ROI of TV advertising is determined by TV spot properties and various contextual factors such as weather conditions, context of the TV program and the target audience. ProSiebenSat.1 tracks an stores a vast amount of such factors. Additionally, ProSiebenSat.1’s proprietary attribution model measures the effect of every TV Spot.
This allows human driven manual hypotheses testing: TV spots with bright colors might work better in bad weather conditions in the afternoon and romantic comedies potentiall increase consumers desire for a dating platform membership late at night. But due to the underlying complexity (e.g. media budget constraints, an enormous number of possible influencing factors) manual optimization falls short of producing scalable, timely and constantly effective growth of ROI. To fully capitalize on this potential, ProSiebenSat.1 moves from manual to an assisted optimization. Through leveraging state of the art A.I. and optimization tools like contextual enriched Deep Reinforcement Learning, TV spots are assigned to the available advertising slot with the maximal predicted ROI.
Ernst & Young GmbH (2019): Artificial Intelligence in Europe. How 307 Major Companies Benefit from AI. Outlook for 2019 and Beyond. Online verfügbar unter https://info.microsoft.com/rs/157-GQE-382/images/EN-CNTNT-eBook-ArtificialIntelligenceinEurope-SRGCM1059.pdf, zuletzt geprüft am 01.11.2020, S. 22.