Politicians spend a ton of money on advertising, taking up space on television that in a nonelection year might be occupied by, say, a 30-second commercial for a brand of soda. Can all that spending answer a persistent, thorny question about advertising effectiveness?
Chicago Booth’s Sarah Moshary and Bradley Shapiro and Princeton PhD student Jihong Song explored the idea of using national elections to measure the payoff of nonpolitical ads, but find that the effectiveness of these commercial ads remains a mystery.
Television advertising seems straightforward enough: run commercials for soda, and sell more soda as a result. But it’s surprisingly hard to attribute sales patterns to ads. In 2016, more than $4 billion poured into US political TV ads, and the influx of cash drove up the cost of advertising. Moshary, Shapiro, and Song wondered if this would allow them to see how those who were priced out fared.
The researchers analyzed Nielsen Ad Intel data for television ads in 130 markets in the contiguous United States from 2010 to 2016 to learn how presidential and midterm elections affected ad proliferation. They find little change in the overall advertising market—a 10 percent increase in political advertising led to only about a 0.06 percent decrease in total commercial advertising. And that minuscule decline was highly concentrated in just a few local television markets, those in states with very competitive political races.
“Political advertising has a concentrated effect on a few categories that advertise almost exclusively locally; for example, advertising by car dealerships, hospitals, household furnishings outlets, and appliance stores is strongly offset by political advertising, while advertising for national brands in consumer packaged goods is not,” the researchers write.
Television advertising may be considerably less effective than published studies suggest.
Why the Power of TV Advertising Has Been OverstatedWhile politicians do dominate television advertising for months at a time, they don’t edge out many would-be advertisers. And because commercial advertising for national brands was largely unchanged, the effect of political ads was inadequate for measuring the efficacy of advertising for those brands, the researchers find.
“The political ads crowd out a specific set of ads in a specific set of places,” Shapiro says. “While this seems to be a promising idea to evaluate the effectiveness of ads, competitive political campaigns that are large enough to be disruptive are happening in so few places and in so short an amount of time that the ‘experiment’ is underpowered for all but a handful of product categories.”
Moshary, Shapiro, and Song see potential substantive takeaways from their results for future research. In particular, if researchers wanted to look for potential deleterious economic effects of political candidates buying up $4 billion in ad space on businesses, the place to look would be among small local businesses such as car dealerships and restaurants in the most politically competitive markets.
Sarah Moshary, Bradley Shapiro, and Jihong Song, “How and When to Use the Political Cycle to Identify Advertising Effects,” Marketing Science, November 2020.
A company that plugs its past can alienate those who were previously marginalized.
Celebrating Corporate History Can Backfire
Research suggests a lesson for developers who want to convert existing users into paid subscribers.
How Do You Get People to Pay to Use a Mobile App?
Evaluating the performance of machine-learning tools isn’t always easily done.
How (In)accurate Is Machine Learning?Your Privacy
We want to demonstrate our commitment to your privacy. Please review Chicago Booth's privacy notice, which provides information explaining how and why we collect particular information when you visit our website.