Webcam eye tracking for consumer research
See where people actually look on screen — using only the camera already in their laptop or phone. No hardware, no lab, no software for anyone to install.
Webcam-based eye tracking — also called browser-based, online, or remote eye tracking — has reached peer-reviewed laboratory-comparable accuracy. Pagegazer designs, runs, and delivers the study end to end, so you get the findings, not a tool to learn.
What is webcam eye tracking?
Eye tracking traditionally required dedicated infrared cameras, chin rests, and a controlled lab. Webcam eye tracking removes all of that: a participant's ordinary front-facing camera is used to estimate where they are looking on their screen, in their own environment, on the real stimulus they would encounter.
The participant opens a link in their normal browser. A calibration sequence of about a minute adapts the model to their setup, and the study begins. Raw camera images and video are processed on the device and never transmitted — only the derived gaze signals leave the browser.
How it works
Modern webcam eye tracking is a four-step pipeline running in the browser.
Face detection
Locate the face and eyes in the camera feed.
Feature extraction
Identify pupil position, eye corners, and head pose.
Gaze estimation
Map eye features to screen coordinates with a calibrated model.
Calibration
A brief per-participant sequence accounts for their setup and viewing distance.
For the full technical breakdown, read how webcam eye tracking works — and how we validated it.
The measurement at a glance
The specification of the platform behind every Pagegazer study. You operate none of it.
| Accuracy | 1.4° visual angle (1.3° at screen centre) — about 1.5 cm on a typical screen |
|---|---|
| Precision | 1.1° visual angle |
| Sampling rate | 30–60 Hz, depending on the participant’s webcam |
| Signals captured | Gaze position, scan path, fixations, time-to-first-fixation, dwell time, areas of interest |
| Devices | Any desktop, laptop, tablet or phone with a camera — nothing shipped or installed |
| Privacy | Gaze is computed on the participant’s device; raw camera video is never transmitted |
| Validation | Peer-reviewed against the EyeLink 1000 (Behavior Research Methods, 2023) |
| Your setup | None — Pagegazer handles calibration, quality control and exclusion thresholds |
How accurate is it?
The question that matters is not whether webcam eye tracking works, but how close it gets to the laboratory standard. That has been measured, peer-reviewed, and independently replicated.
Mean gaze error in visual degrees — lower is better. All three figures from Kaduk et al. (2023).
These figures come from a head-to-head validation in which the webcam system that powers Pagegazer studies and an EyeLink 1000 recorded the same participants simultaneously across five tasks. It is roughly three times more accurate than the next-best webcam approach (WebGazer, the common open-source alternative, sits near 4.2°). The authors conclude that webcam eye tracking now performs roughly on par with mobile eye-tracking devices — in the same regime as research hardware, and good enough for the questions most non-laboratory research is trying to answer.
Kaduk, T., Goeke, C., Finger, H., & König, P. (2023). Webcam eye tracking close to laboratory standards: comparing a new webcam-based system and the EyeLink 1000. Behavior Research Methods, 56(5), 5002–5022. doi.org/10.3758/s13428-023-02237-8 (CC-BY 4.0). Independently replicated for reading by Serrano-Carot et al. (2025) and consolidated in a 2025 methodological scoping review.
Compared to the common open-source alternative
"Webcam eye tracking" is often equated with WebGazer. The system behind Pagegazer studies is a different class of method.
| This system | WebGazer | |
|---|---|---|
| Accuracy | ≈ 1.4° visual angle | ≈ 4.17° visual angle |
| Approach | Pre-trained convolutional neural network | Ridge regression |
| Introduced | 2022 | 2016 |
| Peer review | Journal (Behavior Research Methods) | Conference (IJCAI) |
| Built-in fixation detection | Yes | No |
See the full peer-reviewed body of work using the same platform, the methodology page for per-signal detail and honest limitations, or the underlying platform's technical documentation.
What 1.4° means in practice
At a normal viewing distance (about 60 cm from a 24-inch monitor), 1.4° of visual angle is roughly 1.5 cm on screen. That is precise enough to:
- distinguish between adjacent paragraphs of text;
- identify which product in a row is being fixated;
- determine whether a specific button, headline, or call-to-action is noticed;
- measure dwell time on a packaging element, a video-frame region, or a UI control.
