Definition:
Eyetracking, or eye tracking, is a technique that records where a person looks and how their gaze moves between different points in a visual stimulus. It can be applied to a screen or a physical environment through devices that estimate eye position and movement.
The data describes visual behavior, but it does not reveal thoughts, intention, or emotions by itself. Valid conclusions require the gaze record to be related to the task, the study design, and other evidence such as responses, actions, or test outcomes.
How eye tracking works
Many systems use infrared light and cameras to locate features of the eye, such as the pupil and corneal reflection. They then calculate a gaze point on a screen or within the participant’s field of view.
- Calibration: The system relates detected eye movements to several known points to adjust the estimate for each participant.
- Recording: The device captures successive gaze samples and associates them with timestamps and coordinates.
- Processing: Software groups or classifies the samples to identify patterns that can be analyzed.
There are stationary screen-based systems, glasses for real-world settings, and webcam-based solutions. The choice depends on the required accuracy, freedom of movement, distance, environment, and sample size. Tobii Pro, EyeLink, and Gazepoint are examples of current hardware and software families.
Gaze metrics
Analysis transforms samples into observable measures. Their meaning depends on the algorithm, recording quality, and how each area of interest has been defined.
- Fixations: Periods when the gaze remains relatively stable around an area; their duration does not automatically equal attention or interest.
- Saccades: Rapid movements between fixations that help reconstruct the path followed by the gaze.
- Visits: Entries, dwell periods, and returns to defined areas, which help compare specific elements of an interface or scene.
Results can be represented through sequences, charts, and a heat map. Pupil dilation and blinking can also be recorded, although they require additional control because they change with lighting, effort, movement, and other conditions.
Applications of eye tracking
Eye tracking is used in research, design, advertising, and neuromarketing. It adds a layer of observation that can complement what people do or report, without replacing a complete assessment of causes and context.
- Digital interfaces: It can show whether important controls, messages, or routes are seen during UX and navigability testing.
- Visual communication: It helps compare how people explore advertisements, creative assets, packaging, or content before effectiveness is attributed to them.
- Physical settings: Wearable eye trackers can record visual paths in stores, professional tasks, learning environments, or product interactions.
On a website, looking at a call to action does not prove that it will be used. Eye tracking may identify a visibility problem, but any effect on the conversion rate must be checked with behavioral and outcome data.
Interpreting eye tracking results
A useful study starts with a defined question and task. Gaze must be interpreted within a study design, not as a direct reading of a participant’s mind.
- Objective: The decision to be supported is specified together with the measure that can answer the question.
- Comparison: Equivalent conditions are maintained when variants, participant groups, or moments are contrasted.
- Triangulation: Eye data is combined with clicks, errors, times, interviews, or other observations.
Aggregated visualizations can hide differences between participants or tasks. To evaluate a design change, it is useful to state a hypothesis and, when appropriate, validate it through an A/B test or another suitable method. Combined evidence is more informative than an isolated map.
Limitations and privacy
Accuracy can be affected by calibration, posture, glasses, contact lenses, reflections, distance, or signal loss. A small or unrepresentative sample also limits the generalization of results.
- Validity: Looking at an element does not prove comprehension, preference, recall, or purchase intent.
- Quality: Data loss and classification criteria should be documented before results are compared.
- Privacy: Collection should have a clear purpose, understandable information, controlled access, and limited retention.
Recordings of a scene, face, or gaze may identify a person or reveal sensitive information about their behavior. A responsible study applies data minimization, consent or another valid basis for the context, and safeguards consistent with applicable rules.