The Battle of the Ad Spend - Trump vs Biden Facebook Ads 2020
Follow the money
It’s no secret. There is some serious money being put behind political ads in the US right now. Like, biggest client in a global agency kind of money.
According to NYU’s Ad Observatory, since July 1st Trump has spent a whopping $62.7 million dollars on Facebook advertising to Biden’s comparatively paltry $32.8 million.
What can the money tell us about both parties' intentions, and possibly even the state of the election?
What can Facebook spending really tell us?
The interesting thing about Facebook Ads is that, depending on how sophisticated each candidate’s targeting is, the spend data could either reflect active decisions made by each party OR the audiences’ responsiveness to their messaging. Two very different things.
Facebook’s algorithms are designed to optimise towards users within an audience that take desired actions (i.e. people who click through to a website or watch a campaign video). So let’s say Trump was targeting the whole of the US, then Facebook would start optimising towards those states where his ads got the most traction. Or if Biden was targeting both genders in the same campaign, Facebook might start showing more ads to women if they had higher engagement rates.
In other words, the spend either reveals their tactics or their engagement. Is Biden deliberately spending more on swing states than Trump? Or are those swing states responding more favourably to Biden? These answers to these questions will prove critical to the outcome of the election.
That said, let's dive in.
The big picture
While overall Trump has spent 91% more than Biden since July 1st, the gap has narrowed since mid-August, with Biden actually outspending Trump by 10-15% for the last two weeks of the month.
Biden’s Facebook spend was comparatively non existent through July, possibly hoping to concentrate spend towards the latter stages of the election. Until recently Trump had opted for a generally higher but consistent level of spend with his recent September push looking like a direct response to Biden’s. Will both parties maintain, or even increase, their higher levels of spend right up until the vote? Only time will tell.
Who’s saying what?
The Ad Observatory categorises ad objectives into the 6 buckets above. Assuming these are reasonably accurate, we can see some clear differences in messaging.
Despite COVID, Trump is clearly looking to rekindle the rally-fuelled mentality that won him the election in 2016 with most of his ad spend being funnelled into calls to action to “show up”. Biden on the other hand is focussing on the much more COVID compliant strategy of calling on people to “connect”. Both parties are putting a lot of spend behind “donation” focussed ads, but Biden is spending a slightly higher percentage than Trump.
Interestingly the tool wasn’t able to categorise a sizeable portion of Biden’s spend. This could well be a limitation of the tool, or perhaps an indication that Biden’s messaging is more diverse while Trump is leaning into a few, simpler CTAs?
Everybody’s talking about Trump. Both parties’ highest spending topics focus on the president himself. No doubt just the way he likes it.
Notably, Trump’s second biggest topic is the media, with 22% of his total spend being put behind ads designed to criticise and undermine it. One of his big platforms in the first election, Trump is leaning into fake-news once again to distract from the major issues at hand. Conversely Biden is proportionally putting much more money behind ads aimed at explicitly driving donations and ones that reference the election itself.
Who's targeting who?
So we’ve taken a look at the high-level stuff, but what about the demographics? To whom and in which states are both candidates directing their spend. To answer that question, we’re going to take a look at a different data set.
Bloomberg recently published an article where they discussed the two camps’ ad spend in the week commencing Sep 2nd. On the face of it both parties spent a similar hefty amount, Trump an estimated $5.25 million to Biden’s $5.19 million. But when we break it down, the picture looks very different.
Biden spent 12% more on women, while Trump spent 15% more on men. These spending gaps roughly reflect the gender split seen in the 2016 election, with Trump receiving ~13% more male votes than female, and Clinton ~13% vice versa.
When it comes to age the picture is also fairly predictable. Trump is investing a much higher share of spend into older demographics (45+), while Biden skews much more heavily towards younger votes, spending more than double Trump’s share on 18-24 year olds. Interestingly, however, Biden is outspending Trump in the oldest category (65+), a demographic which leaned heavily towards Trump in the last election. Is this a deliberate campaign effort? Or has the mood shifted among these potential voters? The combined age and gender breakdown can give us some answers here.
The female vote
Biden is spending more on women of all ages, again either a response within the audience or a deliberate attempt to secure the vote. Conversely Trump appears to have all but abandoned the idea of converting the 18-24 female vote, which historically has leant democrat. Biden is also investing more in 65+ women than any other category, hence the overall higher investment in the 65+ demographic.
The male vote
Trump dominates spending on men 45-64 with Biden investing a greater share in young men 18-34. Is this a sign of Trump’s dominance among male voters or perhaps an indication of his insecurity and need to sure up the vote? Trump’s spend is also fairly equally distributed between men aged 25 - 55, while Biden’s reflects a more narrow, millennial demographic.
Where are they spending?
Trump’s smooth curve
Trump’s spend is concentrated most heavily on the swing states, Florida, Michigan, and Texas. Yep, you heard that right, Texas is predicted to be a potential swing state in 2020, despite its unbroken Republican run since the 1980s. A number of recent polls suggest that support there is essentially neck and neck between the candidates.
Trump claims not to be worried about Texas, but the money might suggest otherwise, with 10.4% of his total ad spend there. Although, if the spend is being assigned by Facebook’s algorithms, then this would suggest the opposite, that Trump’s engagement is actually very strong in Texas.
What’s also noticeable is the relatively smooth curve of Trump’s campaign spending by state.
Biden’s swing state focus
In contract to Trump’s curve, Biden’s relative investment is being channelled into a handful of key states: swing states Florida and Wisconsin, as well as Democrat heartland California making up the top 3. It’s likely this California spend is being driven by Facebook optimising towards the no doubt strong engagement Biden is receiving there. This might work against Biden, however, as that’s all money that could have been spent in less secure states. But that’s the Account Manager in me talking...
While both are investing significantly more into widely acknowledged swing states than secure ones, Biden is investing a noticeably higher percentage (73% to Trump’s 64%). Is that a deliberate strategy or not? If so, will it pay off?
So there you have it. Both parties are spending obscene amounts of money trying to influence the outcome of the election. Compared with other media channels, Facebook provides a unique window into each candidate’s campaign because of the way its algorithms work.
And this is, of course, just one tiny part of a mind-blowingly large, omni-channel advertising picture. For insights across all channels, check out this article on CNN.
Who knows to what extent their spending will have the desired effect, or how much Facebook’s new political ads policies will impact their respective strategies.
One thing’s for sure though, I’m glad we don’t allow paid political advertising in the UK…
Note 1: all numbers are estimates with spend share sourced from a mix of the NYU ad observatory and Bloomberg (who sourced from Facebook's Ad Library).
Note 2: “male”, “female” and “unknown” are the only gender categories we can extract from Facebook.
Note 3: the extent to which both parties are using affiliates to circumvent Facebook’s political ads policies is unknown. Check out this article on the prevalence of untraceable Facebook ad campaigns or this one on dubious affiliate spending in India’s elections last year.