the text. Their eyes mostly stopped at the first two words of each line and rarely reached the third. A heading that starts with Population of the United States works. One that starts with Explore our latest figures on population puts the meaning too late.
Then test it. The squint test is quick: blur the page, by squinting or with a blur filter, and see which shape stands out first. A five-second test goes further. Show the page to a few people for five seconds, hide it, and ask what they remember. If they name the population and can say the number, the anchor works.
The same moves apply to any page. On a pricing page, the anchor is the plan most buyers should pick. On a sign-up form, it is the first field. On a phone screen, each screenful usually needs its own anchor, because each one shows less at a time.
072 min
When visual anchors backfire
Even a well-built anchor can stop working. These are the common ways it fails, each with a sign you can check in session recordings, heatmaps or support tickets.
- Too many anchors. Three bright buttons, two large headlines and a moving banner share one screen. Nothing comes first, so people look around without a place to start. Sign: first clicks land in many different places.
- Emphasis everywhere. When half the text is bold, bold stops meaning anything. Sign: people skip the bold text as often as the plain text.
- An anchor on decoration. A large mood photo gets the first look and gives nothing back. Sign: a short glance at the image, then a jump to the search box.
- An anchor that looks like an ad. As on the Census homepage. Sign: support requests for information that is already on the page.
- An anchor people have learned to skip. A banner that has sat in the same place for months, and never helped, stops getting looked at. Sign: fixations on that area fall over time.
- An anchor that says the wrong thing. A large Sale badge on a page whose job is support sends people the wrong way. Sign: people click it and come straight back.
Visitors who go straight to search
Some people skip every anchor on purpose. On the Census homepage, 57% of users looked around the page briefly and then went straight to the search box. For these visitors, the search box is the anchor, whatever the layout says. A page that serves them well keeps the search box easy to find, and makes sure a search for the main question returns the same answer the anchor shows.
Visual anchors and screen readers
A visual anchor only works for people who can see the page. Someone using a screen reader, software that reads the page aloud, never gets the first look. The Web Content Accessibility Guidelines cover this in WCAG 2.2 Success Criterion 1.3.1 Info and Relationships. The structure that sighted users see through visual cues has to be kept when the page is read by a screen reader. So the anchor needs a real heading or a clear first position in the reading order, not only a large size or a bright colour.
082 min
Where the evidence on visual anchors is contested
Most advice on anchors assumes that the biggest element wins. The evidence for that is weaker than designers think.
Faraday's list predicts poorly
In 2018, Jeremiah Still tested Faraday's model with eye tracking in two studies. In the first, simple web pages were built so that the model predicted a clear entry point. In the second, real web pages were coded for the entry point the model predicted. In both studies, the model did a poor job of predicting where people looked first. It picked the anchor mainly by size. Position and distinctiveness, meaning how different an element is from its neighbours, predicted the first look better. Still concluded that designers should stop using the model, because its predictions can mislead.
For a designer, this means a size-first habit is not safe. Test where people actually look, instead of trusting a ranked list of properties.
Visual anchors can bias decisions
In data visualization research, "visual anchor" has a second meaning. It describes anchoring, the bias in which people rely too much on the first piece of information they see, in a visual form. In 2019, Ryan Wesslen and colleagues at the University of North Carolina at Charlotte ran a lab experiment with 94 participants. People used a visual analytics tool, software that shows data in several linked charts, to judge which Twitter news accounts spread misinformation. Some groups first watched a training video that showed one way of using the tool. That video, the visual anchor in this study, changed how active, fast and confident people were, and in some conditions how accurate. The lesson carries over to product design. Whatever a dashboard shows first does more than guide the eye. It can also decide which evidence people trust.
091 min
Visual anchors vs. visual hierarchy and focal points
A visual anchor is one element, so it is easy to mix up with ideas that describe the whole page or a different thing entirely.
| Term | What it describes | How it differs from a visual anchor |
|---|---|---|
| DesignVisual hierarchy | ||
| Not in the library yetFocal point | ||
| PsychologySalience | ||
| PsychologyAnchoring bias | ||
| Not in the library yetAnchor link |
The sharpest line is with visual hierarchy. Hierarchy is the whole ranking of a page. The anchor is where that ranking starts. A page can have a clear hierarchy and still a weak anchor, if the top item is something people have learned to skip.
?3 questions
Questions people ask
How is a visual anchor different from a call to action?
Can a page have more than one visual anchor?
Does the visual anchor need to be at the top of the page?
§13 sources
Sources
Show all 13 sourcesHide sources
Jakob Nielsen, "Fancy Formatting, Fancy Words = Looks Like a Promotion = Ignored", Nielsen Norman Group (2007). www.nngroup.com
Kara Pernice, "Banner Blindness Revisited: Users Dodge Ads on Mobile and Desktop", Nielsen Norman Group (2018). www.nngroup.com
Kara Pernice, "The Layer-Cake Pattern of Scanning Content on the Web", Nielsen Norman Group (2019). www.nngroup.com
Kara Pernice, "Text Scanning Patterns: Eyetracking Evidence", Nielsen Norman Group (2019). www.nngroup.com
Jakob Nielsen, "Photos as Web Content", Nielsen Norman Group (2010). www.nngroup.com
Jakob Nielsen, "F-Shaped Pattern For Reading Web Content", Nielsen Norman Group (2006). www.nngroup.com
Sage Boettcher, Dejan Draschkow, Eric Dienhart and Melissa Võ, "Anchoring visual search in scenes: Assessing the role of anchor objects on eye movements during visual search", Journal of Vision (2018). pubmed.ncbi.nlm.nih.gov
Lea Alexandra Müller Karoza, Sandro Luca Wiesmann and Melissa Võ, "The role of anchor objects in scene function understanding", Scientific Reports (2025). pmc.ncbi.nlm.nih.gov
Soussan Djamasbi, Marisa Siegel and Tom Tullis, "Generation Y, web design, and eye tracking", International Journal of Human-Computer Studies (2010), which summarises Faraday's 2000 model. users.wpi.edu
Jeremiah D. Still, "Web page visual hierarchy: Examining Faraday's guidelines for entry points", Computers in Human Behavior (2018). doi.org
Ryan Wesslen and colleagues, "Investigating Effects of Visual Anchors on Decision-Making about Misinformation", Computer Graphics Forum (2019). webpages.uncc.edu
Android Developers, "Floating action button" (Jetpack Compose documentation). developer.android.com
W3C, "Understanding Success Criterion 1.3.1: Info and Relationships", WCAG 2.2. www.w3.org





