Obama's campaign to retain White House began just weeks after 2008 win

At the same time, the campaign used sophisticated algorithms to crunch data on individual voters drawn from everything from cable television companies and shopping websites to social media sites such as Facebook and even picture-sharing sites such as Instagram and Tumblr.

Voters were classified both in terms of their likelihood to support Obama and the chances that they would vote at all.

Traditionally, campaigns determine likelihood to vote on the basis of vote history – if they voted in 2008 and in the midterm elections of 2012, they were more likely to vote in 2012. Young voters had such a short voting history, however, that traditional methods could not predict future behaviour.

Using statistics, the Obama campaign identified other factors, such as income and housing type, that tended to indicate a likelihood to vote.

They found that voters who lived with others who had a history of voting regularly were themselves more likely to turn out.

Campaign focus

Most campaigns focus on voters that are clearly identified as supporters and who have a medium-range likelihood of turning out – on the basis that those at the top of the range don’t need encouragement and those at the bottom are a lost cause.

The Obama campaign targeted more of those at the lower end of the range, however, by encouraging activists to rate each voter’s responsiveness to their initial contact.

In states that allowed early voting, the campaign focused on getting the least motivated voters to the polls long before November 6th.

By last weekend, Obama’s team had opened 5,117 get-out-the-vote “staging locations” in the key battleground states, located in supporters’ homes, shops, businesses and anywhere else that could serve as a hub. In a memo sent to “interested parties” last Saturday, Bird was already confident that the millions of conversations Obama volunteers had conducted with persuadable voters would bear fruit.

“Our volunteers have made 125,646,479 personal phone calls or door knocks – not counting robo-calls on auto-diallers, mail, literature drops or any other non-volunteer, non- personal contacts,” he said.

“Many field campaigns have historically favoured quantity over quality; we do not. In each conversation we have with a voter, our goal is to make a difference.

“These are not phone calls made from a call centre; they are done at the local level by our neighbourhood team leaders, members and volunteers who are talking to people in their communities.”