Eye tracking

Eye tracking uses infrared cameras to follow the movement and fixation of the eyes, and returns which elements attracted visual attention and in what order, in the survey Melcher BrandãoMelcher BrandãoProposed, with Brazilian colleagues, a method for measuring emotion in the use of digital products.One reference in this work:2025Proposta metodológica para medir emoções em contextos de experiência de uso de produtos di…See in the bibliography → and colleagues carried out in 2025. It is particularly effective for validating visual hierarchy and the effectiveness of critical elements such as buttons and calls to action, and it is the method that directly tests the predictions of exogenous attention that the salience biasSalience biasHighlight one thing per screen. Two is already none.Open the law → describes. It is also the cheapest to justify: a desktop eye tracker costs a fraction of any piece of laboratory-grade equipment, and the same session also delivers the effort measure from the pupil.

There is now a version with no participant at all: plugins that predict a screen’s heat map from a saliency model, with nobody looking at anything. The caveat is a serious one. Those models were trained and validated mostly on photographs of natural scenes, and attention on an interface behaves differently, concentrates in the top left corner, follows reading order and depends on the semantics of the element, distinctions the generic models get wrong (JiangJiangMeasured visual salience in different types of interface, and showed where generic attention models go wrong.One reference in this work:2023UEyes: understanding visual saliency across user interface typesSee in the bibliography → et al., 2023). The top left corner is the corner of people who read from left to right, and the study did not test the opposite. The accuracy advertised by those products was measured in the wrong domain. It serves as cheap triage of hierarchy; it does not replace a human eye when the decision is expensive.