It is not precise enough to resolve individual letters within a word, or to do micro-saccade research that an infrared system supports — but that covers the large majority of commercial research questions.
What it captures
The measurements behind the findings you receive — you get the analysed result, not a raw data dump.
Gaze position & scan path
Where on screen each participant looked, in what order, over time.
Fixations
Where gaze settled, how many fixations, and how long each lasted.
Time to first fixation
How quickly an element, product or message was noticed — if at all.
Dwell time per area
Total attention each region received, and the share of people who ever looked.
Combined signals
From the same recording: facial expression, heart rate (rPPG) and behaviour — gaze in context.
What it answers
The same measurement underpins every Pagegazer research area. Pick the question closest to yours.
Digital Experience
Where attention goes on websites, apps, and prototypes — and what to change.
Packaging & Visual
What gets noticed first on shelf, and whether a redesign stands out.
Ad & Video
Moment-by-moment attention and emotional response in creative.
Behavioral & Custom
Decision and attention studies tailored to a specific question.
Webcam vs. hardware eye tracking
The hardware-eye-tracker era did not end — it became one tool among several. A desktop infrared system is still more precise and is the right instrument for sub-letter reading research or single-trial micro-saccade work in a controlled lab.
What changed is the trade-off. Instead of bringing 20 participants into a facility, you can test 100–200 in their own environments, on the real stimulus, with faster turnaround and a setting closer to ecologically valid. For the questions Pagegazer answers — where attention goes, what is noticed and missed, where engagement drops — the browser is now the right instrument. Where a question genuinely needs lab hardware, we say so before the study starts.
How Pagegazer delivers it
Pagegazer is a research practice, not a tool you operate. The webcam eye-tracking system is the measurement platform built by the team behind Labvanced, used by 35,000+ researchers at 1,500+ universities. Pagegazer clients never touch it — we design the study, run data collection, analyse it, and deliver the findings.
See the full engagement process, or tell us your question and we will scope it.
Common questions
What is webcam eye tracking?
Webcam eye tracking estimates where a person is looking on their screen using only the camera built into their laptop or phone — no infrared hardware, no chin rest, and no lab visit. Participants open a link in their normal browser, complete a one-minute calibration, and the study runs in their own environment.
How accurate is webcam eye tracking?
In a 2023 peer-reviewed validation in Behavior Research Methods, the system that powers Pagegazer studies reached 1.4° of visual angle overall (1.3° at central targets), against 0.91° for a research-grade EyeLink 1000 infrared tracker recording simultaneously. At a normal viewing distance that is roughly 1.5 cm on screen — enough to tell which element, product, or paragraph a person is looking at.
Is webcam eye tracking as good as infrared / hardware eye tracking?
Not identical. A desktop infrared system is more precise and remains the right tool for sub-letter or micro-saccade research. But webcam eye tracking now performs roughly on par with mobile eye-tracking devices and resolves the large majority of commercial questions: what is noticed, what is missed, and where attention and engagement drop.
Do participants need to install anything?
No. There is no hardware to ship and no software to install. The participant opens a link in their existing browser. Raw camera video is processed on their device and is never transmitted — only the derived gaze signals leave the browser.
What can webcam eye tracking measure in a study?
Time to first fixation, whether a region was ever looked at, dwell time on a page area or video frame, scan path and order of attention, and how those differ across audiences or design variants — combined, where useful, with behaviour, facial-expression and heart-rate signals from the same recording.
How many participants can you test?
Because there is no lab, sample sizes are large — typically 100–200 participants per study, recruited from the target audience and tested in their own environments, rather than around 20 in a facility.
Has webcam eye tracking been independently validated?
Yes. Beyond the 2023 EyeLink comparison, independent groups have replicated established reading effects with webcams (Serrano-Carot et al., 2025) and consolidated best practice in a 2025 methodological scoping review. The wider peer-reviewed body of work using the same platform is listed under published research.
Have a question webcam eye tracking could answer?
Tell us what you need to decide. We will scope a study around it.
Submit a project enquiry →