Dr. JT Kostman, a data scientist, mathematician, and psychologist, provided the opening keynote of the ESOMAR Congress 2016 in New Orleans. He has been a paramedic, a rescue diver, and a special operations officer. “I spent the first half of my career looking for serial killers, and the second half looking for killer cereals.” The math and the techniques and the methodology are the same – the way he would triangulate on a killer’s address is how the way to identify where cereal-buying moms live.
“On the Internet, nobody knows you’re a dog.” This classic New Yorker cartoon bothered Dr. Kostman – could you use the data to identify the dogs? “We are being asked to read minds, and we can’t shirk that responsibility. That’s something we can actually do.” Context and profiling are how we start to understand what people are thinking.
Dr. Kostman had been a police officer in Reno, Nevada, where profiling was often simply based on appearance. He later worked with the FBI and the CIA. The FBI has deconstructed the characteristics of serial killers to be able to profile and identify serial killers: surprisingly, they typically own the Bible, The Catcher in the Rye, and John Fowles’ The Collector. Almost every serial killer has read that book. The CIA takes a psychohistorical approach to profiling, looking at items throughout their lives. For instance, the secret wartime report, The Mind of Adolf Hitler, analyzed Hitler’s most prolific behavior, his verbal behaviors. The CIA uses this same technique today, profiling world leaders based on everything they say, publicly and privately. The CIA looks at everything world leaders have said and projects forward.
What if we could do that for the average person, using social media? In fact, Dr. Kostman did that, profiling voters for the 2012 Obama campaign. “We have dataified not just music and words but everything.” For voters and potential voters, social media is the medium to analyze. “We took a bunch of issues and subjected them to machine learning, artificial intelligence and math to distill that to insights into who those people really are.” The data wasn’t about the Republicans or the Democrats but the voters in the middle, and what messages resonated with them.
Are market researchers under siege from data scientists? “No, people who say that are full of beans! Market research is more valuable now than at any time before.” We need both math and an understanding of people. “The numbers don’t tell enough of a story. We need the quant and the qual together.”