Digital Political Advertising in 2020: What We Learned
Welcome to the back side of 2020. This is the point where seasonal affective disorder (SAD) kicks in for campaign staff in 48 states. It’s also when we try to quantify what was spent, what was learned, and how to move forward into the next campaign cycle.
The pandemic made our digital infrastructure tougher, more resilient, and non-optional. We saw demand for first-party data skyrocket and onboarded nearly double the number of files compared to 2018. If anything, the pandemic accelerated the need for onboarding vote by mail (VBM) segments, exclusion files, and custom issue-specific data.
Political digital spending was off the charts. It was literally immeasurable, as we will never know the exact numbers because the FEC doesn’t require sub-vendor reporting, and most states are even more lenient. There will be no precise statements about total political spending because nobody actually knows.
While in the past we could make decent estimates, the ecosystem is too complicated and reporting rules too broken to make accurate statements about political digital spending.
Kantar Media, by far the most credible of those estimating political spends, originally estimated $1.2 billion would be spent on digital. But we expect the estimate will soon be increased to about $1.8 billion. Is that precise? I have no idea, and neither do you.
Show me a social media or Google-generated spend report and I’ll show you an FEC report that doesn’t itemize expenses by tactic or a Demand Side Platform (DSP) that will not release this info unless legally required. Meanwhile, be wary of any platform that uses percent growth rather than hard figures. I ate 200 percent more hamburgers this month than last month. This tells you nothing.
By Election Day, we did collectively confirm once again that political spending is recession-proof. There was no drop in political spending through the 1981, 1990, 2001, or 2008 recessions. There will be no measurable drop in 2020 aside from a reallocation of spending from Q2 to later in the campaign cycle. There is, in fact, always money in our banana stand.
What else did we see, other than the obvious spending increases? OTT took center stage this cycle but has some growing up to do. According to Nielsen, since quarantine living began, total hours spent with CTV devices are up a whopping 81 percent year over year. That increase equates to nearly 4 billion hours of use per week.
Both Republican and Democratic strategists quickly recognized the accelerated viewership trend and invested more ad dollars on OTT than anyone estimated a year earlier. It’ll take a while before the publishers finish counting their money, but we think it will be higher than the $750M we predicted a year ago.
The demand for premium inventory was profound, especially in popular swing states. The Holy Grail of OTT is still scale, especially when voter file, ACR (automatic content recognition) or behavioral targeting is added into the mix. The solution was a programmatic bidding war on the open exchange to knock out competitors — many times on the same side of the aisle — but advertisers cannot force voters to consume more OTT just because there’s budget.
Programmatic buying added another layer to the process in “waterfall” auctions that prioritized certain bidders above others. We anticipate more 1:1 direct publisher deals and programmatic guaranteed inventory next cycle. Still, no one can predict the last-minute and always welcome cash dumps that are cornerstones of our industry, even on a high growth screen like OTT.
Ad fraud and quality issues followed the money. Bad actors including sophisticated invalid traffic (SIVT), app and device spoofing, ghost apps, and unauthorized inventory are real threats to the integrity of political advertising. We saw brand safety become an even larger focus in 2020 for savvy political buyers. OTT is still a relatively nascent medium and has significant limitations for third-party fraud and SIVT verification partners.
The lack of device authentication standards coupled with the inability of connected devices to support VPAID tags, makes it nearly impossible for any third-party vendor to accurately or consistently measure view-ability on OTT. Advertisers must monitor impression quality closely, verify IP addresses, and be on the lookout for red flags like random apps getting high impression counts, too good to be true CPM’s or multiple clicks in click-less environments. As for the future of OTT political advertising, we anticipate continued viewership growth (looking at you, Montana) and demand for transparency around what inventory can be accessed at what cost.
DSPs are the new front line in regulation. The onramps to digital advertising can be divided into those who have undergone severe scrutiny (primarily search and social), and those who may be based anywhere in the world and have thus far escaped the gravitational force of the press, state election law enforcement officials, and federal regulators.
The latter group, DSPs, inherited far more last-minute spending, particularly from independent expenditures. But we are rapidly approaching a point where large DSPs have to weigh the costs and benefits of delivering partisan ads. We saw one of the bigger DSPs, which billed $7.5 billion in 2019, impose a Facebook-like ad blackout after election day. Big DSPs need to get smart on this fast or they will be eviscerated in 2021 by smaller, more nimble advertising onramps which are specialized in our vertical.
Socials disappointed. I cannot emphasize enough how unhealthy the relationship is between Facebook and the political professional community. On one hand, performance is strong and demand for their ad products is healthy. On the other hand, non-presidential campaigns were clearly unwanted. Math matters and local political advertising is a bad business model for any social media giant. Twitter figured it out early enough in the political cycle to manage investor expectations and came out stronger for it. In time, all the socials will recognize that the arc of innovative advertising ideas is long, but it bends away from money-losing ideas like local political advertising.
There’s no such thing as a partisan tech or data advantage. Technology spreads fast, and in cases where there’s no infrastructure buildout required, it spreads really fast. Compare outbound text messaging (fast adoption) to addressable cable (slow adoption). One requires software and the other requires a wholesale rewiring of the majority of homes in America. The same applies to political technology.
If Republicans have a piece of technology that Democrats don’t (not true), you can bet Democrats will figure it out within one campaign cycle. The only people who incessantly complain about partisan tech disadvantages are blinded by their partisanship. That’s not to say one side is better at organizing disparate groups or hiring smart people. If your party or campaign had a tech deficiency this year, that’s a human resource problem.