It’s hard to imagine, given where we are with sophisticated marketing in nearly every aspect of our lives, that only two decades ago political campaigns ran on hunches and gut feelings. Today it seems every word uttered by a candidate and every story that emanates from a campaign headquarters is vetted, focus-grouped, A/B tested, rehearsed, dissected, and media-proofed before it launches into the wild. Political organizations have their fingers grafted to the pulse of American political discourse and they know precisely how to make the American heart beat race. Yet at the same time candidates are rehearsing canned lines that “tested well” in trials they are managing to strengthen their ties with individual people – individual voters and contributors who, more than ever, feel like they have vested interests in the candidates they support. This head-spinning transition was not an accident. Over the last two decades polling specialists, traditional political campaign experts, social scientists, data analysts, and tech-savvy marketing gurus have been increasingly drawn toward the problem of voter engagement. Candidates continue to expand their understanding of the science behind voter choices and technology brings normal voters closer to the political machine. The story of this upending movement is captured brilliantly by Sasha Issenberg in his book The Victory Lab: The Secret Science of Winning Campaigns.
Issenberg takes readers through stories of the campaigns of yesteryear when political hopefuls sought new ways to broadcast messages to the masses, reaching as many people as they could in hopes of swaying voters. He describes how radio and television transformed the political landscape by changing how potential voters received information. Though these technical advances increased a candidate’s reach, they brought little in the way of substantive improvements to how campaigns understood citizens’ preferences. Television and radio became essential components of modern politics because they make candidates louder. In The Victory Lab Issenberg shows us how innovative uses of new media provide message feedback, and in so doing they make candidates smarter.
It wasn’t television that sparked the metamorphosis of political campaigns into two-way discussions. It wasn’t radio that captured citizens’ opinions. The transition to a system where feedback – real scientific feedback – interplayed with candidates’ messaging choices started with people like Hal Malchow. Malchow experimented with the wording and recipient lists of direct-mail political solicitations to produce measurable marginal improvements in election success. Through diligent pedagogic effort Malchow uniquely positioned himself as someone versed in both political campaigns and scholarly research of social pressure. Issenberg recounts Malchow’s work at the collision-point of science and politics through the story of the 2010 Colorado U.S. Senate race. After spending more than two decades learning about voter behavior and micro-targeting on campaigns and in academia, Malchow sent a million letters to carefully selected Coloradan Democrats reminding them to vote. He didn’t endorse a candidate per se, he didn’t press issues, and he didn’t launch an attack on the Republican challenger. Instead, he chose his words, as Issenberg says, “to exploit eternal human vulnerabilities – such as the desire to fit in or not to be seen as a liar – in order to turn layabouts into voters.” The strategy worked. Recipients of Malchow’s letter were 2.5% more likely to make their way to the polls where they nearly unanimously cast votes for the Democratic candidate, Michael Bennet. Simple math shows that Malchow’s work produced 25,000 votes that might not have been cast. Bennet won the election by a mere 15,000.
While Malchow was a pioneer in his field he was not alone. Issenberg also highlights the work of two Yale political theorists, Donald Green, Alan Gerber. An interesting side note in the study of political campaigns is that few candidates dare stray from proven techniques to experiment during elections. Political campaigns are not iterative – these all or nothing endeavors leave little room for risk so candidates rely on practices that have produced success in the past. The science behind voter choices was largely left unexplored because it involved risk-taking and experimentation. Few candidates are willing to risk their own election success for the betterment of the political institution and few opportunities exist for empirical testing outside real-world campaigns. Gerber and Green, sensing that the field of political science was becoming intellectually bankrupt, pushed headlong into this tough environment.
Gerber and Green were not political junkies; they were social scientists. They capitalized on a time when those closely tied to the political establishment were trying to understand why Americans were retreating from civic engagement. In a brilliant move, they allied themselves with non-profit agencies tied to voter engagement rather than engaging risk-averse campaigns themselves. They created field experiments built around genuine scientific rigor to test common methods of voter engagement. Gerber and Green hypothesized that new methods of political engagement, while far-reaching, lacked the connecting-power of more traditional contact. In Issenberg’s words, “Maybe the issue wasn’t just that Americans were being mobilized by campaigns any less, but that even the new forms of individual contact lacked a personal touch. A message that may have once been spoken at the doorstep would now come facelessly by phone or mail.” They proved themselves right – data from Gerber and Green’s experiments found that phone calls urging citizens to vote had negligible impact on turn-out, direct mail increased voter turn-out by 0.6 percent for each postcard sent, but in-person visits produced a staggering 8.7 percentage point increase. As Issenberg keenly notes, this impact is “larger than the margin in most competitive elections.” Candidates certainly cannot discount technologies like radio and television that allow them to mobilize the masses but they must also recognize that elections can be won or lost on the margins where more intimate contact produces the best results.
