The comparative study of traditional and AI eye-tracking: practical utility of ai recommendations and visual attention predictions in digital banner advertising

Abstract

The objective of this study is to clarify the practical utility of AI eye-tracking softwares in predicting visual attention and improving the performance of digital banner advertisements. In order to achieve this objective, the AI eye-tracking software was utilised to predict attention and obtain suggestions for improving the digital banner performance. To clarify the impact of these suggestions, an international digital campaign was executed, employing both non-optimised and optimised banners (A/B testing) . In order to clarify the accuracy of attention prediction, an eye-tracking study was conducted, complemented by semi- structured interviews. The findings of the digital campaign analysis demonstrate that AI design quality metrics are generally not strong predictors of click-through rate, though a few strong correlations were revealed for some countries. The findings of the eye-tracking study indicate that AI predictions exhibited modest alignment with human gaze, yet demonstrated limitations in modelling top-down attention.