Online Political Spending in 2024

Political advertisers in the United States spent more than $619 million on the two largest digital ad platforms between the beginning of 2023 and the end of August of this year, according to a first-of-its-kind analysis jointly conducted by the Brennan Center, OpenSecrets, and the Wesleyan Media Project. Using publicly available data from Google and Meta (owner of Facebook and Instagram) and a transparent methodology, our analysis sheds light on the scope of online spending this election cycle on the two largest online platforms.

As the internet’s importance in American life has risen, much of our political activity has moved online. So too have efforts by hostile foreign governments and other bad actors sowing political disinformation to exploit divisions in the U.S. electorate. Examples of this type of interference include the Russian government’s $10 million Tenet Media scheme to build a network of right-wing influencers on YouTube and other media, as well as its similar effort earlier this summer to influence European elections through covertly sponsored Facebook advertisements.

Despite the importance of online spending in elections, existing transparency rules are extremely weak. As a result, the public data are limited and inconsistent, making it difficult to identify many advertisers and impossible to know the sources of much of the spending. In part because of the difficulty of analyzing and combining the available data, we have focused on Google and Meta. Although they are the largest digital ad platforms in the United States, this means that the numbers reported here represent only a portion of the total universe of online political spending.

• In the 2024 election cycle through the end of August, advertisers who spent at least $5,000 on Google and Meta bought $619 million in political ads, with at least $248 million focused on the presidential race.
• Of spenders on Meta and Google known to advertise in support of a party, spending in favor of Democrats was more than three times the amount of spending for Republicans. Some Republican-aligned spending may have migrated to other platforms not included in this analysis, in part due to lack of public data.
• Almost half the total, $281 million, came from spenders that may hide some or all of their donors.

Ultimately, even with the limited data available, it is clear that online spending is playing a significant and growing role in politics this election cycle, a trend that will almost certainly continue. More than a decade into the era of social media, the need to update campaign transparency rules to take new modes of political communication into account has never been more apparent.

So far in the 2024 election cycle, candidates, parties, and other groups have spent more than $619,090,533 on digital advertising concerning the election and political issues on the nation’s two largest online platforms, Google (which includes YouTube, Search, and third-party advertising) and Meta. Together they account for almost half of the total digital ad market in the United States, but there is not sufficient publicly available data to determine what percentage of the political ad market they have captured.

Even this total for Google and Meta understates the amount of political ad spending on their platforms. For example, neither platform reports expenses for making ads, like research and production, nor payments for influencer endorsements. And due to the volume of advertising activity, we limit our analysis to advertisers who have spent at least $5,000.

Google and Meta also define “political” differently, and thus the scope of the ads included in their libraries differs. Google includes ads that feature elected officials, candidates, parties, and other things the public can vote on. Meta’s definition is broader, including ads by or about candidates, parties, and political figures, as well as social issues.

Despite these data limitations, our analysis provides important insights on digital political advertising in U.S. elections.

The bulk of the spending concerns the presidential election, according to our analysis identifying spenders known to focus on one candidate or contest. Of the ad buys that we know target a specific office, 64 percent, or $248 million, were about presidential candidates. Spending focused on the Senate accounted for 19 percent, or $73 million. Another $232 million targeted multiple races, discussed political issues rather than elections, or otherwise could not be categorized.

The biggest targets outside the presidential contest include Senate races in Arizona, California, Montana, and Texas, as well as a California ballot measure on homelessness and the 2023 Ohio abortion rights initiative. The table below breaks down the top 20 candidates and ballot measures in terms of spending. In most cases, this spending is done by the candidate’s own campaign committee, but for some — especially the presidential candidates — it includes spending by other known groups supporting the candidate. Since Vice President Kamala Harris took over President Joe Biden’s campaign committees in late July, some pro-Biden spending is counted in Harris’s total.

Across the two digital platforms we surveyed, Democrats and their allies spent more than three times their Republican and Republican-aligned counterparts on federal races. The chart below combines spending by candidates, parties, and outside groups that are known to spend in favor of one party. The pro-Democratic spending advantage at the presidential level on these two platforms is in stark contrast to past elections, when Donald Trump relied heavily on digital spending on widely used platforms, such as Facebook. Some of this spending may have migrated to other platforms.

The origins of much of the money funding online political advertising are not publicly known. As with other forms of electioneering, spenders can use “dark money” nonprofit organizations to keep their donors hidden. But there is even less transparency online than in other media due to a lack of regulation.

Of the $619 million in spending in our sample, 45 percent came from spenders that may conceal some or all of their donors. Nonprofits that typically don’t disclose any of their donors spent $106 million, while $65 million is “gray money” from political committees that can partially conceal their donors.

