Fifteen terms are enough to read any eye tracking report: four describe how the eye moves, five measure attention, three turn it into pictures and three tell you whether the data can be trusted. This guide defines each one in a single sentence, explains its role in a study and flags the most common misreading. It is written for the people who receive the report: marketing, UX and research teams.
How the eye moves
Sharp vision covers a tiny part of the visual field. To read a label or scan an ad, the eye jumps from point to point and pauses at each one. All eye tracking starts from that alternation.
Fixation
A fixation is the interval during which the gaze stays almost still on one point, and it is the moment the brain processes visual information. In a study, the count and duration of fixations show how much reading effort each zone received. Common misreading: treating a long fixation as interest. Long fixations also appear when an element is confusing or hard to read.
Saccade
A saccade is the rapid jump between two fixations, during which the brain suppresses visual perception. It lets researchers reconstruct the order in which a person scanned the stimulus. Common misreading: looking for information inside the saccade itself. Nothing is processed mid-jump; the useful data sits at the start and end points.
Regression
A regression is a backward saccade over content already seen. In reading and in interfaces it signals re-reading, or a search for something missed the first time. Common misreading: treating every regression as a problem. In a long text or a price comparison, going back is a normal part of the task.
Blink
A blink is the brief closing of the eyelids, during which the device loses the pupil signal. It is logged to clean the data and, in some designs, as an indirect indicator of mental load. Common misreading: leaving blinks in the analysis. An unfiltered blink counts as lost attention and distorts the metrics of whichever zone it landed on.
Attention metrics
Metrics are computed over areas the research team defines in advance. Without areas of interest, a report is just pictures.
Area of interest
An area of interest is a region of the stimulus the researcher outlines to measure the attention it receives: the logo, the price, the face, the buy button. It turns gazes into numbers that can be compared across versions and participants. Common misreading: drawing areas that overlap or are too small, so that the device's own error decides which zone each fixation belongs to.
Time to first fixation
Time to first fixation is the interval between the stimulus appearing and the first fixation inside an area of interest. It measures how quickly that element draws the eye. Common misreading: comparing it across participants who never looked at the element. Anyone who never fixated the area has no value and should be reported separately as a reach percentage.
Dwell time
Dwell time is the sum of all fixations and saccades inside an area of interest during exposure. It shows how much total attention that element gathered. Common misreading: reading it without accounting for size. A large area accumulates dwell time by sheer surface; a fair comparison normalizes by size or compares versions of the same layout.
Fixation count
Fixation count is the number of times the gaze stopped inside an area. Together with mean duration, it separates a quick read from a detailed inspection. Common misreading: using it as a preference measure. Many fixations can mean interest or difficulty; the follow-up survey and the task resolve the ambiguity.
Gaze sequence
Gaze sequence is the order in which a person visited the areas of interest. It answers which element was seen first, second and third, a central question for packaging and landing pages. Common misreading: averaging sequences from people with different strategies. Group the patterns and report the most frequent one.
How attention is pictured
Visualizations compress many participants into a single image. They open the analysis and, more often than not, they are the only part that travels afterwards.
Heatmap
A heatmap colors the stimulus by the number or duration of fixations each zone received, aggregated across participants. It is the fastest way to spot the most and least viewed regions. Common misreading: taking it as proof of comprehension or liking. The map shows where the gaze went, and a red zone can be the clearest element in the design or the most confusing one.
Opacity map
An opacity map darkens the zones that received no attention and leaves visible the ones that did. It is the most useful version for creative teams, because it shows the ad exactly as it was seen. Common misreading: tuning the visibility threshold until the image confirms a hypothesis. The threshold is set before the results are seen.
Scanpath
A scanpath draws, for one participant, a circle for each fixation (size shows duration) connected by lines that stand for saccades. It lets you follow a single case in detail. Common misreading: generalizing from one scanpath chosen because it looks good. A scanpath illustrates; the evidence lives in the aggregated metrics.
Data quality
Three terms decide whether everything above deserves trust. Our article on eye tracking accuracy and how it is measured covers them in depth.
Accuracy and precision
Accuracy is the average distance between the point a person actually looks at and the point the device records; precision is the spread of successive measurements of the same point. Accuracy sets the minimum size of an area of interest; precision sets how stable the detected fixations are. Common misreading: using the two words interchangeably. A device can be precise and inaccurate at the same time, with every measurement clustered far from the true point.
Calibration
Calibration is the preliminary step in which the participant looks at known points so the system can fit a model of their eye. Every measurement in the session inherits its accuracy. Common misreading: calibrating once and trusting it. Posture changes, lighting and fatigue degrade it; rigorous studies validate calibration during the session and discard recordings that fall below the required level.
Sampling rate
Sampling rate is how many times per second the device records gaze position. It determines which events can be detected: a high rate resolves short saccades, while a low rate is enough for fixations and heatmaps. Common misreading: demanding lab-grade sampling for a study that only needs attention by zone. Screen-based eye tracking in the lab and eye tracking glasses in real environments serve different needs at different rates.
A 2026 study in the European Journal of Information Systems combined screen-based eye tracking with physiological measures to link attention to charity campaign images with donor behavior. The metrics it relies on are the ones in this guide: areas of interest, fixations and dwell time. Source: Blicher, Gleasure, Constantiou and Clement, doi.org/10.1080/0960085X.2026.2631718.
Frequently asked questions
Which metric should I check first in a report?
The share of participants who fixated each area of interest at all. If few did, every other metric for that zone describes a minority, and the issue is layout before it is content.
Does a heatmap replace per-area metrics?
The map guides and the metrics decide. A map aggregates participants and loses order and variability; per-area metrics let you compare versions with a margin of error.
Do these definitions hold for webcam eye tracking?
Yes. The achievable accuracy changes, and with it the minimum size of areas of interest. The metrics and how to read them stay the same.
An eye tracking study designed with rigor
Remote measures attention with webcam eye tracking, with areas of interest defined before data collection and quality control for every participant. When the question calls for a lab, Databrain's scientific team runs it with screen-based eye tracking or glasses.
Book a meeting