Discover PerformanceHP Software's community for IT leaders // April 2013
The DNC leverages big data for a big win in 2012
Find out how the Democratic National Committee tackled its big data opportunity and helped President Obama keep the White House in 2012.
The mandate to tame big data is spreading quickly in the business world, but business is not the only arena where competition and big data collide. As the Democratic National Committee (DNC) proved in the 2012 U.S. presidential election, big data can make a big difference in politics, as well.
The DNC recognized early on that big data analytics could be a critical asset in its efforts to win the re-election of President Barack Obama. “From the beginning, Jim Messina, the campaign manager, had a vision—a mandate, really—that we should … demand data on everything, measure everything, and create an analytics team that would make sure we were being smart about things,” Chris Wegrzyn, director of data architecture for the DNC, says in a recent HP Vertica webcast.
Eighteen months before Election Day, the DNC began an initiative to create a big data analytics platform that could accomplish that vision. In the process, it learned a few best practices for building an organization that uses big data effectively:
1. Build a comprehensive view of data. Messina’s vision to “measure everything” implied a comprehensive view of election data that crossed disconnected channels within DNC operations. Though data mining had been a part of the DNC’s election strategy for a while, these prior efforts were conducted in data silos—a fragmented approach that limited the usability of the data.
The DNC is organized into five distinct channels: Field, Digital, Communications, Media, and Fundraising. “Over the past decade their operations have become more sophisticated in varying ways,” Wegrzyn says, “… but they’ve all done it independently in these channels, and it led to a fragmented technology platform.”
By contrast, the 2012 election was the first time the DNC had pulled together data from all channels to get a holistic picture of the electorate.
2. Support analyst-driven decision-making at every turn. To create actionable intelligence from this comprehensive data set, the DNC set a goal to create an analyst-driven organization. The DNC hired more than 100 experienced data analysts with diverse backgrounds. Their primary task was to experiment with data to create new strategies to persuade undecided voters and to get Obama supporters to the polls.
To do that, the DNC team required both simplicity and high performance.
“We needed to create an environment to let smart people freely pursue their ideas,” Wegrzyn says. “We didn’t want the technology to be a limiter.”
3. Create a positive feedback loop. To create the new programs that would get more Democrat votes at the polls, analysts would need to create a positive feedback loop, so that as they learned more about the data, they could quickly share results and pursue new ideas. The DNC built a system that combined speed and collaborative features to remove the barriers to innovation.
“The work that our analysts did could build on top of each other, creating this increasingly powerful platform that, from the very beginning, was functional,” Wegrzyn says, “but as we used it more, it became even more powerful.”
Finding the competitive advantage
The DNC had chosen the HP Vertica Analytics Platform after a thorough evaluation of many alternatives. The solution immediately let the DNC make correlations in the data that would have been impossible to make in the siloed system used in the 2008 election. A few examples:
- More personal follow-ups after field contact. Specific voter contact information reported by door-to-door field workers could be matched to email addresses collected by the digital team. When a field worker reported that a certain voter might be persuaded with more information about Obama’s health care policies, the digital team could then email both health care information and a personal invitation to contact the appropriate field organizer.
- More effective utilization of television advertising budget. By using predictive modeling in Vertica, the DNC built an application to optimize when and where to buy television ad time. Analysts quickly learned which media outlets delivered the best coverage of the people they most wanted to reach. This optimization technique helped the DNC spend its budget more effectively, ultimately making twice as many cable buys as Mitt Romney’s campaign team.
Such powerful results were achieved quickly, with minimal development time and training required for analysts. This allowed the DNC to begin rolling out new programs almost immediately, making the most of a fixed timeline to create results—a key to success in any competitive arena.
HP Software’s Paul Muller hosts a weekly video digging into the hottest IT issues. Check out the latest episodes.
Welcome to a new reality of split-second decisions and marketing by the numbers.
Looking toward the era when everyone — and everything — is connected.
Introduction to Enterprise 20/20
What will a successful enterprise look like in the future?
Challenges and opportunities for the CIO of the future.
Dev Center 20/20
How will we organize development centers for the apps that will power our enterprises?
IT Operations 20/20
How can you achieve the data center of the future?
What the workforce of 2020 can expect from IT, and what IT can expect from the workforce.
Preparing today for tomorrow’s threats.
Data Center 20/20
The innovation and revenue engine of the enterprise.