Political advertisers spent $1.9 billion on online ads for the 2024 election on the four largest digital platforms (Meta, Google, Snap, and X) that publish analyzable spending data, according to a new analysis by the Brennan Center, OpenSecrets, and Wesleyan Media Project. Although this is the most complete accounting of online spending to influence the 2024 elections to date, it is an underestimate since no law requires platforms to publish information about political spending. Some platforms publish no data on this, and the voluntary disclosures of others are unstandardized and likely incomplete.
Our new examination of political ad content in the general election period expands on our summer 2024 and postelection analyses of online ad spending, identifying significant differences in the strategies used by spenders. Parties and outside groups were much more likely than candidates to use negative ads, and their ads focused largely on persuading voters. Candidates’ advertising goals, by contrast, tended to be evenly split between persuading voters and fundraising.
There were partisan differences, too: While both sides of the aisle spent on efforts to persuade voters, spending in favor of Democrats was more likely to have fundraising as a goal, and spending in favor of Republicans was more likely to include get-out-the-vote efforts. Additionally, pro-Democratic spenders put a somewhat greater portion of ad money toward contrasting their party’s candidates with their opponents compared with pro-Republican spenders, who spent more on simply promoting their own candidates.
Understanding political advertising online is difficult due to weak disclosure rules. This analysis necessarily relies on voluntary disclosures by tech companies, which are not standardized and may be significantly incomplete. With political activity continuing to shift online and dark money surging to record highs, the need for greater transparency and stronger regulation is clear.
Spending by Platform
Meta, owner of Facebook and Instagram, was the largest ad seller, with more than $1 billion in reported political ads. Google (which includes YouTube, Search, and third-party advertising) accounted for $846 million. Two other platforms, Snap and X (formerly Twitter), reported far smaller totals of $27 million and $24 million, respectively.
It is highly unlikely that these reported totals represent the full universe of online political ad spending. The quality of the data that platforms voluntarily publish varies greatly. For instance, media investigations have found political ads that platforms failed to include in their ad libraries; one 2023 report led X to release data showing its revenue from political ads to be 50 times higher than it had previously reported. And some online political ads are placed through independent brokers, which means they may not show up at all in a platform’s database. Analyzing multiple platforms together is especially difficult because companies’ definitions of what counts as political spending vary, and they do not share any common method of identifying ad buyers.
Other platforms do not release any information about political spending. Reddit, for instance, only provides ad spending estimates that are in a format that is not searchable or downloadable, preventing analysis. Another platform that does not make any information public is Truth Social, which is partly owned by President Trump — although its overall ad revenue is still too low to make a noticeable difference in this analysis.
No Transparency for Influencer Payments
There is little information about emerging categories of online political spending, such as paid promotion by influencers. These payments are not treated like traditional ad buys and are not required to be disclosed by platforms, advertisers, or the paid influencers. Growing numbers of Americans get their news from social media influencers, which makes this spending an increasingly important avenue for political messaging.
More than a quarter of digital content creators reportedly were approached about promoting political content in the 2024 election as campaigns funneled millions of dollars to influencers. The Democratic National Committee and the Harris campaign, for example, paid more than $4 million to influencer-focused social media firms including Village Marketing Agency, Good Influence, and People First Marketing. Turning Point USA, a pro-Trump youth group, claims to have partnered with hundreds of online content creators throughout the 2024 election. And Trump’s 2020 campaign paid almost $1.8 million to an influencer marketing company. Payments to influencers generally vary based on the size of the account’s following and the level of engagement the post receives. But the amounts can be substantial, ranging from a few thousand dollars per post to $100,000 for influencers with millions of followers.
The increase in paid political influencing creates an avenue for big spenders to try to influence elections anonymously. In the months before the election, for example, a mysterious network of paid influencers engaged in a coordinated attack promoting unfounded sexual allegations against Harris. One participant reportedly made more than $20,000 for their efforts. And in September, the Department of Justice released an indictment alleging that Russian operatives funneled $10 million to Tenet Media, which promoted right-wing content through its network of influencers. As spending from undisclosed sources (known as dark money) continues to shatter records, the lack of transparency around the origin, amount, and purpose of these payments leaves voters in the dark and makes elections more susceptible to manipulation.
Content of Ads
Our first-of-its-kind analysis of 2024 political advertising content on Meta that targeted federal races in the weeks leading to the election (from Labor Day through Election Day) found significant differences in how candidates use online ads compared to other spending entities, as well as partisan differences. (See the Methods section at the end of this analysis for more information on the data and how it was processed.)
Online spending from candidates differed from other entities in several notable ways. In terms of tone, candidates spent the majority (57 percent) of their online ad dollars promoting themselves rather than attacking their opponents. Other entities, by contrast, spent most of their money on ads with negative elements, either attacking or contrasting themselves with their opponents. Ads with negative facets accounted for nearly two-thirds of outside groups’ spending (65 percent) and the majority of spending from national parties (59 percent) and state and local parties (56 percent).
Candidates also tended to pursue different goals than other spenders, focusing much more heavily on fundraising appeals. While candidates spent nearly equal amounts trying to raise money (42 percent) and persuade their audience to vote for them (43 percent), three-quarters of party spending and almost all spending (90 percent) from outside groups went toward persuasion. Get-out-the-vote efforts were a small but significant focus for candidates (14 percent) and state and local party committees (14 percent), but less so for national parties (7 percent) and outside groups (4 percent).
