MM2RTB: Bringing Multimedia Metrics to Real-Time Bidding
This addresses the problem of enhancing user ad experiences in display advertising for stakeholders like advertisers, publishers, and users, but it appears incremental as it adapts existing metrics to a known system.
The paper tackles the problem of poor user ad experiences in real-time bidding by proposing a framework that incorporates multimedia metrics like contextual relevance and visual saliency, showing that advertiser and user benefits can be significantly improved if the publisher slightly sacrifices short-term revenue.
In display advertising, users' online ad experiences are important for the advertising effectiveness. However, users have not been well accommodated in real-time bidding (RTB). This further influences their site visits and perception of the displayed banner ads. In this paper, we propose a novel computational framework which brings multimedia metrics, like the contextual relevance, the visual saliency and the ad memorability into RTB to improve the users' ad experiences as well as maintain the benefits of the publisher and the advertiser. We aim at developing a vigorous ecosystem by optimizing the trade-offs among all stakeholders. The framework considers the scenario of a webpage with multiple ad slots. Our experimental results show that the benefits of the advertiser and the user can be significantly improved if the publisher would slightly sacrifice his short-term revenue. The improved benefits will increase the advertising requests (demand) and the site visits (supply), which can further boost the publisher's revenue in the long run.