When to Play Your Advertisement? Optimal Insertion Policy of Behavioral Advertisement

Digital advertisements offer a full spectrum of behavioral customization for timing and content capabilities. The existing research in display advertising has predominantly concentrated on the content of advertising; however, our focus is on optimizing the timing of display advertising. In practice, users are constantly adjusting their engagement with content as they process new information continuously. The recent development of emotional tracking and wearable technologies allows platforms to monitor the user’s engagement in real time. The proposed optimal policy regarding the timing of behavioral advertising is based on a threshold policy with a trigger threshold and target level. Analogous to the familiar idea of “price discrimination,” the methods we propose in this study allow the platforms to maximize their revenue by “discriminatory” customization of the timing and length of the advertisement based on the behavior of individual users. Finally, we quantify the benefits of the proposed policy by comparing it with the practically prevalent policies (i.e., preroll, midroll, and a mix of the two) through a simulation study. Our results reveal that, for a wide range of settings, the proposed policy not only significantly increases the platform’s profitability but also improves the completion rate at which consumers finish viewing the advertisement. Digital advertisements offer a full spectrum of behavioral customization for timing and content capabilities. The existing research in display advertising has predominantly concentrated on the content of advertising; however, our focus is on optimizing the timing of display advertising. In practice, users are constantly adjusting their engagement with content as they process new information continuously. The recent development of emotional tracking and wearable technologies allows platforms to monitor the user’s engagement in real time. We model the user’s continuous engagement process through a Brownian motion. The proposed optimal policy regarding the timing of behavioral advertising is based on a threshold policy with a trigger threshold and target level. Specifically, the platform should insert the advertisement when the user’s engagement level reaches the trigger threshold, and the length of the advertisement should let the user’s engagement level drop to the target level. Analogous to the familiar idea of “price discrimination,” the methods we propose in this study allow the platforms to maximize their re