We also found significant differences between the strategies of spenders that we were able to identify as favoring either Democrats or Republicans. Republicans and their allies tended to devote a greater portion of their spending to promoting their candidate (63 percent) than Democratic spenders (53 percent), and Republicans devoted nearly three times as much of their total spending on get-out-the-vote efforts (23 percent vs. 9 percent). Meanwhile, the percentage of spending among Democrats on fundraising appeals was more than twice the Republican percentage (38 percent vs. 15 percent).
Partisan differences also carried over to the tone of ads, with Republicans spending proportionately more to promote their candidates and Democrats spending proportionately more to contrast themselves with their opponents. Nearly two-thirds of Republican spending went to positive, promotional ads compared to just under half of Democratic spending. Meanwhile, Democrats spent 37 percent on contrast ads, while Republicans devoted just 22 percent. Both sides spent roughly the same (relatively small) proportion on attack ads: just 15 percent for Democrats and 14 percent for Republicans.
Due to issues with the acquisition and availability of content data from other platforms, our content analysis is limited to ads on Meta. Moreover, the comparisons we draw between Republican- and Democratic-aligned groups comes with the caveat that their spending was lopsided and driven by the strategic decisions of the presidential campaigns. As with our prior findings, the spending in our content analysis pool shows that Democrats and their allies outspent their Republican counterparts nearly three-to-one, with most of the money coming from groups affiliated with the Harris and Trump campaigns. Still, our findings provide a window to assess how different groups used online advertising in 2024, which may offer valuable insights for future analyses and policies.
Methodology
This analysis relies on political advertising data reported by Google, Meta, X, and Snap between January 1, 2023, and November 9, 2024. As with our prior summer 2024 and postelection analyses, we downloaded data from Google’s weekly Transparency Reports and Meta’s daily Meta Ad Library Reports and used the same methodology to identify political ads and unique spending groups. We similarly downloaded data from X’s Political Ads Disclosures and Snap’s Political Ads Library. But given limitations with this data, particularly difficulties identifying and matching ad sponsors to other known entities, our analysis of political spending on X and Snap is limited to the platform-level totals they report. X’s library contains two types of ads: “political content” ads, which reference a candidate, party, government official, election, referendum, ballot measure, legislation, regulation, directive, or judicial outcome, as well as “political campaigning” ads, which include ads from registered PACs and super PACs, fundraising appeals, and ads that advocate for or against a specific candidate, party, or electoral outcome. Snap’s library includes similar election- and advocacy-related content, plus ads urging people to vote or register to vote and issue ads concerning issues or groups that are the subject of local, national, or global debate or importance.
The analysis of the content of ads focuses on federal election political advertisements on Meta between Labor Day and Election Day 2024. To identify the relevant pool of advertisements, we conduct extensive keyword searches across the Meta Ad Library to identify ad sponsors whose ads (or page names or disclaimers) reference federal candidates. We supplement this keyword approach by analyzing the Meta aggregate report to identify additional sponsors known to air political advertisements on television, and where possible, we match groups’ identifying information with OpenSecrets’ existing campaign finance collections. The entities we identify through these searches collectively account for approximately $400 million in spending. We then process all advertisements (including metadata and video and image content) that appear on these sponsors’ pages, regardless of the funding entity, using automatic speech recognition, optical character recognition, and facial recognition. The pool we identify through this processing — and ultimately include in our content analysis — totals over $250 million and consists of ads from any sponsor, except known downballot candidates or ballot measure groups, that mention or picture a federal candidate or sitting federal officeholder, and ads from federal candidates or national parties, regardless of whether their content features federal candidates.
We assess each advertisement in our content analysis pool on two metrics: tone and goal. The method for categorizing tone depends on the type of ad sponsor. For candidate sponsors, we focus solely on whom the ad features. If the ad mentions or pictures only the sponsoring candidate or neither candidate, we classify it as promotional. If the ad mentions or pictures only the opposing candidate, save for any “paid for by” or “approved by” disclaimers, we classify it as “attack.” And if the ad mentions or pictures both candidates, we classify it as “contrast.”
For non-candidate sponsors, which more regularly include multiple candidates for multiple offices, we conduct aspect-based sentiment analysis to determine the tone of each reference toward each candidate and then roll up multiple references to the same candidate into an aggregate ad tone for each candidate referenced. If the ad features candidates for multiple offices, we first determine which office is the primary focus of the ad based on which candidates, party leaders, and races it identifies and where the ad has the most impressions. Once we define which office is the race of focus, if the ad features both candidates for that particular office, we classify it as “contrast.” If the ad mentions only one candidate, we classify it based on the aggregate tone toward the candidate from our aspect-based sentiment analysis. The method for categorizing goals relies on computational classification constructed from a manually labeled training set.
Authors and Acknowledgements
Ian Vandewalker and Eric Petry
Brennan Center for Justice
Brendan Glavin
OpenSecrets
Erika Franklin Fowler, Michael Franz, Pavel Oleinikov, Breeze Floyd, Travis Ridout, Yujin Kim, and Meiqing Zhang
Wesleyan Media Project
We are grateful to Grady Yuthok Short, Roberto Cordova, Adriana Begolli, Saul Ferholt-Kahn, and Nathan Weisbrod for their contributions to this report.
