What we must learn from the Trump campaign and 2019 to prevail in 2020
But the absolute dollar amount does not tell the full story. This report highlights how the Trump campaign’s digital operation is more sophisticated, integrated and focused than any Democrats’ – and why that will influence the outcome in November. From message testing to targeting, to collecting voter contact data, there are key lessons to learn now, early in the cycle, from both the Trump campaign and our work on ad programs across 90+ campaigns since 2017, including 19 in Virginia’s 2019 state elections.
Since 2016, Trump’s team has relentlessly A/B tested different aspects of their ads – varying image, copy, ask, and targeting – to find the most engaging and best-performing messages. As a proxy for a commitment to message testing, one can look at the number of ad variations a campaign is running. From May 2018 through January 2020, the Trump campaign created 58,000 (25%) more individual ads than all the Democratic front-runners combined. The Trump campaign then incorporates those winning digital messages into other mediums, including speeches, email, and television. Fervently focused on embracing digital advertising for message testing, rather than just like an ATM where one goes for campaign funding, the Trump campaign has worked at a totally different scale than any Democratic candidate (although it’s worth noting that Bloomberg is catching up, but only very recently).
By examining the language used in each ad, we can see the outcomes the campaigns are optimizing for with their ad campaigns. On average almost 40% of Democratic ads ask for a donation, with Bernie and Mayor Pete hitting close to 60%, while only 25% of Trump campaign ads make direct fundraising asks. The lower ratio of direct fundraising asks reveals a clear strategy to trade up-front fundraising asks for data collection – obtaining emails, phone numbers, and issue preferences to be used in the future.
Some of the Trump campaign's most popular calls-to-action (i.e. asks) are:
After collecting and refining its supporter profiles, the campaign is then able to communicate with supporters privately and perpetually for fundraising, persuasion, or volunteer recruitment.
The Trump campaign is leveraging cost-effective message testing and a long-term focus on the general election into building a foundation of voter relationships that will be increasingly difficult for any Democratic campaign to beat. Can Democrats match this effort? That depends on the timeline to pick a nominee, the digital mindset embraced by its leadership, and the Party – but Trump has a big head start. And the implication is that if there is a lack of these practices at the Presidential level, it is happening all the way down the ballot more often than not.
If you’re asking yourself who’s doing digital well on the Democratic side, there is no perfect way to judge from the outside. But one interesting proxy for level of sophistication is looking at the number of ads compared to overall ad spend. Over the last 90 days, as the primary and his campaign heated up, Mike Bloomberg not only has ramped to enormous spending levels but also a high number of individual ads, likely reflecting a belief in more robust testing. Buttigieg and Warren also show promising levels given spend, while the Biden and Klobuchar campaigns lag.
In addition to the official campaign arms, super PACs and national campaign committees, a network of far-right sites, including profit vendors like “I’ll go ahead and keep my guns, Thanks” or “Conservative Gear” spent $18.5M on advertising that pushes Trump’s image and brand emblazoned on t-shirts and coffee mugs and pro-gun imagery. They contribute extra earned media with every item sold.
Learning from the lessons of 2019 is not just about the Presidential candidates. In the last two years, we’ve worked with 90 campaigns to execute their digital media programs: here are some of the key lessons we learned.
The press often focuses on the dangers of “microtargeting”: showing ads to small, specific audiences. In political practice, targeting is often treated as a panacea to persuade voters: however, our data shows that creative visual style and the messaging topic can have a far larger impact – up to 3.4x greater – than targeting on how audiences engage with ad content. This is not to say that targeting isn’t valuable; rather, it’s just one piece of a successful digital campaign.
And speaking of creative styles, authenticity in creative and messaging is fundamental to effective ads on social media. In 2018, Tech for Campaigns found that selfie-style videos, like those shot on mobile phones, were about 20% cheaper than more highly produced, television-style content – a mark of positive relevance on Facebook – and that's before factoring in production costs of non-selfie videos. In 2019, boosted “organic” posts, ads that look like personal Facebook posts, continued to do well, with costs per click 60% lower than average.
Online consumers demand not only authenticity and fitting creative, but also freshness. Cost per click, a measure of the cost of online engagement, increases 18% within 7 days of an ad running and nearly 50% within 20 days. Digital is not a medium where a campaign can post a TV ad as creative and leave it up for weeks.
As found in our 2018 report, Facebook’s ad click-through rate for users 55+ is still >2x higher than for those under 55. Click-through rates for 18-24s did trend upward in our campaigns in 2019 due to the implementation of creative strategies designed to appeal to those demographics (also known as memes!).
In political campaigning, Presidential campaigns are an anomaly in almost every way, from their access to resources, to the attention they garner, and to the level and sophistication of their media programs. However, while part of the digital disparity can be explained by a primary versus general election focus, another significant part is a differing mindset. A mindset that digital (and technology) are still not primary parts of campaigning, instead of one where digital is integrated so that the whole is greater than the sum of the parts. That mindset is not an anomaly. It permeates up and down the ballot. It’s not rational to think that different methods are being used in non-presidential Democratic races.
Working on hundreds of campaigns in a cycle and access to ad platform transparency data has afforded us the opportunity for more frequent learnings and iteration. The biggest travesty would be if we kept waiting to act on these learnings. 2020 has so much on the line: yes, it is a Presidential year, but it’s also the last major election before the census and hence redistricting which will affect policy and electoral outcomes for at least a decade. Redistricting will largely be decided by state (not federal) legislatures, so it is time to walk and chew gum at the same time - applying these learnings up and down the ballot.
The Trump campaign shows what is possible when a campaign centralizes and internalizes the importance of digital, data and technology. We hope you’ll join us in taking action to not only bring this year's learnings, but new ones for election cycles to come, to Democratic campaigns.
This project was made possible by the efforts of some of the members of our amazing 11,500-person skilled volunteer community. Thank you to:
Matt Nickerson, Bret Frangipane, Anela Chan, Stephanie Saunderson, Juliet Logan, David Tester, Kirill Bylkov, Liz Hoffman, Bernadette Herrera, Tim Cullen, Jami Butler, and Stephen Cummings.
State of Political Digital Ads: TFC teams worked to research and analyze data from Facebook's Ad Library and Google's Transparency Report. For the 2018-19 cycle spend data, data is from November 10th and 11th through February 5th, 2020.
For ad creative data from the Facebook Ad Library, except where otherwise noted, the time range is each campaign’s inception on Facebook through January 15th, 2020.
Campaign Learnings: Data sourced from Google and Facebook's advertising analytics platforms.
For the “Impact on Ad Performance: Targeting vs Creative” analysis, a regression analysis was used to isolate the effects of different messaging topics, creative choices, and Facebook ad targeting options on the on-platform performance metrics of TFC ads. These metrics included click-through rates, eCPM, and costs per click. The results were plotted to show the distribution of the outcomes of the different options in each category.
