Dstillery Unveils Distinguishable Characteristics of Iowa Caucusgoers

Read the full article on USA Today

or read below:

  Who needs exit polls when you can track caucusgoers’ phones? That’s what one company did. Dstillery, which has been called “Picasso in the dark art of digital advertising,” turned its intelligence-collection capabilities to the Iowa caucuses last week. The company used location data to identify more than 16,000 devices at caucus locations across the state. “We can take a population in a discreet location — in this case a polling, a caucus site — and sample that population and go and then look at characteristics of that population that no one’s been able to discern before, because we have this incredibly rich behavioral view of American consumers based on all the digital behaviors we observe,” Dstillery CEO Tom Phillips said in an interview. The results are interesting, if scientifically inexact. The company could not tell how individual caucusgoers came down by candidate but could determine, in counties decisively won by certain candidates, the dominant online behaviors of attendees: In a harbinger of more (slightly spooky) technologies that could be ahead for political campaigning — if they aren’t already in use — Dstillery also cross-referenced which devices had been used on university campuses during the previous two weeks to determine how many caucusgoers were students — roughly 5.4%, according to the analysis. And those voters showed up in greater numbers where Sanders and Rubio scored decisive wins. Phillips said caucusgoers should not be concerned that their privacy was compromised. “All this is totally anonymous,” he said. “We don’t really care who the person is and we don’t have any knowledge of who the person is… We just want to know what the consumer behaviors are attached to that particular device.” Dstillery, which has not worked with political clients, mainly sells targeted advertising to large companies, including Microsoft, Citibank, Comcast and Verizon. “We find their audiences for them using machine learning and massive data sets,” Phillips said. “That’s what we do. This is a byproduct of what we do, it’s a fun byproduct of what we do.” He said the company is at the forefront of taking intelligence data and mapping it to broader digital behaviors. “That’s pretty new. No one’s really doing that, and there’s lots of reasons for that. No one really gets that right,” Phillips said. “So we’ve been able to do that, and we like to show it off.”