Editor at Large
Published: 08 Jan 2013
The use of technology in the 2012 presidential election has generated scores of articles, and without a doubt will be analyzed for years to come by political historians. The campaign, however, could just as well serve as a lesson book for CIOs engaged in capitalizing on a company's most valuable asset aside from money -- data. The use of "big data" and data analysis was an important factor in securing a win for one side and documenting a defeat that, by most accounts, the other side never saw coming.
One of the first campaign lessons for CIOs is that any data analytics program, no matter how successful, can't rest on its laurels. The 2008 Obama-Biden campaign organization had been touted as running the most technologically advanced presidential campaign in history, with particular praise heaped on its use of social media to connect with voters and its application of data analysis to gauge voter behavior. Just two years later, the Democrats' pounding in the 2010 midterm elections showed that this vaunted voter connection had frayed -- and that the Obama administration either didn't know it or was ill-equipped to deal with the change -- or both.
Driven by a cadre of data enthusiasts, the Democratic campaign upped its tech game, making a radical break from the traditional methods used to track voter sentiment. As described by journalist Sasha Issenberg in "How Obama's team used big data to rally voters," the 2012 Obama-Biden reelection campaign not only built new tools for collecting and analyzing data, but also figured out how to put this vital data in the hands of people -- and a president -- who were in a position to change the behavior of the electorate, person by person. The campaign devised a strategy that took advantage of the volume and variety of data out there to "microtarget" the populace -- in real time. It recognized the velocity with which public opinion changes in an always-on culture. Rather than collecting data to test assumptions, data pointed the way, according to Issenberg.
"It was almost as if one group was trying to make things happen with the use of their data, and the other group was trying to figure out what to do, based on what had already happened," said Todd Thibodeaux, president and CEO at CompTIA Inc. in Downers Grove, Ill., referring respectively to the Obama and Romney campaigns. "Historical data may not be as useful as it used to be in some of these contexts."
The use of big data and analytics to change voters' minds was due in large part to Dan Wagner, a veteran of the 2008 presidential election who took on the mantle of chief analytics officer in the 2012 reelection campaign, according to Issenberg's account. "His techniques marked the fulfillment of a new way of thinking, a decade in the making, in which voters were no longer trapped in old political geographies or tethered to traditional demographic categories, such as gender, depending on which attributes pollsters asked about or how consumer marketers classified them for commercial purposes. Instead, the electorate could be seen as a collection of individual citizens who each could be measured and assessed on their own terms," Issenberg wrote.
Microtargeting and the power of personalization
Thibodeaux had a bird's-eye view of this new way of microtargeting voters. His assistant, Patti Hansen, was among the army of volunteers who collected data from prospective voters and fed it into a custom-built customer relationship management (or in this case, "constituent relationship management") system. Some of the calls she made were designed to "psych up" people identified as "ready to jump on board." Early on, however, most interactions were "cold calls" to gather information on people -- voters and unlikely voters, Democrats and Republicans. Volunteers collected all kinds of information, from what the person at the other end was reading to how often he talked about politics to friends. "They were trying to get this granular stuff on people and put all that information into data sets," Thibodeaux said.
It was almost as if one group was trying to make things happen with the use of their data, and the other group was trying to figure out what to do, based on what had already happened.
The profiles were refreshed constantly. Each volunteer drew on what the previous caller had entered into the system to expand and refine the data set; the dynamic profiles allowed the Obama-Biden team to constantly refine its message to individual voters. The break with tradition was not only in the kinds of information collected but also in how it was collected. The campaign allowed volunteers to leave some fields empty.
"When people are filling out data in a traditional database structure, they want to fill out all the fields on everybody to homogenize customers. So, a lot of times that means going to the lowest common denominator of fields," Thibodaux said. "In this world, you were allowed to not have the same amount of information on everybody, but instead get the most important pieces of information on each person. It is more unstructured."
Multidimensional view requires big data and predictive analytics
Analyst Mike Gualtieri, who covers big data and predictive analytics at Cambridge, Mass.-based Forrester Research Inc., said the Obama-Biden campaign accomplished exactly what companies want to do in their marketing and product development: "They want to create even more personalized experiences with their customers."
To do that, companies need to know "a lot more" about their customers and would-be customers, Gualtieri said. "In the past, we talked about the '360-degree view' of the customer," he said. That view, however, tends to focus on standard demographics (age, gender and ethnicity) and is limited to interactions with the company, he added. "It is a very company-centric, transaction-based view of the customer. To make it personal, you have to know even more about customers -- maybe their aspirations, their mood at any particular time, their location at a given moment, what they are doing with other companies, even competitors."
Instead of a company-centric 360-degree view of the customer, CIOs should be helping to provide their companies with a "multidimensional view" of each customer, Gualtieri said. "The only way you can find that deeper meaning is by using big data and predictive analytics on top of that." The election served to solidify the need for this.
To be sure, there are challenges on the technology side, Gualtieri said. CIOs need to worry about storing, processing and accessing data. (He calls it "SPA.") And that means all the data. "In the past, companies would make decisions in a conference room about what data is valuable and what data isn't. You don't know, so you have to store it all," he said, adding that companies are investing in Hadoop clusters and other tools that allow them to store big data sets. Processing requires that traditional cleansing and enriching techniques be performed, such as extract, transform and load, and other integration tools. In this new effort to get a multidimensional view of the customer, however, "it also means doing analytics, using machine-learning algorithms to find the question you didn't know to ask," he added. To provide access to the data, CIOs need analytics platforms that help people visualize the data.
But the real challenge for CIOs, according to both Gualtieri and Thibodeaux, will not be in accumulating or even processing data but in developing the mind-set, skills and organizational governance to take advantage of big data and data analysis tools. "Companies are just not set up to operate in a way that allows them to see customers as individuals," Thibodeaux said. The Obama-Biden campaign serves as model for how it's done.
As journalist Issenberg pointed out, few events in national life besides a presidential election "touch 126 million adults … on a single day." With the help of big data and analytics technology, the Obama-Biden campaign did just that. "Obama did so by reducing every American to a series of numbers. Yet those numbers somehow captured the individuality of each voter, and they were not demographic classifications. The scores measured the ability of people to change politics -- and to be changed by it," he wrote. And that is a lesson to be learned by every CIO in this nation.
Let us know what you think about the story; email Linda Tucci, News Director.