inspirational candidate, and harnessed the energy of Facebook and Twitter. But the hard work – the important work – was traditional database management: merging the donor database with the field-activist database with the voter database. As Thomas says, “At that point you suddenly begin to look at people in multi- dimensions, whereas before they were handled incredibly transactionally.” The money flowed in again: $500 million online in 2008, $690 million in 2012. Where was it spent? On traditional TV, giving a huge share of voice advantage. This is a particularly poignant part because in the 2016 election the insurgent Donald Trump relied on a huge earned media advantage: “authenticity” fed out via Twitter. But, more importantly, millions of people got vested in the campaign. The manager of a local field office in Cuyahoga, Ohio, could see that voter X who had given a few hundred dollars and knocked on some doors was a remarkably valuable target to create a whole experience around, with one aim – getting his or her labour. So the money financed the biggest field operation ever – tens of thousands of paid staffers, millions of volunteers, all singing from the same playbook – and socially enabled. (One technical note for non-American readers: the US “Voter File” and the long campaign cycles do make this recipe difficult to export in full – although it has been much demanded from all shades of the political spectrum in many countries.)