The problem, as Issenberg notes, is that direct contact with every potential voter is cost prohibitive. With the entrance of the Internet and its related social technologies, however, Issenberg argues that political hopefuls can now identify exactly which voters to contact for the greatest effect. Alexander Gage, a long-time pollster and opinion analyst who observed marketing strategies in commercial firms and politics for decades, noted how commercial entities used data they had collected and homogenized on their customers to influence decisions and he helped bring these techniques to the political arena. Through sophisticated and creative data fusion, Gage and others in his business found they could identify pockets of potential supporters at a granularity far below the precinct level. In so doing, Issenberg says “Gage imagined his job as developing ‘search and rescue’ tools, using data to spot sympathetic bodies in a disaster zone and plucking them out to the polls.” This level of two-way engagement and micro-targeting has changed the face of politics in ways television and radio never imagined.
As we blast forward in time to the middle of 2013 we look in retrospect to the wildly successful data-driven practices of President Obama’s 2008 and 2012 campaigns. The President and his team carved out a seat at the decision-making table for those responsible for their data strategy in the elections and the results are simply unprecedented. The combination of micro-targeting technologies driven by tremendous amounts of data with an inclusive grassroots organizing campaign was a recipe that flocked supporters to the polls like never before. This formula – something like Marshall Ganz meets Experian – was able to pinpoint voters in such a detailed and tailored way that individuals felt as if they were valued by the campaigns. This important cultural shift cannot be understated: individuals felt that the support they provided to their candidate and the votes they cast were important. In turn the candidates truly understood what mattered to people and they directed resources toward those pressing issues. This two-way communication between candidates and constituents opened doors to understanding and participation that never even existed. As the technical apparatuses and organizing methodologies continue to improve we can expect the scientific edge in political persuasion to grow ever sharper.
When we analyze the astute historical context presented by Issenberg we are given a framework from which to launch into future discussions on the effects big-data will bring to politics. The enormous data sets produced through habits and social media interactions online coupled with the immense processing power now available to interpolate this data could potentially redefine micro-targeting once again. For example, James Fowler, a Professor of Medical Genetics and Political Science at the University of California, San Diego, has recently conducted large-scale political experiments on the edge of this field. Fowler’s unique combination of academic skillsets drives him toward the scientifically rigorous testing and experimentation that Issenberg focuses on in The Victory Lab. Using enormous data sets from Facebook (to the tune of 61 million user profiles) Fowler showed how voter habits can be influenced by their Facebook friends’ actions. The entire meaning of the term “statistics” with data sets of this size must be re-examined. The insight data like this can produce in terms of engagement strategies and measurable offline behavior are beyond what the early political and social scientists every dreamed when they carved out their experimental path in this field.
But was the 2012 election an anomaly? Certainly, we saw incredible increases in the use of new media and information technology to facilitate innovative methods of shaping voter behavior. Was the election something indicative of a new trend in politics or was it a case of the right candidate, the right circumstances, and fortuitous timing? Critics argue that President Obama was a candidate who, regardless of the methods used, was going to draw unlikely voters to the polls given his background as a community organizer and his ethnicity. President Obama’s campaign staff certainly used innovative techniques ranging from the way they organized their army of volunteers to how they targeted voters, but can we assign causality between these methods and the margin of victory? Using the momentum created by Malchow, Gerber, Green, and Gage in their explicit application of scientific principles, Obama’s team wasn’t guessing on what was going to work; they knew what moved people. They collected empirical evidence through their mountain of A/B tested messages and their correlated increases in both fundraising and turn-out. The team built an unmatched technical infrastructure and, more importantly, they reshaped processes to take full advantage of their technical lead. Between the technical advantage, the reshaped processes, and proof that the money the campaign spent on technology was justified I see the 2012 election as a benchmark as opposed to an anomaly.
