Predicting Adoption of Ad-Blocking Technologies for Ad Ecosystem Response Strategies


Speaker


Abstract

Digital advertising has been steadily growing and this trajectory is expected to continue for years to come as both consumers and advertisers have been increasingly shifting their focus to the online world. Nonetheless, a potential threat to the online advertising ecosystem has recently emerged: ad-blockers. Prior research has already discovered adverse effects of ad-blockers on publishers, advertisers, and consumers. However, our understanding of the possibility of proactively predicting and managing ad-blocker adoption decisions of online consumers to avoid such adverse effects remains limited. In this study, we use a combination of datasets to address this important question and employ a wide variety of data-science approaches. We find that using state-of-the-art meta-modeling machine-learning methods can achieve high generalization performance in predicting ad-blocker adoption decisions of online users (i.e., 89% in terms of AUC). This suggests, for instance, that platforms can proactively —rather than reactively — manage or even prevent online users’ ad-avoidance strategy of installing ad-blocking software. To further foster accountability and fairness in the ecosystem, our analyses also uncover the most influential website visits in predicting individual ad-blocker adoptions, providing additional insights into the ecosystem regarding irresponsible publishers. Additionally, to reinforce responsibility, we also examine various preventive interventions that ad networks and browser developers could potentially deploy to proactively manage likely ad-blocker adoptions and compare their economic effectiveness. The findings highlight the promising potential of personalized intervention policies, which could have an annual impact of up to $15.8 billion. This study offers important and timely implications for firms and suggests several avenues for future research to further promote social responsibility and participation using data science for advertising management.

This seminar will take place online. See below to join:
https://eur-nl.zoom.us/j/96886971957