The average (non-FiveThirtyEight-reading) American’s main exposure to political campaigns probably comes from TV. Since the days of Dwight Eisenhower, television ads have been a central part of our elections; by 2016, campaigns and organizations were spending $2.4 billion to put a total of 3.3 million broadcast TV ads in front of Americans’ bloodshot eyes. These ads can stir up viewers’ emotions, educate low-information voters, manipulate attitudes on race and gender, and — however modestly — persuade voters and win elections.
In short, they are important enough that — along with national polls, state polls, endorsements and fundraising — FiveThirtyEight is now tracking television advertising data for the 2020 presidential election. Our newest elections tool, which uses data from Kantar Media’s Campaign Media Analysis Group, lets you watch every ad that a 2020 candidate has aired on national cable or in any broadcast media market in the country. For each ad, we also have data on how many times it played, the dates and states where it aired, and how much a campaign or group spent to broadcast it. You can also filter the ads to see only those paid for by a specific candidate, those that aired in a specific state1 or those that deal with a specific issue, such as health care or guns.
Despite the hit our productivity might take if we get sucked into a political-ad-watching black hole, we’re excited about this data’s potential to help us understand the 2020 campaign from yet another angle. Which Democratic primary candidates are going all in on which early states? Now their television ad-buying patterns can give us another clue. What issues are driving the conversation of the campaign? These TV ads will show us not only what candidates are talking about, but also how they are talking about it — what stances they take, what framing they use, and so on.
The ads also tell us something in aggregate, like when individual candidates are making their move and how the campaign overall ebbs and flows.
So far, TV is telling a simple tale. Up until July, the airwaves were pretty quiet. Then philanthropist Tom Steyer jumped into the race. Saying he planned to spend $100 million of his own money on his campaign, he has aired at least 1,000 spots per week ever since, dwarfing the volume put out by other candidates. As of Sept. 26, Steyer had accounted for $12.0 million of the estimated $15.4 million that had been spent on TV ads in the presidential race.
Steyer has been focusing his ads on the first four states to vote in the Democratic primary: He’s aired 15,095 spots in Iowa, 4,588 spots in New Hampshire and Massachusetts (the Boston media market also covers much of New Hampshire), 6,943 spots in Nevada and 10,442 spots in South Carolina.2 And this campaign strategy appears to have paid off, as he does better in early-state polls than he does nationally. For example, according to Morning Consult’s latest weekly tracking poll of the Democratic primary, Steyer has 6 percent support across the four early states but just 1 percent support nationwide. And all seven of his qualifying polls for the October debate (those in which he got 2 percent support or higher) are from early states as well. Arguably, Steyer has his aggressive TV strategy to thank for making the debate stage.
The other candidates have not yet started seriously spending on TV. To date, most candidates have been committed more resources to Facebook and Google ads than to television ads (Pete Buttigieg, for example, has spent $5.3 million on digital vs. just $302,200 on TV). After Steyer, the active candidate who has spent the most on TV is Joe Biden, who has aired 882 spots for an estimated $384,220, almost all of it in Iowa.3
But that’s likely to change. In the 2016 cycle, campaign spending really started to ramp up in fall 2015; the same pattern could emerge this cycle too. Just this week, in fact, Elizabeth Warren announced plans for a $10 million TV and digital ad blitz in the early states. So keep an eye on our tracker, which is updated every weekday, to see the television primary evolve in real time.