Big Data • Big Analytics • Big Insight

January 14, 2013

Big Data Assists Obama Win

Ian Armas Foster

Hopeful phone calls, emails, and people standing at street corners are common to campaign season. For a long time, these tactics were more or less random, maybe fueled by a gut instinct of a couple of campaign managers.

For Obama’s 2012 re-election campaign, that could not have been further from the truth. After the election, we looked at how an advanced view of statistics changed political punditry from a gut-guessing game to a mathematical exercise. As profiled by this Time piece, big data analytics also helped ensure four more Obama years in the White House.

According to Obama campaign manager Jim Messina, they felt their data operation was one of their strongest advantages over the Romney campaign. As such they implemented drastic measures to ensure its secrecy, including placing it under media embargo until after election results were released and having the data team work in a windowless room separate from the rest of the campaign.

The key was to disregard any assumptions and let the information do the talking. Big data analytics are about letting statistics and predictive models help efficiently make use of one’s resources.

For example, the Obama campaign had George Clooney and Sarah Jessica Parker ready and willing to make speeches or public appearances to increase awareness and funding to the campaign. The campaign determined through data analysis that Clooney tests highly with affluent California women aged 40-49. Those women were thus likely to donate sums of money if a chance to have dinner with Clooney were provided as an incentive. Likewise, the potential for dinner with Obama and Parker opened the checkbooks of the East Coast affluent women.

The goal was to raise a billion dollars—a seemingly monumental task they were able to accomplish thanks to their data-driven efficiency.

That efficiency was put to use again in decisions on how to allocate those funds with a focus on increasing voter turnout in high-reward areas. Messina noted that the campaign ran over 60,000 election simulations every night and used the results daily to identify and bolster weak areas.

For example, after Obama had a poor showing in the first presidential debate, the analytics team found that the majority of voters lost were those who were likely to vote for Romney anyway. After some complex modeling, the team was able to determine where ads should be bought to regroup. “We were able to put our target voters through some really complicated modeling, to say, O.K., if Miami-Dade women under 35 are the targets, [here is] how to reach them,” commented an Obama official.

While there were still some instinctual calls to be made, they too were informed by the large database. A large portion of young voters supported Obama but could not necessarily be counted on to vote. In an attempt to mitigate that trend, Obama took the unconventional route of holding an AMA (Ask Me Anything) on Reddit. In doing so, he bolstered his youth support. “Why did we put Barack Obama on Reddit?” an official asked himself. “Because a whole bunch of our turnout targets were on Reddit.”

The power of big data can be unlocked for those like the Obama campaign with the funds to carry it out. In the end, it helped him retain a powerful job.

Related Articles

Obama Win Reinforces New Tech Era

The NSA, Big Data, and “Total Information Awareness”

The Algorithmic Magic of Trendspotting