REPORTS

2020 Political Digital Advertising Report

How COVID, massive fundraising, and the culture war played out online, up and down the ballot

The Digital Election (...it had to be)

2020 was hailed as a “digital election” – after all, it had to be. With COVID-19 as a driving force, the political battlefield moved from mainstays like door-knocking and in-person events to organic social media, Zoom fundraisers, and, crucially, digital advertising. Political digital advertising spend grew 4.6x from 2018 to 2020, outpacing the growth in federal election spending over the same period, which grew 2.4x. but it is still less than 20% of total spend. In this report, we use digital advertising’s public datasets and our own experiences running digital for hundreds of campaigns to seek to answer critical questions about how COVID-19, massive fundraising, other factors, and the culture war impacted candidates up and down the ballot.


Part 1: The Digital Electoral Landscape

Screen time for Americans exploded in 2020, with 24% growth from 2018 to 2020

The pandemic helped to drive up Americans’ digital screen time to nearly 8 hours/day, according to eMarketer, while traditional media consumption, like television, radio, and print, remained steady. From 2018 to 2020, there was a 24% growth in digital screen time, making it a necessity for political campaigns to shift spend towards where voters were spending their time.

DIGITAL SPEND ACCELERATED BUT REMAINED FAR BELOW THAT OF COMMERCIAL ADVERTISERS

Digital spending as a percentage of total media spend

Campaign culture is still based primarily in offline media, although we saw some change in 2020 with COVID as an accelerant. In 2018, TFC found that incumbent U.S. Senate campaigns only spent 4-7% of their media budgets online. This low media allocation came two years after the Trump 2016 campaign attracted much attention for its online prowess.

In 2020, we saw a three-fold increase in political media spend share, a notable jump, but still far below commercial standards. As a representative set, competitive Senate races spent 20% of their media budgets online, still far below commercial advertisers who spent more than 60% of media budgets online. It’s also worth noting that this was an unprecedented year in terms of dollars raised, so the largest Senate campaigns had extra cash to burn; much went online.

TRUMP LEANED INTO THE ‘CULTURE WAR, ’ SPENDING 5X MORE THAN BIDEN DID ON SIMILAR MESSAGING

Topic splits on Facebook advertising: Trump spent 21.4% of his Facebook on "Culture War" topics versus 4.2% for Biden.

2020 was defined by COVID-19, social justice movements and increased popularity of fringe groups like QAnon. The text of Biden and Trump’s ads don’t reflect this. It wasn’t just a “normal” fight about issues, but instead two entirely different perspectives on the country. Most noticeably, the Trump campaign leaned in much more heavily to creating and perpetuating the “culture war.”

Trump allocated 5x more to culture keywords – radical, conservative, patriot, and socialism – than on topics the right portrayed as Biden’s  equivalents: diversity, equality, racism, black lives matter, hate group, qanon, and dark money
Three Facebook ads from Donald Trump depicting Kamala Harris, Hillary Clinton, and the concept of Cancel Culture.

While "Defund the Police" got strong organic coverage, Trump spent 6% on explicitly police, riot, and Antifa-related topics, while Biden spent close to zero. COVID-related messaging was moderate, even from Biden. Notably, both candidates used high-concept phrases that compared their elections to an existential crisis for the country. For Biden, the election would determine "the soul of our nation.” For Trump, his supporters were the "heart and soul of our beautiful nation.” 

SOME OF THE LARGEST ADVERTISERS FOCUSED MORE ON PERSUADING DONORS THAN VOTERS

Many large advertisers spent more than 40% of their Facebook funds on fundraising ads.

In April 2021, the New York Times reported on the Trump campaign's aggressive and misleading use of recurring donations, many likely originating with a Facebook ad. While others may not have acted as aggressively as the Trump campaign, many of the largest spenders focused significantly on fundraising. The Super PAC Progressive Turnout Project and its associated entity, Stop Republicans, was the fourth-largest spender on Facebook and Google in 2020, with 63% of its spend going to fundraising. U.S. Senate campaigns made massive use of the tactic, too, with Lindsey Graham spending 93% (>$7m) of his total Facebook budget on fundraising ads and the average competitive Democrat spending almost half their funds.

Fundraising ads from Facebook that are divisive in tone.

On the flip side, thousands of smaller campaigns with lesser-known brands, for whom fundraising ads do not work efficiently, focused on voter persuasion and turnout instead. 

Part 2: 2020 Campaign Learnings

DIGITAL’S RAPID DISTRIBUTION ALLOWS NEWER CANDIDATES TO BREAK THROUGH

Facebook ads from Charles Booker and Amy McGrath

The Internet’s ability to connect like-minded voters can give newer and lesser-known candidates the ability to break through even against well-funded opponents. This dynamic often bolsters challengers in primary elections, which have classically rewarded smaller but more engaged voter blocs. For example, in 2020, Charles Booker emerged from relative obscurity in the Kentucky State House to come within 15,000 votes of beating Amy McGrath in the US Senate primary, the immensely funded ‘mainstream’ choice, and is now the front-runner for 2022. Andrew Yang came from obscurity to becoming a notable voice in the Presidential primary. In 2018, Ayanna Pressley succeeded in beating an establishment candidate for her primary election, leveraging a digital-first strategy.

DIGITAL’S LOW COSTS AND PRECISION ALLOWS SMALLER CANDIDATES TO COMMUNICATE EFFECTIVELY

Digital is much cheaper than television for many state legislative candidates due to geographic mismatches

In 2021 the power of state legislatures has become clear, as Republican voter suppression bills methodically become law in many places. The candidates for these offices receive less organic coverage, though, and creating name recognition is critical to winning these elections. For these candidates, especially newer and smaller ones, making five- or six-figure direct mail and/or television buys ranges from daunting to impossible. 

