Areas of Interest in Eye Tracking: What They Are and How They Work

The short answer

An area of interest (AOI) is a region of a visual stimulus, such as a button, headline, face or product image, that a researcher marks so gaze data can be summarized for that region. The eye tracker records where a person looks, and the AOI turns those raw gaze points into results such as whether people looked at the region, how soon, how long and how often. An AOI is an analysis choice, not something the eye tracker detects by itself, so results depend on how carefully each region is drawn. AOIs are also only meaningful when the tracker is accurate enough for their size, which matters most with lower-precision setups such as webcam-based eye tracking.

About the evidence: no product-specific or study-specific data was supplied for this article. It therefore gives no figures for accuracy, recommended margins or software features. What follows describes widely documented principles of AOI analysis. Check the documentation for your own tracker and analysis software for specific values.

How an AOI differs from raw gaze data

Raw eye-tracking output is a stream of gaze coordinates over time. Software usually groups these coordinates into fixations, which are brief pauses where the eyes stay relatively still, and saccades, which are the fast movements between fixations. On their own, these coordinates describe where on the screen or scene the eyes pointed. An AOI adds meaning by tying coordinates to content. For example, a fixation at a certain screen position becomes a fixation on the price label. This is why AOIs are the usual way to answer design and research questions, while heatmaps and gaze plots mainly offer a visual overview.

Common AOI metrics

Metric | What it describes

Time to first fixation | How long after the stimulus appears the first fixation lands inside the AOI

Fixation count | How many fixations fall inside the AOI

Dwell time (total fixation duration) | The total time gaze stays within the AOI

Visits and revisits | How many separate times gaze enters the AOI, which can show returning attention

Hit ratio or proportion of viewers | The share of participants who looked at the AOI at all

Transitions | How gaze moves between AOIs, often summarized as a transition matrix or sequence

Names and exact definitions differ between software packages. For example, some tools count a visit from the first gaze sample in the AOI, while others count from the first fixation. Confirm the definitions in your tool before comparing results across studies or platforms.

Static and dynamic AOIs

Static AOIs stay fixed on a still image or an unchanging screen layout. Dynamic AOIs follow content that moves or changes, such as a person in a video, a scrolling web page or an object in a scene recorded with wearable glasses. Dynamic AOIs usually need keyframes, object tracking or manual frame-by-frame mapping, so they take more effort to set up and check. When the content scrolls or changes, a fixed screen region can silently start measuring the wrong thing.

Defining AOIs well

• Decide AOIs before collecting data where possible. Drawing them after seeing heatmaps can bias the results toward what already looks interesting.
• Match AOI size to tracker accuracy. If a region is smaller than the tracker's typical error, gaze meant for it may land outside it, and gaze near it may count wrongly.
• Consider a margin around small targets, and apply the same margin to every AOI you compare.
• Avoid unintended overlaps. If AOIs overlap, decide in advance how a fixation in both is counted.
• Keep comparable AOIs similar in size where possible. Larger regions naturally collect more gaze.
• Track gaze that falls outside all AOIs as well. Time spent elsewhere is useful context.
• Use consistent names so AOIs can be grouped and compared across stimuli and participants.

Why accuracy matters more with webcam-based eye tracking

Webcam-based eye tracking estimates gaze from an ordinary camera image rather than dedicated infrared hardware. It is generally described as less accurate and less precise than research-grade trackers, and it is more sensitive to lighting, head movement and camera placement. In practice, this tends to favor fewer, larger and well-separated AOIs over many small ones. Treat fine-grained comparisons, such as between adjacent words or small icons, with caution. No measured accuracy values were supplied here, so check the stated accuracy of your specific system and, if possible, run a validation step during each session.

Where audio capture fits in

Audio does not define or measure AOIs. It becomes relevant when gaze studies also record speech, for example in think-aloud sessions where participants describe what they are doing, or in research that links spoken references to where a person was looking. In these setups, the useful step is making sure the audio and the gaze data share a common timeline, so a spoken comment can be matched to the AOI being viewed at that moment. If the webcam's built-in microphone is used, make sure the recording software captures audio from the intended input device and that speech is clear enough to transcribe. A webcam microphone placed farther from the speaker usually picks up more room noise than a close microphone. No evidence was supplied on how specific webcam microphones or eye-tracking platforms handle this synchronization, so confirm it in your software's documentation.

Practical next step

List the questions your study or design review needs to answer. Draw one AOI for each element those questions mention, and size the AOIs to suit your tracker's accuracy. Then confirm how your software defines each metric before you interpret the results. If you record speech alongside gaze, test the audio input and timeline alignment in a short pilot session before full data collection.