Langer and Peterson described how the Obama campaign gathered data in novel new ways and used segmentation, regression-based modeling and Bayesian analysis to improve and isolate conversion for voters and donors. They shared data strategies and tactics that businesses could replicate, including:

• Testing. Langer described how the campaign build a “testing culture” to ensure the reliability of the data it used to guide correspondence and canvassing. He described how they uncovered flaws in one consumer database by cross-referencing it with voter registration information.

• Democratization of data. The campaign didn’t restrict access to its database to a few central organizers. Volunteers could scour these names and contacts to create phone trees or organize canvassing teams in their neighborhoods. When canvassing in a new neighborhood, volunteers could target the homes of likely Obama supporters, and often knew the issues of concern to people before knocking on their door.

• Vendor competition. The campaign signed short contracts with multiple vendors for phone canvassing and other services, and closely tracked the success rate of each vendor. The vendors with the best results received contract extensions.

• Mobile. Exclusive content offered to mobile subscribers – including the text-based announcement of Joe Biden as Obama’s vice presidential nominee -- was enough to convince large numbers of people to provide their mobile phone number and other information to the campaign.

• Social Networks. In addition to MySpace.com, Facebook, Twitter, LinkedIn and BlackPlanet, the campaign reached out to smaller, interest-based social networks. Peterson described how the campaign, for example, generated significant engagement with targeted ads on a network for Goths.

To leave comments, please login or sign up to the Engage Network at the top of this page.