Research Recap
TV Ads and Search Spikes: Toward a Deeper Understanding
Key Takeway
For TV advertisers who seek to maximize consumers' online information gathering, post-ad search spikes offer a causal attribution measure that is responsive, reliable and readily available.
Consumers frequently use television and smartphone simultaneously, especially during the prime time TV viewing hours. One consequence of television/smartphone multitasking is well established: TV ads lead some viewers to search online for the advertised brands. In fact, a small industry has recently developed to help marketers measure the effectiveness of TV ads through post-ad spikes in online searches. However, limited research is publicly available beyond the main effect of TV ads causing search spikes.
In this report, Rex Du, Linli Xu, and Kenneth C. Wilbur seek a deeper understanding of how television advertisers should (or should not) leverage post-ad search spikes when evaluating and purchasing TV ads.
They report two empirical studies of how minute-by-minute TV ad insertions influence minute-by-minute brand searches online. The first is a pilot study analyzing two daily fantasy sports brands over a period of three months. Post-ad search spikes are readily detectable for these brands, even for ads with small audiences. Spillovers to competing brands' search volume are positive and approximately one-fifth as large as the effect on the advertised brand's search volume. After controlling for audience size and brand fixed effects, ad content and media factors explain an incremental 54% of the variation in post-ad search spikes.
The second study is carried out with three pick-up truck brands; it merges brand search data from half a million minutes with audience data for more than 40,000 national TV ad insertions. This main study replicates the results of the pilot study, and further shows that (1) about 75% of incremental search occurs either in the minute the ad begins or the following minute, (2) the moderating effects of media factors are brand-specific, (3) ad audiences who are more likely to buy in the advertised product category produce larger search spikes, (4) brand-focused ads produce larger search spikes than price-focused ads, and (5) search spikes are smaller after more informative ads.
In this report, Rex Du, Linli Xu, and Kenneth C. Wilbur seek a deeper understanding of how television advertisers should (or should not) leverage post-ad search spikes when evaluating and purchasing TV ads.
They report two empirical studies of how minute-by-minute TV ad insertions influence minute-by-minute brand searches online. The first is a pilot study analyzing two daily fantasy sports brands over a period of three months. Post-ad search spikes are readily detectable for these brands, even for ads with small audiences. Spillovers to competing brands' search volume are positive and approximately one-fifth as large as the effect on the advertised brand's search volume. After controlling for audience size and brand fixed effects, ad content and media factors explain an incremental 54% of the variation in post-ad search spikes.
The second study is carried out with three pick-up truck brands; it merges brand search data from half a million minutes with audience data for more than 40,000 national TV ad insertions. This main study replicates the results of the pilot study, and further shows that (1) about 75% of incremental search occurs either in the minute the ad begins or the following minute, (2) the moderating effects of media factors are brand-specific, (3) ad audiences who are more likely to buy in the advertised product category produce larger search spikes, (4) brand-focused ads produce larger search spikes than price-focused ads, and (5) search spikes are smaller after more informative ads.
Put into Practice
Taken together, these findings show that, for TV advertisers who seek to maximize consumers' online information gathering, post-ad search spikes offer a causal attribution measure that is responsive, reliable and readily available.
However, TV advertisers who have additional goals (e.g., information provision, communicating current pricing, converting prospects to purchasers, etc.) should exercise caution in treating post-ad search spikes as the single metric of interest in assessing TV ad effectiveness. In fact, a larger search spike could even be a negative signal in some cases, as it might convey reduced effectiveness in communicating product information.
However, TV advertisers who have additional goals (e.g., information provision, communicating current pricing, converting prospects to purchasers, etc.) should exercise caution in treating post-ad search spikes as the single metric of interest in assessing TV ad effectiveness. In fact, a larger search spike could even be a negative signal in some cases, as it might convey reduced effectiveness in communicating product information.
Related links
Harnessing Marketing Analytics for Business Impact (2017) [Conference report]
Social TV, Advertising, and Sales: Are Social Shows Good for Advertisers?
Beth L. Fossen and David A. Schweidel (2016) [Report]