What’s your project about?

My research is about TV advertising effects. Because advertisers still spend around 25-30% of their advertising budgets on TV advertising (second after online advertising), this is an important topic to study.

What’s the research problem?

These days, people increasingly engage in multiple activities at the same time by using their mobile devices while watching TV. This behavior results in people spontaneously visiting an advertiser’s website or the online shop in response to a TV ad. But how do TV ads affect people’s online shopping behavior, both immediately after the ad airs and in the long-run?

Which solution does your paper bring to the problem?

TV advertising likely increases visits to an advertiser’s website, but people attracted to an advertiser’s website by a TV ad might be less “valuable” for the advertiser in the long-run because they have been nudged to visit the online shop.

How did you study this?

To study the effects of TV advertising on online shopping behavior, we combine a TV ad dataset with an online shopping dataset. But since our data does not tell us whether someone visited the advertiser’s website in response to a TV ad, we develop a novel model to estimate the probability of a website visit being TV-induced. We use the estimated probabilities to study how they affect online shopping behavior over a 90-day window after the visit.

What did you find?

First, we find that TV ads bring more traffic to the online shop of the advertiser, and that the number of new visits depends on the device and channel used to visit the online shop. Second, we find that visits more likely to be TV-induced are less likely to end with a purchase, and when that purchase is made, shopping baskets are usually smaller (compared to purchases resulting from visits that are less likely to be TV-induced). Third, we find that visitors are more likely to be TV-induced and make fewer and smaller purchases in the long-run (over a 90 day-period). 

What can practitioners learn from these results?

We present a novel approach to estimating the probability of being TV-induced since this is impossible for advertisers to observe once the visit occurs.

What we learn is in line with some previous research on the effectiveness of TV advertising: Customers who are more likely to be TV-induced are less valuable to the advertiser, even in the long-run. We argue that they bring less revenue to the advertiser because visits made as a reaction to the TV ad are more spontaneous. Visitors who spontaneously reacted to a TV ad seem to be less loyal customers. These findings can point marketers to rethink their spending on marketing efforts and approach visitors differently depending on what nudged them to visit the advertiser’s online shop.