Part 1
Elite Utilization
1
Strategic Communication in a Networked Age
DANIEL KREISS AND CREIGHTON WELCH
In the weeks before the 2012 election, President Barack Obamaâs supporters using the campaignâs Facebook application received messages asking them to urge select friends in key swing states to vote, register, or volunteer. An estimated 5 million voters responded positively to the requests of their friends, many of them 18- to 29-year-olds who could not be reached by phone (Judd 2012d). On quick glance, it may appear that there is not much new here. The 2008 Obama campaign notably pioneered the use of social media platforms for political organizing, leveraging Facebook to mobilize and coordinate supporters for electoral ends (Kreiss 2012a). What was different in 2012 lay in the campaign asking its supporters to contact only select members of their social network. The campaign matched parts of its massive voter databases, including one managed by the Democratic Party and encompassing more than 500 points of data on every member of the electorate, to data on the social networks of its supporters on Facebook to help it contact priority voters.
This social media targeting was premised on voter modeling, which entails assigning numerical scores representing likely political attitudes and behavior to every member of the electorate. These scores are the outgrowth of an enormous proliferation of data about the electorate over the past decade and, as importantly, new analytical techniques that render data meaningful. The Obama campaign used four scores that on a scale of 1 to 100 estimated the probability of votersâ likelihood of supporting Obama, turning out to vote, being persuaded to turn out, and being persuaded to support Obama on the basis of specific appeals (Beckett 2012a). The campaign and the consulting firms it hired calculated these scores by continually surveying the electorate and looking for patterns within the massive databases of political data they had access to. These modeling scores, in turn, became the basis for the entire voter-contact operation, which ranged from making âpersonalizedâ voter contacts on the doorsteps (Nielsen 2012) and through the social media accounts of voters to running advertisements on the cable television screens of swing voters (Rutenberg 2012).
Although the use of political data for electoral advantage has a long history dating back to at least the middle to late 1800s (McGerr 1986; Kazin 2007), there is both qualitatively new data and new means of using it as the basis for targeted strategic communications in contemporary campaigning. Data lies behind the extraordinary growth in online advertising, which allows campaigns to target specific groups and even individual voters with highly tailored appeals. Data is also the foundation of the targeted outreach on social media platforms that both support widespread citizen social and symbolic participation in electoral politics and provide campaigns with new ways to subtly influence the electorate. These practices enable campaigns to create âindividualized information flowsâ (Barnard and Kreiss 2013) to members of the electorate that are difficult, if not impossible in many cases, to open up to public scrutiny.
In this chapter, we take an in-depth look at the data practices of contemporary campaigns and the new forms of targeted communications they support, through the lens of the 2012 electoral cycle. We focus on President Obamaâs reelection bid given that practitioners on both sides of the aisle cite the campaign as the most advanced application of data and analytics to strategic communications in electoral politics to date and as the standard that Republicans are currently striving to meet (Confessore 2013). We first discuss the history of political data, placing particular emphasis on the rise of the sophisticated forms of voter modeling that underpinned the 2012 electoral success of the Obama campaign. We then show how data and modeling work in online advertising and strategic communications using social media. We conclude with a discussion of the implications for democratic practice, detailing the interplay between the decentering of political communication and the ways that campaigns have adopted new means of finding and appealing to members of the electorate. The data presented in this chapter is drawn primarily from a survey and analysis of journalistic articles, as well as open-ended qualitative interviews with campaign staffers active during the 2012 cycle.
In sum, we argue that the explosion of data on the electorate has provided campaigns with new ways to control their message in a networked media environment, from targeting and tailoring online advertising to using supporters as the implements of campaign communications on platforms such as Facebook and the doorsteps of voters. Data and analysis lie at the center of all aspects of contemporary campaigningâfrom the voter modeling that probabilistically determines who a candidateâs supporters are and who is undecided to the continual tracking of the outcomes of voter contacts and behavior online in order to optimize messages. And yet this is far from complete professionalized control of the message of candidates such as Obama, as social media also provide unprecedented opportunities for political discourse by nonelites (Chadwick 2013). This chapter explores the contours of contemporary political communication in the context of presidential campaigning, revealing both new capacities for controlling the message and the limits to this control in an age of digital social networks.
The Long History of Big Data
Data has increasingly become central to political campaigns, but it is deeply ahistorical to posit that this change originated with the 2012 cycle, as many popular accounts suggest (Scherer 2012; Sifry 2011).1 There is a gap in the scholarly literature as well when it comes to historical analysis of contemporary forms of mediated politics. The dominant approach to considering the effects of the Internet, and new media more generally, has failed to account for the fact that the Internet of 2012 does not look like the Internet of 2008 and even less so like the Internet of 2004 (for this argument, see Bimber, Flanagin, and Stohl 2012; Karpf 2012). Meanwhile, a body of literature compares campaigning across electoral cycles to account for technological change but offers few explanatory accounts of why and how practices and technologies change or grow more sophisticated.
Scholars need to take a historical approach to understand the particular arrangement of tools, practices, and techniques campaigns used during the 2012 cycle. Indeed, the accumulation of data by political parties, the specific systems that make this accumulation possible, the practices of analyzing and using data, and the forms of strategic communications that data supports all have a history (Kreiss 2012a). The two presidential campaignsâ specific techniques and tools, and differing capacities, for gathering, st...