Measuring the impact of digital out-of-home advertising on purchase decisions: A study at high-traffic urban stations with exposure rate as mediator

Abstract

Digital Out-of-Home (DOOH) refers to a centralized advertising network managed through real-time analytics dashboards and remote control systems, enabling dynamic content delivery and performance tracking using metrics such as impressions, reach, and engagement. Campaign effectiveness increases significantly when content is tailored to nearby audiences, affirming the value of a data driven approach. This study aims to explore the influence of ad frequency, exposure duration, and face detection technology using the YOLOv8 algorithm on audience visual engagement, with exposure rate acting as a mediating variable toward purchase decisions. A qualitative exploratory descriptive approach was adopted to understand audience behavior in high mobility public areas. Data collection was conducted through direct nonparticipant observation at three key transit stations in Jakarta including Sudirman, Manggarai, and Kalibata, combined with documentation from a face detection system capable of identifying gaze orientation and attention span. The YOLOv8 algorithm integrated with lightweight CNN based face recognition was employed to enhance real time tracking accuracy in dynamic environments. Data were analyzed thematically to identify patterns of visual attention and the contextual factors affecting ad engagement. Findings reveal that although DOOH exposure volume is high, actual engagement remains low. This suggests the need for improved evaluation metrics and adaptive strategies to optimize DOOH campaign effectiveness.