EXPRESS: Dynamic Effects of TV Ad Suspension on Keyword Search: Evidence from the U.S. Telecom Industry

Abstract

This paper investigates the dynamic effects of temporarily suspending TV ads on online keyword search through a large-scale quasi-experiment in the U.S. wireless telecom industry. One top carrier, which spends stably around one million USD per day on TV advertising, halted its TV ads for a randomly selected week over our (6-month) study window. The primary challenges in this event study — continuously changing treatment variables and establishing valid control groups to account for time-varying confounders—are addressed using the generalized synthetic control estimator (GSC). Findings reveal significant immediate and lasting declines in the focal brand’s search volume, with reductions of 5-15% per day persisting for two weeks after TV ads resumed. The effects varied significantly across search categories: informational searches related to TV ad content experienced drops of up to 30% per day but recovered more quickly, while navigational searches showed smaller but more prolonged declines. Furthermore, the intervention had modest, short-lived spillover effects on three competitors. Interestingly, the suspension tended to hurt the competitor most similar to the focal brand, while benefiting leading competitors. These spillover effects were particularly pronounced in informational and channel searches, suggesting that competitors’ visibility was significantly affected by the focal brand’s absence.