In what appears to be a political media-buying first, the majority of digital buys made during the run-up to the 2018 midterm elections were bought programmatically.
The finding, which comes from an analysis of buys made via Centro's Basis system, is anecdotal, but likely is indicative of underlying shifts taking place among political media-buying specialists and their campaign teams.
Basis is a SaaS-based platform enabling agencies, clients and media specialists to buy digital media either direct of programmatically. For some users, it's a self-serve interface, while others use it as a managed service operated by Centro's experts. But the net of all the political media buys is that 60% were purchased programmatically, while only 16% were "direct" buys negotiated with publishers. Twenty-three percent were social, and 1% search.
Centro executives said they were surprised by the shift to programmatic and expect it to continue to grow as campaign spending begins to heat up for 2020, including the Presidential campaign primaries and general election, but also across a host of local, state and federal elections.
What was surprising from the 300-plus campaigns analyzed in the 2018 Centro database, is that relatively few focused on connected TV inventory.
That was surprising because political media campaigns usually favor TV and digital video campaigns, according to Grace Briscoe, vice president of the Candidates + Causes group at Centro,
While Centro's Basis database does not process linear TV buys, Briscoe said only about 5% of total digital ad spending went toward connected TV buys. In terms of actual media weight, the connected TV buys were even less representative, because the format is disproportionately higher in terms of CPMs, and accounted for only 3% of total political media impressions.
"I would be shocked if connected TV isn't a much bigger factor in 2020," Briscoe says, adding, "It's just a question of how much we see it grow."
One thing she doesn't expect to see change is the timing of political media buys.
Though 2020 spending likely will heat up earlier due to the unprecedented number of candidates campaigning in the run-up to the primaries, Briscoe says the vast majority of spending will be in the last couple of months before Election Day (see 2018 month-to-month spending above).
"Programmatic" is too broad a term, because it includes both tracking-based and non-tracking-based methods. This means that people using tracking protection (e.g. Privacy Badger, or browser settings that that thwart tracking—some of which, such as Apple's, are turned on by default) won't see programmatically based political ads that rely on tracking.
@Doc Searls: Different segments of the industry have different terms of art meaning different things. Generally, MediaPost uses the term "programmatic" the way big agencies, advertisers, and media-buying platforms use it, which is media that is bought automatically using machines vs. negotiated by people.
But you're right, some of those media utilize various tracking methods, or not. So do some media that are bought manually.
Thanks, Joe.
The problem is that tracking itself needs to be called out. "Programmatic" by itself masks the difference, which is immense.
It's no accident that ad blocking, which had been available for browsers since 2004, took off in 2013, which was when the advertisers, agencies and publishers, led by the IAB, made clear that they would not respect Do Not Track, which was nothing more tha a simple request expressed in a person's browser. At last count (though nobody really knows), perhaps as many as two billion people now block ads online. That would never have happened if Do Not Track had been respected. We also would not have the GDPR today, at least not in its current form. Same goes for the CCPA in California.
By the way, I don't block ads, but I do block tracking in all my browsers (I use four), and I have not yet seen a single political ad. I know that's a sample of just one, but it sugest to me that most programmatic political ad placements are tracking-based.
Interesting item: we stopped using Google Analytics, which tracks users, at Linux Journal, when it was clear that a large percentage of our readers blocked tracking, and Google Analytics was therefore inherently inaccurate.