1 2 3 4 5 CTA reached 5th, not 2nd eye-tracking · gaze path · heat map · where users actually look · attention before assumption

Eye-Tracking Studies

See exactly where users look — not where they say they look. Eye-tracking records the precise path and duration of visual attention, revealing which elements are actually seen, which are skipped, and in what order.

Visual HierarchyCTA PlacementNavigation DesignLanding PageForm DesignAdvertisement Placement

Two sentences.

Eye-tracking uses hardware or software to record the precise location, duration, and sequence of visual fixations across an interface — producing heat maps, gaze paths, and fixation data that reflect actual visual behaviour rather than self-reported attention. The fundamental value is the gap it closes: interfaces that look balanced in design review are routinely revealed by eye-tracking to direct all attention to a single quadrant, leave key messages unseen, or fail to achieve the intended reading path.

Eye-tracking became more accessible in the 2000s with tools like Tobii producing lab-grade data and webcam-based tools enabling remote testing. Its advantage over recall-based methods is objectivity and granularity: attention measured to within a few pixels, at millisecond resolution, without relying on user recall.

Apply this when…

A specific element is not performing and the team suspects it is not receiving visual attention
A complex screen with competing elements needs attention-sequence evaluation
Two layout alternatives need objective attention comparison
A usability test showed users missing critical elements and you need to know exactly what they looked at instead
Scanning pattern research is needed for a specific user group where standard assumptions may not apply

When NOT to apply it

Skip it when the question is about task completion or comprehension (eye-tracking shows where, not why), when budget constraints rule out lab equipment and webcam accuracy is insufficient, when the interface involves frequent mobile scrolling or movement, or when the team needs findings within days.

The mechanism

Eye-tracking measures infrared light reflection off the cornea to calculate gaze direction at 60-1200 samples per second. Raw data is aggregated into heat maps (fixation density) and gaze plots (fixation sequence), making attention patterns legible.

01
Visual attention changes dramatically with task
Yarbus's 1960s research showed the same image produces completely different gaze patterns depending on the viewer's task. A user navigating to complete a task looks at entirely different elements than one exploring. Eye-tracking data is only interpretable in context of a defined task.
02
Heat maps reveal hierarchy failures invisible to review
The most common finding: elements the team believed were prominent receive almost no fixation, while secondary elements dominate. Designers know what to look for — eye-tracking reveals what users actually look for, which is different.
03
Looking does not mean understanding
Fixation does not guarantee reading, comprehension, or encoding. Users frequently skip words, fixate without processing under cognitive load, and fail to recall fixated content. Combining eye-tracking with comprehension questions or task completion data is necessary.
04
Fixation count, dwell time, time to first fixation, AOI analysis
Fixation count: how many stops at the element. Dwell time: total fixation duration. Time to first fixation: how quickly the element attracts attention. Area of interest analysis: metrics aggregated for defined regions rather than single points.

Heat maps are not proof of success

A hot spot on a CTA means users looked there — not that the design is working. A CTA with extensive fixation but low click-through may be receiving attention because it is confusing, not compelling. Eye-tracking diagnoses attention; it does not diagnose intention, comprehension, or conversion.

NNG's F-pattern research and twenty years of hierarchy guidance

In 2006, Nielsen published eye-tracking research with 232 participants showing users scan web content in an F-pattern: reading the first horizontal band fully, scanning a shorter second band, then moving vertically down the left side. Content in the lower-right quadrant received almost no fixation.

The F-pattern has since been replicated in specific contexts and qualified in others — it reflects poor content design more than inherent scanning behaviour. But it demonstrated eye-tracking's ability to surface systematic patterns across large samples, producing findings specific enough to inform concrete layout decisions. Twenty years of subsequent research represents the method's cumulative value.

Nielsen Norman Group · F-Pattern Research · 2006
Large-sample eye-tracking surfaces systematic scanning patterns
Low attentionFirst horizontal scan — highest density↓ vertical scanF-pattern: first band → second band → left vertical → lower-right dead zoneNNG F-pattern · 232 participants · systematic scanning · hierarchy implications
232 participants → systematic pattern confirmed

Test yourself & see real examples

No examples yet — be the first.

Spotted a layout where a key element is in a visual dead zone — or one where the hierarchy perfectly matches scanning patterns? Submit what you observed.

✓ Reviewed before publishing✓ Your name on every example you submit✓ Violation or fix — both welcome

Seen Eye-Tracking insights ignored in a real product? Help grow the evidence base.

Where teams go wrong

Running without defined tasks or areas of interest. Task-free eye-tracking shows exploratory scanning — where attention goes without a goal. Because task changes gaze patterns dramatically (Yarbus), task-free data cannot be extrapolated to real usage.
Treating heat maps as the complete finding. High fixation on a CTA with low click-through means the button is being looked at but not acted on — a comprehension problem, not visibility. Heat maps must be combined with task completion data and qualitative observation.
Using eye-tracking when cheaper methods would suffice. If the question is "did users click the CTA?" rather than "where did their gaze go before clicking or not clicking?" — five-second testing, first click testing, or usability observation is sufficient.
Overgeneralising findings across contexts. The F-pattern applies to content-heavy pages, not navigational interfaces or mobile screens. Each context requires its own study or an explicit assessment of whether conditions are similar enough to justify the transfer.

Connected ideas

Eye-tracking is the only method providing objective, millisecond-level attention data.

The most important pairing is eye-tracking with usability testing. Eye-tracking within a usability session reveals the specific attention pattern associated with each navigation decision, error, or success. Detached from task context, eye-tracking produces descriptive findings; embedded in usability sessions, it produces diagnostic findings.

Run it right now

⏱ 10 minutes · Solo · No prep

The Blur Test

1. Take a screenshot of your most important screen. Apply a Gaussian blur (8px) until text is illegible but shapes and colour blocks remain.

2. Note which elements are still visually dominant — shapes, colour blocks, contrast regions. These are receiving the most pre-attentive visual weight.

3. Compare dominant elements in the blur to what you intend to be most prominent. If the wrong elements dominate (a decorative image, a background element), your hierarchy has a mismatch eye-tracking would confirm.

4. List three elements that should receive first attention. Check each in the blurred view. Any that disappear into the background are competing against larger, higher-contrast elements — and users are fixating on those competitors first.

10 minutes