Time Magazine's political site Time Swampland has an excellent behind-the-scenes report from the Obama campaign's data team, detailing improvements the campaign made from the president's successful 2008 bid.
By consolidating the databases and sharing information more efficiently, the campaign was able to monitor action in swing states more closely. In October alone, a senior official told Time the team ran the virtual election 66,000 times every night, which amounts to slightly more than 2 million simulations in the campaign's final month. The ensuing data would be processed the next morning to determine which states and districts needed the most resources. In Ohio, the swing state whose 18 electoral votes effectively dictated the winner, the Obama team leveraged data from roughly 29,000 voters.
"This was a huge advantage: when polls started to slip after the first debate, they could check to see which voters were changing sides and which were not," the report says.
All this information was used to help identify more useful campaigning methods, ranging from social networking to television ad buying to identifying people who would be more likely to donate to the campaign. The result was a re-election.
"In politics, the era of big data has arrived," the article concludes. (seriously, go read it if you haven't yet).
Where the politlcal use of big data goes from here is more interesting. Aside from the obvious - that both the Republicans and Democrats tasked with mounting the next campaign will soon begin poaching MIT grads and politically motivated data specialists from Silicon Valley - the publicity of this success could spawn a new brand of startup. Congressional and even local elections will undoubtedly benefit from similar services, making up a new market for entrepreneurs looking to siphon funds raised for the next major election.
Obama's team consisted of just "dozens of data crunchers," according to the report, which was enough to monitor voter behavior nationwide. Given the smaller scope of congressional and local elections, it'll be easier for a group of data scientists to launch a startup in this market without driving costs too high. If a few VCs see it this way too, the campaigns themselves could become "job creators," at least in the big data sector.