Can computer vision cure display blindness? An investigation into the impact of tailoring advertisements to demographic attributes of passers-by
(A2021-93471)

AUTHORS

Saar Bossuyt, University College Leuven-Limburg; Floris De Feyter, KU Leuven; Kathleen Custers, University College Leuven Limburg

ABSTRACT

Display blindness, or the lack of attention for digital advertising displays due to low relevance expectations, is a major issue in the growing digital signage market. In this paper, we explore if recent developments in computer vision algorithms that allow to detect and classify demographic attributes of people passing by digital displays increase the relevance and hence the impact of digital signage. We set up two mobile eye-tracking studies to examine if digital advertisements tailored to the passerby’s gender (Study 1; N=117) or age (Study 2; N=181) are considered more relevant, receive more visual attention, and are better recalled. Both studies yielded similar results: while advertisements congruent with the participants’ demographic attributes were considered significantly more relevant than incongruent advertisements, this increased relevance was not translated into considerably more visual attention or better recall.

REFERENCES

We are thankful for Hilde Walravens, who helped to carry out both studies.