One of the most intriguing aspects of the transition to digital campaign strategies is found in the concept of message control. In my opening paragraph of this entry I noted how it seems every message candidates release is run through the wringer before it is allowed into the wild. Issenberg’s methodologies actually seem to suggest that control is an imperative component of experimentation. Michael Slaby, the Chief Integration and Innovation Officer for “Obama for America: 2012” tells us the concept of strict, top-down message control is all but dead. As I described in my recent analysis of Daniel Kreiss’ book, Taking our Country Back, technology is allowing more people, namely volunteers, to be given responsibilities traditionally reserved for formal campaign staff members. Digital reach allows more people to participate but this additional depth also allows more people to speak for the campaign. Slaby stresses the need for all volunteers and staff members alike to share a common framework of values as opposed to hopelessly sticking to pre-approved scripts in their interactions with potential voters. In an interview with Ethan Zuckerman, Slaby said ““If people share your mission and passion, it’s fine to let them speak – they won’t say exactly what you want them to, but they’ll pass the ‘genuity’ test.” So while modern campaigns may ruthlessly test messages to maximize intended outcomes campaign staffers are entrusted with the campaign’s voice more than ever.
Another important observation I took away from Issenberg’s work is the importance of cross-pollination between political campaigns and commercial industry. Many of the key experts and actors Issenberg analyzes flowed back and forth between the two spheres and I feel this will be an even more important exercise over the next 8-12 years. The commercial sector is developing social media and digital marketing expertise that absolutely must make its way to the political arena if the scientific edge of politics is expected to keep pace with commercial contemporaries. At the same time, organizational practices and targeting strategies are emerging through online political campaigns that will bolster how industry views potential customers and supporters in the future. The onus is on campaigns to incentivize talented, innovative thinkers in the field of digital sociology to break from their lucrative positions in the commercial sector to participate in civil affairs.
One question that must be answered as we move forward with data-driven politics is this: who owns the data? Barrack Obama collected 180 million contacts through his two presidential campaigns and we have discussed how this information was vital to his success. While in office, the Obama team parked these voter files in their offshoot structure, Organizing for Action, but it remains to be seen whether this trove of personal information – interests, issues, demographics, connections, and contribution histories – will be made available to any of the candidates who emerge in 2016. It is not a stretch to foresee a time where the ultimate statement of endorsement from one candidate to another is not conveyed in speeches but instead in the ceremonial passing of data.
Issenberg’s work may focus on the science of political engagement at the surface-level but the implications of this field reach much further into the digital realm. After spending more than a decade in the communications and cyber warfare arenas, I have found one of the largest challenges in the business is deciphering usable data from unimportant background bits and bytes. For example, when a network infiltration or anomaly occurs, the cause of the event is almost certainly captured in a log file somewhere. Granted, local log files may only contain disguised addresses and aliases that must be traced back through other logs on different networks to uncover layer upon layer of masked identities before ever finding and identifying a culpable actor, but the information exists. Computing power and storage challenges severely limit the amount of time network administrators can keep logs of every transaction that occurs on a network. Even if logs are complete, finding the proverbial needle in the haystack may expend far more resources than the task is worth. The scenario is analogous to a miner who has a mountain of worthless dirt that contains a handful of valuable gold nuggets. The gold might be worth a million dollars but it will cost the miners two million to get to the profits so the nuggets remain nestled in the earth. Through the lens of voter activation, Issenberg keenly describes new ways of getting to the nuggets and other industries, including my own, should take note of this way of thinking. While the data trail to a malicious actor may quickly grow old, online identities and histories of activity may very well lead to profiling strategies to identify those wishing to do harm in cyberspace. Issenberg describes how to locate potential voters but his framework and strategies could be used to profile just about any type of behavior as people live more of their lives in cyberspace.
Issenberg provides a poignant view of election strategy that has been in development for decades but has only recently been accepted by mainstream political campaigns. The rigorous scientific approach he describes through the work of people inside and out of the information technology arena is one that should be noted by any candidate wishing to run for office in today’s era of online engagement. The practice is anything but finalized, however. Several questions remain about how best to organize campaigns within this new reality, how to efficiently handle unimaginably large data sets, and how best to pass voter data from one candidate to another but the trajectory toward greater online engagement in future campaigns is unmistakable. The Victory Lab holds answers for political strategists but anyone who works at the intersection of information technology and traditional society will find the book and its many lessons extremely valuable.