It is no coincidence that in both 2004 and 2012 the engines of radical innovation were the campaigns of incumbent presidents. We tend to underappreciate how radically different a presidential re-election is from any other enterprise in American political life. It is the rare chance for candidates to disrupt the cycle of short-term, election-year priorities and invest in their own research agendas instead of being forced to follow a consultant-driven marketplace.
For Bush, this proved a unique opportunity to synthesize information from consumer-data warehouses with voter registration records and apply some of the same statistical modeling techniques that companies used to segment customers so that they could market to them individually. In Obama’s case, the continuity provided by a re-election campaign encouraged a far broader set of research priorities, perhaps most important the adoption of randomized-control experiments, used in the social sciences to address elusive questions about voter behavior.
Following their 2004 loss, Democrats found it relatively easy to catch up with Republicans in the analysis of individual consumer data for voter targeting. By 2006, Democrats were at least at parity when it came to statistical modeling techniques, and they were exploring ways to integrate them with other modes of political data analysis. Already the public-opinion firms of the left saw themselves as research hubs in a way that their peers on the right didn’t, a disparity that stretched back a generation. When polling emerged in the early 1980s as a new (and lucrative) specialty within the consulting world, the people who flowed into it on the Republican side tended to be party operatives; former political and field directors who had been consumers of polls quickly realized that it was a better business to be producers of them.
Those who went into the polling business on the left were political consultants, too, but many of them also possessed serious scholarly credentials and had derailed promising academic careers to go into politics. Now that generation—Stan Greenberg, Celinda Lake, Mark Mellman, Diane Feldman, among others—preside over firms that see themselves not only as vendors of a stable set of campaign services but patrons of methodological innovation. When microtargeting tools made it possible to analyze the electorate as a collection of individuals rather than merely demographic and geographic subgroups, many of the most established Democratic pollsters in Washington invested in developing expertise in this new approach. Their Republican rivals, by contrast, tended to see the new tools as a threat to their business model.