For many years, political campaigns have used basic demographic data in order to predict voter behavior. But in 2012 President Obama’s campaign team took big data and advanced analytics to a new level. While details of “the cave”, as the campaign called it, were kept secret until election night, recent articles in Time and the Chicago Tribune shed light on the new and innovative ways they were using data and advanced analytics. The Obama campaign hired an analytics team of 54 individuals and spent millions of dollars in the effort. No other Presidential campaign not even Karl Rove in 2004 or Dick Morris in 1996 has so fully embraced advanced analytics and data-driven decision-making.
Most importantly, advanced analytics helped scan the swing states for voters most likely to support President Obama. It was used for local media buys, email personalization, and solicitations for campaign contributions. The analytics team went beyond traditional demographic data and built models based upon unique combinations of demographic, geographic, behavioral, and attitudinal attributes for each stage of the campaign. The Soccer Moms’ and NASCAR Dads’ of past campaigns gave way to micro-segmentation at a local and individual level. And to top it all off, these models were applied and re-applied as new data was collected, so Campaign officials were able to make major decisions based upon near-real time insights from the data.
Another critical part of the process was their extensive use of testing to collect data for modeling and analysis. Although there was no primary challenger in 2011, the Obama campaign still made tens of thousands of calls and sent millions of emails to voters nationwide to determine whether or not they supported the President. They tested campaigns against lapsed donors, sent mailings to previous subscribers to their email lists and ran hundreds of local ad campaigns to test messaging and targeting tactics. Each test left a trail of response data they could mine against hundreds of individual attributes to build highly predictive models of voter behavior. The campaign was fully committed to these testing activities and saw them as an investment in data and learning for the general campaign to come.
While the analytics that the campaign pulled off is certainly innovative and impressive, it is also a story of big data integration. Their breakthrough was that the team was able to integrate previously disparate data sets like voter registration files, DNC donor lists, social data, and third-party databases containing demographics, psychographics and media habits to create a mega file that provided insights all the way down to the household and individual level. Once integrated, the analytics and insights were only limited by the team’s time, energy, and imagination.
We at West Monroe Partners have extensive experience in customer data integration, converting that data to actionable analytic insights, and operationalizing those insights to drive business results the same types of data integration processes ,technologies and analytic models used by the Obama campaign. If advanced analytics via micro-targeting can help re-elect a President, just think what it can do for your business to help reduce customer churn, make product recommendations for cross-sell, or increase customer advocacy.