Government intelligence agencies have their own galaxies of acronyms worthy of ad tech.
The IC, for example, as the intel community refers to itself, has HUMINT (info collected by humans), GEOINT (geospatial data collection) and OSINT (open-source online info), to cite just a few.
And here’s another: ADINT, or information gathered through commercial ad tech and programmatic data, which, over the years, has been unofficially added to the mix.
Mike Yeagley, an independent contractor who has scouted and acquired commercial data and technology on behalf of intelligence agencies, is one of the earliest evangelists of using ADINT to identify and surveil targets – as in drone targets, not audience targets.
But there are also other “champions and godfathers” within the IC who saw the intelligence value of the OpenRTB and SDK system, Yeagley told AdExchanger.
Yeagley is a named source in the recently published book by Byron Tau, “Means of Control: How the Hidden Alliance of Tech and Government Is Creating a New American Surveillance State.” In the book, Yeagley describes funding for ADINT-related initiatives going back at least a decade, which is when the CIA’s venture capital arm began looking into location data and ad tech audience attribution services as a data source.
AdExchanger caught up with Yeagley to discuss his work on behalf of intelligence agencies and how (or if) this data might be used considering the rise of new privacy standards.
AdExchanger: How did you begin working with the intelligence community?
MIKE YEAGLEY: I was coming from a marketing background, where it was all data driven and about connecting data sets to record and understand behavior based on an ad ID. At first, it was about demonstrating the ability to do these kinds of things over time, that you don’t have to spend tens of millions of dollars on new satellite tracking to be able to put dots on a map and append them to an ID.
And as an outsider, I could sort of dislodge people from their attitude that unless you steal, hack or engineer data through signal intercepts, it’s not “real” intelligence data.
With consumer data privacy changes, is this type of data collection still feasible?
For years that’s been a thing I would tell Commerce Committee staffers, that by the time they took action [to fund this type of data collection with additional budget] the industry would have changed things.
The volume of data, the kinds of signals and the indications of accuracy are definitely impacted by Apple and other privacy rules. Data brokers have become the new Philip Morris, and mobile device carriers are more aware of privacy settings, particularly in first-world countries.
This is the kind of thing where – and this is typical of these types of government programs – it works well until it doesn’t. And we will use it until we can’t.
Is there always a next thing coming up to focus on instead?
There is always a next thing.
But it was good for many years. We get a lot of good value out of it, the taxpayer gets good value, and the operational community gets good value. There are volumes of data in the program that still persist, so it’s not at the point of deprecating entirely.
As adversaries, or an audience, depending on what you’re looking at, become more knowledgeable, those data sets become less and less valuable. That’s just part of the lifecycle of how the intelligence community, law enforcement or any of these other agencies use and deprecate programs over time.
How was this data used?
For example – and as an example of how these programs come and go – there was a time when a conventional terrorist would use a mobile device to communicate. That was until they learned that we were intercepting those communications. So they stopped using mobile phones.
But their wife, their kids, their drivers, the people around them were still using tablets and mobile devices. We began to track them using ad IDs and location SDK data.
We would identify safe houses for adversaries by tracking known IDs that were at previous known safe houses, and now we see they’re all at a new location. Okay, so that’s a new safe house. An intelligence agency might have spent millions and millions on satellite tracking and to develop human on-the-ground assets and still not have discovered these secondary locations.
It’s a sort of culture of how people learn about the optimal use of a device as a means of communicating. And yes, as privacy becomes a thing, the efficacy of a program changes – but there’ll be other programs.
When you work with commercial data vendors, do they know it’s on behalf of an intelligence agency or do they think it’s just an unknown business entity?
When I would approach a company, the use case was not attributable directly to the community or the DoD. It was billed as audience detection capabilities around, say, refugees and other humanitarian types of use cases where it would be applied as well.
I also wasn’t trying to activate any IDs. We weren’t doing audience activation, which makes it less complicated to acquire what we were looking for, which was bulk data processing through the RTB network and SDKs.
But whether it’s about trying to detect an audience of soccer moms or people that are into gadgets, it’s the same concept.
This interview has been lightly edited and condensed.
