The 2016 Election: Big Data and the Politics of Influence

From dysfunctional debate microphones to mismanaged e-mail servers, technology has played a somewhat inglorious role in the 2016 election.

But these debacles detract from the real technological breakthrough of this election cycle: big data analytics delivering hyper-focused campaign messaging to voters.

History as told by U.S. presidential elections contextualizes the relationship of technology and culture conveniently every four years. Famous benchmarks like Kennedy vs. Nixon in 1960 ushered in the age of politics on camera. Internet crowdfunding and social media propelled Barack Obama to the White House in 2008 at a time when his peers understood the Internet as a series of tubes invented by a former vice president.

While the old facade of campaign politics remains firm (robocalls, mailboxes stuffed with fliers, attack ads), data tools have quietly advanced social media targeting to the next level.

Behind the scenes, elaborate custom-built CRMs and petabytes of database files culled from online user behavior deliver more precise messaging tailored to heavily vetted segments of the electorate. This is an entirely new approach for attracting voters, volunteers, and donors.

Big data is big business for election tech

The goal of any campaign is to reach targeted segments of the electorate as efficiently as possible. It inevitably takes owned media to achieve this objective. The media landscape has never been more complex and fragmented, but data science does the math to identify channels that influence voters if delivered the right message at the right time from the right person. Most of the time this person is a peer, not a politician.

Data firms earn big money by making this kind of targeting possible for campaigns.

A solution might look like this: Data firms harvest data from social media platforms and store it in cloud-hosted databases. They use RDD tools to sort and analyze the information. Data programmers make further refinements: they develop sets of voter profiles and may flesh them out with business intelligence visualization tools. Profiles are warehoused in another cloud database connected to a client-facing mobile application which a campaign would use to compose and distribute its messaging.

Four cycles ago, politicians had only voter history, party affiliation, and polling metrics to direct campaign efforts. Today there is unprecedented information about the constituency. Who do they follow, what do they like, what do they engage with and share, what words do they use in comments? Which devices do they use to do this? What site did they come from pre-engagement, and where did they navigate next?

These, and many other metrics, give strategists a way to drill down voters and hit them with the right messages and the right time. It is the same data science approach that corporations use to market products and services in 2016. It is called “personality based microtargeting” by persuasion experts, and it works because it turns out previously unengaged voters.

Building apps for microtargeting

There are dozens, maybe hundreds, of digital campaign tools that use a scalable cloud-hosting platform like Amazon Web Services to park the database and host a user-facing application. The apps are typically based on Python and coded in open-source query programming like SQL NoSQL, R, and Java. Some apps integrate Business Intelligence programs like Tableau for data visualization.

Small campaigns have a plethora of low-cost applications for tracking what amounts to a handful of Excel files of voters, donors, and volunteers. Presidential campaigns assemble data science teams to build custom programs to processing servers full of this type of data. They hire out digital marketing firms for website building, social media management, and e-mail marketing. These teams base creative messaging from profiles designed by data specialists. Candidates craft talking points from these materials.

That’s how it’s supposed to work, at least. Candidates aren’t message bots. Certain candidates are predictably unpredictable, and humanism resonates with some voters--which makes sense when the electorate is 100% human beings.

Conclusion

Throughout history, electoral politics notices how technology persuades the populace and attempts to influence behavior in the same manner. Candidates attempting to harness big data in the 2016 election--like Kennedy and Reagan did with television, and Obama did with social media crowdfunding in 2008--is political business as usual.