No matter how you feel about US President Donald Trump and the way he goes about his business, his campaign's use of digital advertising to boost its messaging - and Facebook ads in particular - is a fascinating case study for social media marketers.
Trump's use of Facebook has been a major topic of debate, with the most recent controversy around Trump's posts prompting hundreds of businesses to boycott Facebook ads in protest over the company's inaction in addressing what many see as hate speech.
But while Trump's personal posts are often divisive, its how his campaign has used ads, and Facebook's granular targeting options, that's potentially even more powerful in activating his supporter base. And it will likely play a key role in determining the eventual winner of the November poll - and this week, we've been given even more insight into the Trump team's 'divide and conquer' Facebook ads approach, including the fact that team Trump has posted 21x more Facebook ad variants than the Biden camp, again highlighting its more intricate, targeted approach.
As reported by Bloomberg:
"Since entering the presidential contest in April 2019, Biden has spent $21 million on Facebook ads compared to $33 million for Trump over the same period, according to the Center for Responsive Politics. But the two campaigns are spending the money very differently."
The variation in spending is significant, but as we saw with Michael Bloomberg's failed campaign, buying reach, alone, won't win you favor.
Bloomberg, at one stage, was actually outpacing Trump's Facebook ad spend, reportedly dumping more than a million dollars per day into Facebook ads during his short-lived campaign. Bloomberg saw Trump's approach and sought to take him on with the same tactic - but where Bloomberg fell short was in this next element highlighted within (somewhat ironically) Bloomberg's analysis:
"The Trump campaign does constant testing of its ads and makes small changes on a daily basis. Through the morning of July 8th, Biden’s campaign had bought about 23,000 distinct ads on the platform compared to more than 489,000 for Trump, according to a Bloomberg analysis of data made available by Facebook. About 68% of the president’s ads are seen fewer than 1,000 times compared to 34% for Biden, suggesting much more specific micro-targeting by Trump’s campaign."
This is the key to Trump's approach, which has been documented in previous profiles of how his campaign utilized Facebook ads. It's not just that Trump spends a lot on maximizing reach, it's the specifics of how his campaign does it.
Earlier this year, Trump campaign chief Brad Parscale spelled this out in clear terms:
"The campaign is all about data collection - if we touch you digitally, we want to know who you are and how you think and get you into our databases so that we can model off it and relearn and understand what’s happening.”
This approach is particularly interesting in its specifics - back in 2017, Facebook came under massive public pressure after it was revealed that Cambridge Analytica, a mysterious firm which claimed to utilize 'psychographic targeting', had essentially stolen user insights from Facebook, under the guise of academic study. Cambridge Analytica then, allegedly, used those data points to formulate complex data profiles of specific audience subsets, which then enabled it to create more targeted, and in essence, more emotionally manipulative Facebook ads, which subsequently influenced how people voted.
How true that is is difficult to say - Facebook actually says that Cambridge Analytica was not effective at all. But while the issue sparked a whole range of investigations, and a whole new wave of debate over how people's data can, and is being used, what many seem to have missed within the discussion is that the role that Cambridge Analytica claims to have played was actually not essential at all.
Yes, since then, Facebook has shut down access to its data, its put new limits in place on academic study, Facebook has significantly limited how people can be targeted through the utilization of its insights. But Facebook itself still has all that data - and way, way more. And even without the data analysts of CA, businesses can still utilize the same precise targeting, without having to go to so much extra effort.
This is the most interesting element in how the Trump campaign uses Facebook ads.
As per The New Yorker's profile of Brad Parscale from earlier this year:
"One of Parscale’s favorite Facebook marketing tools was called Lookalike Audiences. “I mean, it’s why the platform’s great,” he said in an interview with “Frontline,” in 2018. [...] If you have a Custom List of three hundred thousand people, Parscale explained to “Frontline,” you can use Lookalike Audiences to find another three hundred thousand Facebook users with attributes similar to those in the first group. One of the most difficult tasks of a political campaign - distinguishing likely supporters from the undifferentiated mass of the American electorate - can now be accomplished instantly through artificial intelligence. When the “Frontline” interviewer asked how accurate Lookalike Audiences was, Parscale called it “pretty amazing.”
Whereas CA claimed to be able to profile user groups based on certain leanings or ideology, the Trump campaign now essentially does the same without their assistance.
The process works like this - the Trump campaign will publish ad like these, which promote a divisive topic.
Then it all comes down to data collection - the CTA on each of these ads is an email sign-up, which essentially enables the Trump campaign to collect data on who, specifically, is passionate about a specific topic. If a user was incentivized enough to provide their email address, that's a good indicator that this is a key pain point for them.
Collect enough email addresses and you can build a Lookalike Audience based on that info, which can then double, or more, your target audience pool, adding people that Facebook's data matching suggests will likely be equally engaged in the same issue, and will have the same leaning. Run another variant with this new audience and you collect more email addresses, hen you add them to your list of people for that specific subject.
Over time, the Trump campaign - through almost half a million Facebook ad variants - has been able to build very specific custom audiences for its Facebook ads, which enable it to poke each group whenever an issue of specific relevance to them comes up. And when the election nears, they'll poke them again with more specific, targeted messages designed to pull them closer to Trump - whether that's through highlighting his successes or stoking fears of a Biden-led Government on each front.
In this process, you don't need to complex data analysis of Cambridge Analytica, you simply need to use Facebook's existing ad tools to advantage. That comes through data collection, through segmentation, and eventually, through divisive messaging.
Whether you agree with the tactics or not, you can imagine that they would be highly effective.
That's why it's so significant that the Trump camp has posted so many more ad variants, and is spending so much on Facebook ads. It's also why if Facebook were to ban or restrict political ads, it would have such a massive impact. Facebook has argued that by blocking political ads, that would only benefit those with an existing media platform, as it would make it much harder for smaller candidates to get their message in front of more people. That may be true, to a degree, but without Facebook ads to fall back on, the Trump campaign's divisive tactics would be significantly impacted.
That could, arguably, be a better outcome, more so than the potential impact on smaller candidates. We'll likely never know - Facebook continues to resist all calls to ban political ads, though it is, reportedly, now weighing a ban on such in the 72 hours prior to the election.
But the damage would be done by that stage either way - if Facebook was serious about addressing concerns with the misuse of political ads, it would take a look at how its systems facilitates divisive politics, and assess how it could restrict such. Maybe by eliminating Lookalike targeting for political campaigns.
It's too late for this campaign cycle, but when you look at the data, it's clear that stoking fears, and prompting division, is key to the Trump campaigns approach. It's divide and conquer - with the focus on conquest, above all else, now overshadowing the true aim of political process. Which should be in electing the people that are best suited to lead a nation towards greater prosperity.