Eye tracking for usability and user experience testing

The short answer

Eye tracking in usability and user experience testing records where participants look while they use a product. It shows what draws attention, what gets missed, and in what order people scan a screen. It does not explain why they looked there or whether they understood what they saw. For that reason, gaze data is usually paired with spoken commentary, task results and follow-up questions. If you run a study with a webcam, that commentary is captured by a microphone. The quality and timing of that audio matter as much as the gaze recording.

A note on evidence: no study-specific data, product specifications or measurements were supplied for this guide. The points below rely on widely documented principles of eye tracking and audio capture. They include no figures for accuracy, sample sizes or costs, and none should be inferred.

What eye tracking measures in a usability test

Eye trackers estimate the point on a screen where a person is looking and record how that point moves over time. Researchers then summarize the raw data with a small set of standard concepts.

• Fixations: moments when the eyes stay relatively still on one spot, generally read as attention to that spot.
• Saccades: the rapid jumps between fixations, during which little visual information is taken in.
• Areas of interest: regions of a page, such as a button, menu or price, that the researcher defines in advance so gaze can be counted per region.
• Time to first fixation: how long it takes before a participant first looks at an area of interest.
• Heatmaps: aggregated views showing where attention concentrated across participants.
• Gaze plots or gaze replays: the sequence of fixations for one participant, useful for seeing scanning order.

What eye tracking cannot tell you

Looking is not the same as noticing, reading or understanding. A long fixation can mean interest, but it can also mean confusion. Peripheral vision lets people register elements they never fixate on directly. Heatmaps can also hide individual differences, because averaging many sessions blends very different behaviors into one picture. These limits are why practitioners treat gaze data as one input among several rather than a verdict on a design.

Dedicated eye trackers versus webcam-based tracking

Lab-grade eye trackers typically use infrared illumination and specialized cameras to locate the pupil and corneal reflections. Webcam-based eye tracking estimates gaze from an ordinary camera image using software. That makes remote, unmoderated studies possible on participants' own computers. The trade-off is generally lower and less consistent precision. The table compares the two in general terms; it does not describe any specific product.

Aspect | Dedicated (infrared) eye tracker | Webcam-based eye tracking

Hardware | Specialized device, usually in a lab or shipped to the session | The participant's existing webcam

Precision | Generally higher; better suited to small targets and detailed reading analysis | Generally lower; better suited to larger areas of interest and broad attention patterns

Sensitivity to conditions | Less affected by room lighting | More affected by lighting, camera quality, glasses, head movement and screen distance

Setup | Controlled calibration with a moderator present | Calibration done remotely, often without a moderator

Scale | Smaller, moderated studies | Larger remote samples are practical

In practice, webcam-based tracking is a reasonable fit for questions like whether people notice a banner or which half of a page gets attention. It is a weaker fit for questions that depend on exactly which word or small icon someone fixated.

Capturing speech alongside gaze

Most usability studies ask participants to think aloud, so there is a spoken record of what they expected, what confused them and what they decided. With webcam-based or remote sessions, that speech is often recorded through the webcam's built-in microphone or another input on the participant's computer. Several audio decisions affect whether the recording is usable.

• Choose the input device deliberately. Operating systems and testing or conferencing tools can default to a different microphone than the one you expect. Ask participants to confirm the selected input before the session starts.
• Check levels with a short spoken test. Speech that is too quiet is hard to transcribe. Speech that is too loud clips, meaning the peaks are cut off and distort.
• Reduce room noise and echo. A built-in webcam microphone usually sits farther from the mouth than a headset microphone, so it picks up more of the room relative to the voice.
• Be aware of automatic processing. Conferencing software often applies noise suppression, echo cancellation and automatic gain control. These features help live calls, but they can clip quiet speech or pump background levels up and down in a recording.
• Keep audio and gaze in sync. Commentary is only useful if it lines up with the gaze replay. Recording both in the same tool, or including a clear sync point, makes later analysis easier.
• Cover consent explicitly. Recording a participant's face, gaze and voice is personal data, so participants should know what is captured and how it will be used.

Concurrent versus retrospective think-aloud

Talking while working, known as concurrent think-aloud, gives immediate commentary. It may also change where and how long people look, because describing a task competes with doing it. An alternative is retrospective think-aloud: participants complete the task silently, then watch a replay of their own gaze and explain what they were doing. This keeps the gaze recording closer to natural behavior and moves speech capture to a separate, calmer step.

There is also a possible practical concern with webcam-based tracking. Speaking involves head and facial movement, and these trackers are sensitive to head position. That suggests concurrent think-aloud could reduce tracking stability. This is an inference from how the technology works, not a measured finding supplied for this guide. Treat it as something to test in a pilot session rather than an established effect.

When eye tracking is worth adding

Eye tracking earns its place when the research question is about visual attention. Typical questions include whether people see a key element, what they look at before acting, or how they scan a layout. If the question is about whether people can complete a task or what they think of it, a standard usability test with good audio and screen recording often answers it without gaze data. A short pilot session is the most reliable way to confirm the setup works before a full study. That includes calibration, lighting, microphone selection and audio levels.

Gaps in the available evidence

No sources were supplied comparing tracker accuracy, recommending sample sizes, or describing particular eye-tracking tools or webcam microphones. No US government guidance was supplied either, so this guide does not characterize what official usability resources recommend. Check those details against current, authoritative documentation before relying on them.