Was the best ad campaign of 2008 really the Obama-Biden presidential campaign? It was good enough to win the ad industry’s Cannes Lion grand prize, so maybe the 2008 campaign’s practitioners could provide some insight on a brief media contest that affected the whole world.
No campaign could knock on the door of every voter in the United States, or even give every one of them a phone call. The key to winning was to identify those voters who were persuadable, to target a portion of the electorate, a strategy called Micro Targeting (MT) that could be as valuable to business advertisers as it was to both parties in the 2008 election.
“I define MT as taking data you have about people and using it to make an informed judgment about who they are and how they’ll act,” said Alexander Lundry, TargetPoint Consulting vice president and director of research and an advisor to Republicans, in a webinar hosted by analytics firm SPSS.
Campaigns obtained data by interviewing a cross-section of the electorate and they extrapolated that data to find out who was persuadable. The results were something you cannot buy.
“There was this idea that you could buy commercial lists, but you get the best results with multiple algorithms. You combine the strengths of multiple techniques to overcome their individual weaknesses,” said Ken Strasma, president of Strategic Telemetry and an advisor to Democrats, in a Webinar hosted by analytics firm SPSS.
“You need gold standard phone interviews for you survey data,” he said. “Then you add social network data and unstructured data.”
“Get the voter file and append as much as you can,” agreed Lundry. “Get lifestyle data, consumer data, and geographical data.”
Using the data
Once you have the data, you need to use it well.
Lundry explained the results with data that was plausible but not real.
If you know someone is male, the possibility that they’ll vote for the GOP might be 53 percent. If you also know that they read the bible, drive an SUV, live in the state of Louisiana, are a businessman, and are 50 years old, that probability could rise to 85 percent.
Change the data and you change the result. Lundry said that the if the person was still a 50-year-old male businessman but was black, lived in California, and was interested in the arts, the probability that he would vote for the GOP could drop to 15 percent.
“We customized appeals to strictly defined groups,” said Lundry. “For flag and family Republicans, we would contact them by phone and talk about a proposed flag burning amendment. For rural non-GOP gun voters, we would focus on an opponent’s second amendment votes.”
Targeting is important, agreed Strasma. “If you’re selling a car, you don’t want to send your message to people who don’t drive,” he said.
MT is a process, according to Strasma. “Admittedly, when we started in
2002 … it was more like a project. Plan, execute, and that was it. Now, we see a process that continually feeds into itself, so we’re constantly making better judgments.”
He added that the whole organization needs to be constantly sharing and processing data.