By Lloyd Marino
Here’s a quick quiz? Who among the current crop of presidential candidates is the most successful creator of Big Data on Twitter? Hint: It’s not Hillary Clinton. In fact, it would be one Donald Trump whose presidential campaign is almost entirely built on social media platforms.
While the Republican front-runner’s quickly fading rivals still cling to the hope that The Donald lacks sufficient skill to entice supporters and ultimately lure them to the polls, it appears the opposite may be true. As the popular website Politico recently reported, several months ago Trump assembled an experienced data team to “build sophisticated models to transform fervor into votes.”
The Donald’s data team, headed by two low-profile former Republican National Committee data strategists, Matt Braynard and Witold Chrabaszcz, and includes assistance from L2, a political data outfit, focuses on integrating information Trump has collected, through his campaign website and at voter rallies, on non-traditional and unregistered supporters. It also includes “commercial data obtained from the [Republican National Committee] and other sources, in an effort to mobilize voters in key early states,” reports Politico.
Though Trump isn’t the only Republican candidate to make use of Big Data—his main rivals Ted Cruz, Marco Rubio, and Jeb Bush combined have spent tens of millions of dollars on data—he’s seem to gotten the most mileage out of the data collected. Though there’s no election in the world more expensive than the race for 1600 Pennsylvania Avenue, a billionaire candidate like Trump, who doesn’t need to grovel for money and has personally financed his White House bid, will find that data is far more valuable than dollars.
Following in the wake of President Obama’s successful data-driven 2012 re-election campaign, Big Data has become a critical player in big time politics. From building and maintaining massive voter databases to piecing insights into voter behavior, data is the key to everything from targeting and motivating voters to dynamically determining resource allocations.
Big Data works the same in politics as it does in business; the only difference is that marketing efforts target voters instead of customers. Indeed, the era of candidates knocking on doors and greeting shoppers at supermarkets also seems to be part of our bygone past. With Big Data firmly entrenched in the political fray, campaigns and candidates have to figure out how to generate enough data for effective analysis, how to cull data from multiple sources, and ultimately how to apply predictive analytics for the best results – all while dealing with that nasty little problem of voter privacy.
Having set the stage, it might be useful to see how the 2016 presidential campaigns can use Big Data to influence the outcome of the race.
Using social media to identify and connect with voters
Social media has literally revolutionized the 21st century political process. Just look at the millions of Twitter conversations and thousands of viral campaign videos that surfaced in recent campaigns. By 2012, both President Obama and his Republican rival, Mitt Romney, had become well-versed in social media and the use of social data, enabling them to further refine their digital messaging to more accurately target voters. And with continued growth in social media use— 2.13 billion users in 2016, up from 1.4 billion in 2012 — the importance of social media in the 2016 election is sure to increase.
Political engagement via social media presents campaigns with an unprecedented opportunity to collect information and reach out to voters. Like businesses that post content to capture their target audience’s attention, political campaigns post similarly engaging content. However, political content isn’t just a voter resource; it’s a way of engaging voters in active conversations that mimic real-life dialogue. Social media fanatics also share campaign content, creating a wealth of user-generated information that campaigns use it in their traditional advertising.
Social media can also feed campaign databases with an information avalanche generated when voters provide information about themselves to campaigns. Bloomberg Media ran an interesting story on how Hillary Clinton’s presidential campaign has been data mining like crazy; collecting voter information via social media to build “ as big a list as humanly possible” for her bid to become the country’s first ever female president.
The Big Challenge: Data unification and data privacy
Of course, Hillary Clinton isn’t the only candidate who’s using social data to build a database of potential voters. But no matter who’s doing it, there are two important concerns for any campaign that collects voter data:
- How to analyze voter data culled from multiple sources
- Protect voter privacy
Of course, the key to making data useful is making sense of the information collected. Campaigns are gathering data from a variety of social media channels – and not only that, they’re also collecting it via email lists, and at campaign events, among other sources. The real challenge isn’t collecting. Rather. It’s analyzing the data and then determining how to act on it.
Data privacy also is an important challenge, especially in an era of targeted and individualized voter information and communication. As former New Jersey deputy attorney general Joel S Winston told ZDNet, “Modern campaigns have an enormous task to protect the big data they give to staff, vendors, and campaign volunteers.”
Analytics drives strategy
Collecting data is one thing. However, driving strategic decisions is Big Data’s true value in the political arena. As I’ve said in multiple columns, data is ubiquitous; it’s continuously created as people share events and information on social media, through their mobile devices, or via any number of online portals. Applying Big Data algorithms to collected information allows campaigns to target voters more precisely than ever. They can also make more targeted decisions about resource allocation, e.g., determining how to spend media dollars to get the biggest impact.
Big Data also is incredibly useful in identifying and targeting undecided voters (sometimes referred to as “swing voters”). Similarly, data can be used to identify the most hardened supporters who can then persuade those in their own networks to vote for their favored candidate.
Data analytics plays a predictive roll as well. That’s what brought so much attention to a man named Dan Wagner and his work first as the Democratic National Committee’s Targeting Director and later for his role as Chief Analytics Officer on President Obama’s 2012 reelection campaign. Wagner used a variety of data science tools, drawing on multiple data sources and statistical models. While pundits have historically used national polls, and cherry picked the collected data, Wagner, as part of his overall analysis, employed the results of hundreds of state-level polls. And poll data wasn’t the only source: other factors that can influence elections, like economic variables, demographics, and party registration figures were also incorporated. Wagner’s efforts were seen as a triumph of data science over intuition. It’s a terrific example of how Big Data is increasingly shaping the political landscape.
None of this implies that Dan Wagner’s methodology is fool proof. Many factors influence election campaigns, some of which are not so easily quantifiable like bad weather. Still, Obama’s 2012 triumph is a reminder that we in data are constantly challenged to re-examine and refine our methods to use data to our best advantage. It’s also a reminder that in politics, anything can happen.
Image by: Ronda Darby