How Obama Used Big Data to Lock Up His Reelection in 2012

March 31, 2017 • Blog Post • BY ATLANTIC RE:THINK


  • In 2012, President Obama set a new standard for the use of Big Data in political campaigns
  • Today, data analytics firm BlueLabs, along with HPE Vertica, unifies multiple databases and targets voters with custom messaging based on past behavior

Candidates are using data analytics to understand and persuade voters

When Daniel Porter and Erek Dyskant began working on President Obamas campaign in 2012, they proceeded to refine the science of electioneering on the basis of a factvoters change their mindsand a theoryits possible to identify, target and influence them through sophisticated, data-driven outreach.

By all accounts, the method they pioneered (called uplift modeling) played a major role in winning President Obama his second term in office and set a new standard for the use of microtargeting in U.S. politics.

They began by identifying voters whose historical data indicated that they might be conflictedi.e., voters who had voted for or donated to a combination of Democratic and Republican candidates in the past. Voters captured by that search became targets for advertising and phone calls, whose messages were then further refined according to the responses they generated. The goal was twofold: identify which voters needed a message to change their minds and which voters needed a message to turn out to vote.

In order to predict who those persuadable voters were, Porter and Dyskant relied on Democratic Party databases.

The party was sitting on a vast wealth of information, but it was siloed and fragmented, incapable of producing useful insights. There was fundraising data, data around emails, etc., but the databases werent talking to each other, Porter says. Even within the finance team, different departments had their own databases. The big concept and the hub of all our activity in 2012 was to bring all this data together into one unified database using HPE Vertica.

After the 2012 election, Porter and Dyskant joined up with Chris Wegrzyn, who led Obamas analytics-tech team, and Elan Kriegel, now Hillary Clintons chief analytics officer. Together, they founded BlueLabs, which provides data and analysis services to Fortune 100 companies, nonprofits and political campaigns at both the national and local levels.

Some experts wonder if microtargeting may adversely affect the electoral process since it involves prioritizing some voters over others and could conceivably cross the line from suggestion to manipulation. At the same time, there is no question that engaging voters with the issues theyre likely to care about is an effective way of increasing political participation. Obamas election in 2012 proved that, as have other campaigns have in the years since then.

Data analytics can also help candidates better understand their constituents concerns. As Andrew Hay, chief information security officer of DataGravity, told CSO, The analysis of data simply helps campaigns match the needs and wants of clusters of voters to a particular message.

Dyskant emphatically agrees. The most valuable data that a campaign has is the conversations that theyve had with voters over time, Dyskant says. If a volunteer knocks on your door and you have a really engaging conversation about the economy, then the follow-up method should be about the economy, not healthcare reform. The ultimate purpose of data analytics in politics, he adds, is to have more engaging, relevant conversations with voters and to be responsive to the trends of what voters are telling us.

His BlueLabs cofounder Dan Porter suggests that data-driven direct marketing has benefited from the disruption of traditional media and the proliferation of digital devices. The way I look at it, weve come full circle, he says. As the media space gets more fragmented, there are more opportunities to have individual, one-to-one communication and engage voters in a way that wasnt possible 50 or 60 years ago.

Just after the last election, another data-science expert, Gurbaksh Chahal, made this bold prediction: Social data drove the 2008 presidential elections, and Big Data drove the 2012 election. In 2016, it will be the marriage of the two that will determine the next president of the United States.

As founder and CEO of RadiumOne, a data-driven marketing firm, Chahal was talking his book, but given that Hillary Clintons data operation is reportedly far more sophisticated than Donald Trumps, well know if hes right in November.


Using Data to Predict Soccer Outcomes

Blog Post

Playing to win: New tools unleash massive opportunity for HPE partners

Blog Post

New HPE Pointnext Capabilities Accelerate Transition to Memory-Driven Computing

Press Release