These campaigns may only care about reaching a few thousand voters. In contrast, broadcast television reaches hundreds of thousands of voters in a geography, often far outside their district. At the congressional level, Priorities USA’s 2020 election analysis noted how the shift of competitive seats into the suburbs and away from these pre-set zones made television less cost-effective in 2020. 

Of the 63 state legislative campaigns where TFC ran ads in 2020, voters recall their ads for under $0.20, using Facebook’s estimates. In the chart above, we show that Facebook can be 12x more cost-effective than TV in inefficient geography zones. To this end, a February 2021 Wesleyan study found that more than ten times as many state legislative candidates advertised on Facebook than ran television ads, a fact that the authors attributed primarily to this cost-precision tradeoff. The lower cost of digital advertising allows many more candidates to access paid media and make their case to voters.

ENGAGEMENT DECREASED AS ELECTION DAY ARRIVED, POINTING TO NEED TO TEST AND EXPAND MESSAGES EARLIER

Political click-through rates fell sharply as election day approached

During the last eight weeks of the election, we observed a 4x decline in engagement (as measured by click-through rate). Voters were tuning out political messaging, either from already having made a decision or election oversaturation.‍

On Google, almost half (47%) of the entire year’s spend took place during those eight weeks! While this is understandable as searches increase when ballots arrive and attention increases, there was only a slight decrease in this phenomenon since 2018. This imbalance indicates missed opportunities to test messages before voters tune them out and more aggressively persuade and get out the vote in a year where more than 65% of the electorate voted early or via mail.

AUTHENTICITY STILL WINS

Cost per click went down 13-40% for candidate selfie videos

In addition to social media giving candidates more efficient access to voters, we can use data to improve upon that further. Even in a hypercompetitive election year, voters continued to reward “authentic” messaging from candidates. Authenticity is often amorphous, but two formats continue to stand out. 

Selfie videos – videos shot from mobile phones without significant post-production – continue to produce higher engagement online for a lower cost. In 2020, costs per click ran 13-40% lower for selfie-style videos, a marker of positive attention from the audiences that saw them. Boosted posts, Facebook ads that look like regular social media posts and often more informal in tone than typical political ads, continue to be effective. It cost 26% less for a boosted post to deliver a voter likely to remember the ad versus regular ads.

Boosted posts outperform regular Facebook ads, reducing costs by 26%

Taken together, the outperformance of “boosted posts” and selfie videos as “authentic” formats that seem less like traditional ads suggests a desire among voters to connect with political content differently. Down ballot candidates should continue finding ways to uplift and use their voices to reach voters.

Automobiles and why this matters

For smaller and newer campaigns like primary challengers and state legislative campaigns, digital is an incredibly effective way to communicate. Voters reward “authentic” communication like selfie-style videos and boosted news. Testing different messages and learning based on performance leverages the Internet’s efficiency and two-way communication power to hone a candidate’s message. It helps to give candidates - like Charles Booker and Ayanna Pressley -  that would not otherwise have a chance, a real platform.

At the same time, the political environment everywhere has become incredibly divisive. The division has created adverse knock-on effects, like the rise of conspiracy theories and extremist tendencies on the left and the right (often manifesting in those same primary elections!) Commentators like Yale Law’s Tracey Meares have compared the current era of the Internet to the early emergence of automobiles: an incredibly useful invention that brought with it real dangers. Over time, roads were made safer, seatbelts and other tactics were implemented, considerably reducing traffic fatalities. Riding in cars remains a dangerous activity, but one for which society has accepted the downsides. The Internet is paving a similar path: it helped birth the Arab Spring and Black Lives Matter, but also QAnon and Donald Trump.

These dynamics reached new power in 2020. As we look to the next election cycle, we may see online platforms continue to restrict the reach of political messages. Drastic restrictions would be a mistake - they would surely lead to an outsized, negative impact against smaller and challenger candidates. Improved “nutrition labels'' for funding disclosure for advertising and controls on the misinformation that often spreads via organic social media are examples of needed reforms. Furthermore, Facebook and Google’s political advertising datasets are the most complete and transparent ones available in political media. Drastic measures like bans won’t have a big impact on divisiveness, but rather will cause ad dollars to move to less transparent channels like independent digital ad operators, direct mail, and television. The amount of money needed to do so will end up hurting the chances of precisely the type of candidates many in the Democratic Party want to run.

Digital politics is here to stay: sometimes uncomfortable, but always powerful, and success for progressive and centrist candidates hinges on embracing that reality.

sign up for updates

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Thank you

This project was made possible by some of the members of our amazing 16,000+ person skilled volunteer community. Thank you to:

Cristina Poindexter, Divya Goel, Jenny Wen, Joe Goldbeck, Paulina Leperi, Steven Ellis, Bryce Eakin, Gulya Radjapova, Mark Trapani, Winston Featherly-Bean, and Zach Winston.

Methodology and sources

  • Topic analysis: TFC classified individual ads as belonging to a topic using a keyword-based classifier using the copy and link text provided by Facebook. A single ad can belong to multiple topics. The share of ad spend is determined by averaging the minimum spend and maximum spend parameters by Facebook.
  • Spend figures: Facebook spending data sourced from the Ad Library. Google spend data linked from the Transparency Report and corresponding BigQuery databases.
  • Campaign expenditure data: Sourced from the FEC, the Center for Responsive Politics, and FollowTheMoney.org.
  • Population data: 2019 U.S. Census American Community Survey 5-year projections
  • "Competitive U.S. Senate races": campaign spending from Arizona, North Carolina, Maine, Montana, Iowa, Colorado, Georgia, South Carolina, and Kentucky.