An example of “gray money” spending is Americans for Prosperity Action, one of the top online spenders with more than $7 million in ad buys. As a super PAC, it is required to report its donors, but one of its biggest donors is an affiliated dark money group that contributed $25 million from unknown sources.

Another $109 million came from groups whose form of organization is unknown. It is likely that the great majority of the almost 2,700 groups in this set are dark money spenders, but the data available from the platforms does not allow them to be readily categorized.

The remaining spending in our sample, $339 million, was spent by campaigns, parties, and other groups that are required by state or federal law to fully disclose their donors.

Our ability to analyze online political spending is limited in two important ways: Transparency rules in federal campaign finance law leave much online activity unregulated, and no regulations require platforms to publish data about political ads. Both issues could be addressed by Congress.

Federal law requires mass media ads that mention candidates in the weeks before an election to be disclosed as “electioneering communications.” But the decades-old list of media covered by the rule does not contemplate today’s internet, allowing online spenders to influence elections without disclosing their spending or their donors to the public. The law should be updated to make online ads as transparent as those on television or radio.

Because there are few regulations governing the sale of online political ads, whether and how platforms collect and publish political ad data is left to their discretion. The platforms’ voluntary disclosure policies vary from each other and change over time. For example, Truth Social, which caters to conservatives, does not disclose ad data. And after it changed ownership, X (formerly Twitter) reversed a previous ban on political ad sales. The platform announced last year that it would share data on political ads, but it initially omitted most ads from its disclosures. As already noted, Meta and Google, which do disclose political ad data, define what constitutes a political ad differently. Moreover, studies have cast doubt on the accuracy and completeness of platforms’ political ad data. A new law that required major platforms to consistently disclose certain information about their political ad sales would improve the public’s understanding of attempts to influence elections online.

This analysis relies on political advertising data self-reported by Google and Meta between January 1, 2023, and August 31, 2024, which we downloaded from Google’s weekly Transparency Reports and Meta’s daily Meta Ad Library Reports.

To identify unique spending groups among the data, we match the ad sponsor information available from Google and Meta to advertiser information from the Wesleyan Media Project and existing database information from the Federal Election Commission using OpenSecrets’s extensive campaign finance collections that include state and federal candidate databases as well as additional information on outside spenders, industry, disclosure status, and organizational and individual donors. We then merge the results from each platform to calculate each group’s total online spending.

Given the large volume of advertisements, we limit our analysis to U.S. advertisers that have spent at least $5,000 this election cycle.

We also confine our analysis to groups whose spending is primarily election related. Google and Meta define political spending differently, but both definitions can capture ads that may not be intended to influence elections. Google’s Ads Transparency Center policies apply to ads that “feature” current officeholders and candidates, political parties, ballot measures, initiatives, and proposals. Meta’s Ad Library similarly tracks ads “made by, on behalf of or about” candidates, political figures, political parties, political committees, and individuals who advocate for a particular election outcome, as well as ads “about” an election, referendum, ballot initiative, or “any social issue.”

As a result, the reports from both platforms include some commercial ads that merely mention candidates or political issues but not for an election-related purpose, such as those from news media and commercial sources. To be sure, some commercial advertisements may simultaneously have the intent or effect of influencing the election. For example, three retailers have combined to spend almost $10 million advertising clothing with messages supportive of former president Donald Trump. But because these advertisements are primarily commercial, we code them as such and exclude them from our analysis when we conduct our entity identification and matching process.

Finally, we take a conservative approach when coding groups for characteristics like target and disclosure status. Some groups spend money in multiple races across state and federal elections. Determining an ad’s target therefore often requires an extensive analysis of its content, which is not possible on a pre-election time frame. For this analysis, we assign a candidate or issue target only for groups that are known from the existing databases described above and groups whose name indicates clearly that their advertisements target a specific race. For disclosure status, we similarly rely on the existing databases of known groups to assign spenders to forms of organization, such as candidate, super PAC, 501(c) nonprofit, etc. This process results in a pool of spending that we cannot yet identify, as noted in the text.

Ian Vandewalker and Eric Petry
Brennan Center for Justice

Brendan Glavin and Olivia Buckley
OpenSecrets

Erika Franklin Fowler, Michael Franz, Pavel Oleinikov, Breeze Floyd, and Travis Ridout
Wesleyan Media Project

We are grateful to Josh Bell, Erin Geiger Smith, Grady Yuthok Short, Roberto Cordova, Yujin Kim, Meiqing Zhang, Katherine (Katie) LoCascio, Emmett Perry, and Nathan Weisbrod for their contributions to this report.