Out with the guru, in with the geek

What can the UK learn from US politics' number-crunching and precision-targeted campaigning?

People walk past a poster of Barack Obama. Photograph: Getty Images

Wherever Barack Obama’s target voters were, his campaign team knew how to reach them. It knew which magazines they read and even which bus routes they travelled on. The precision-guided adverts that it placed helped propel Obama to victory against Mitt Romney in battleground states.

The US president’s 2012 campaign was the fullest expression yet of what the American journalist Sasha Issenberg calls a “scientific revolution” in the way elections are fought. In his book The Victory Lab: the Secret Science of Winning Campaigns (Crown Publishing) he describes how the hitherto unimaginable quantities of data assembled by US political parties have enabled them to “micro-target” voters with ever greater sophistication. The “gurus” who rely on hunches are being supplanted by the “geeks” who rely on numbers. Already required reading inside the Beltway (the US website Politico described it as “Moneyball for politics”), the book is now attracting attention in Westminster as all parties search for the elixirs that will deliver victory in 2015.

When I spoke to Issenberg, who is part of a new cadre of stats-savvy US journalists, he told me that the tipping point came in 2004, when: “People in politics realised that the corporate world knows a lot more about consumers than they do about voters.” By acquiring data on people’s shopping and viewing habits and matching it up with their existing canvassing records, parties “were able to develop new statistical models and to look for patterns in the hundreds, sometimes thousands, of individual data points that resulted”. It was after Obama’s number-crunchers discovered that swing-voter households containing teenagers were more likely to support him that the campaign decided to buy ad space in video games. George Clooney and Sarah Jessica Parker were invited to host fundraisers after they were found to have the strongest influence on 40-to-49-year-old women, the demographic group most likely to donate.

As well as harvesting data, the parties are performing what Issenberg describes as “political versions of drug trials”, in which a new campaigning technique is tried on one group of voters while a control group is left untouched. Before a Michigan primary in 2006, the political consultant Mark Grebner sent citizens a copy of their voting history (already publicly available), along with their neighbours’, and informed them that an updated set would be sent to all residents after the election. Turnout increased by 20 per cent among those who received the message.

The words “creepy” and “scary” are the most common responses to such techniques, but Issenberg argues that they are evidence of a fuller, healthier democracy. “These types of nudges work because they make voting meaningful to people and one way they make it meaningful is by appealing to someone’s desire to fit in or not be shamed.”

Rather than relying on crude stereotypes such as “Worcester Woman” and “Mondeo Man”, data-mining allows campaigns to treat voters as individuals. By the 2016 presidential election, if you visit an environmental campaign website you will see an advert on the Democrats’ climate change policy; visit a jobs site and you’ll receive one on their employment policy. In the new era of big data, where the voter leads, the party will always follow.