When to Combine Television With Online Campaigns

ABSTRACT This work investigates under what circumstances a television campaign should be complemented with online advertising to increase combined reach. The authors first proposed probabilistic models to derive necessary and sufficient optimality conditions for the best media mix. They then relied on roughly 26,000 television campaigns to train classification models to decide whether a campaign should add online advertising. Linear and support vector regression models are used to predict optimal budget allocation, cost savings, and additional reach. The resulting meta-study yields simple, interpretable, and actionable rules that improve the understanding of media-mix advertising.