Every day, human beings create 2.5 quintillion bytes of data – data that is generated, accessed and shared on laptops, mobile devices and social media. In the political realm, data drives today’s campaigns – from studying voter demographics and conducting outreach to tailoring and testing specific messages.

Data, Data, Everywhere Data

Technology and data are closely intertwined. In fact, politicians will allocate almost 10 percent of their media budgets toward social media – or approximately $1 billion. We need look no further than the current presidential campaign with candidates’ use of Twitter, Facebook, YouTube video ads, sponsored Google search links, and Snapchat to see evidence of this. Campaigns have realized they must harness technology to gain valuable feedback from specific audiences about how their messaging resonates and how voters are responding to key issues. Also, campaign managers need technology to recruit volunteers, plan voter outreach, boost membership and raise money. Basically, real-time data drives campaign management decisions and has changed the way political campaigns are organized, executed and analyzed.

The Humans Behind the Data

How do campaigns work with data to reach the humans behind the votes – to monitor and listen to people’s feedback as well as communicate with them? Here are three ways people put campaign big data to use to listen to and talk with voters.

1.Analyzing audience data allows campaigns to drive the right content to the right people via the right channels.

Campaigns use various resources to capture and manage data, including polls, surveys, email contact and phone calls. Once a campaign starts collecting a lot of data on voters (likes, dislikes, cars they own, jobs they hold, etc.), content can be directed to them via the most appropriate channel (printed media, Twitter, Facebook, TV). Campaigns also examine “live” data from people, including Twitter, Facebook and Instagram data, as well as inbound data from voters. For example, people may visit a candidate’s site and send an email or fill out the “contact info” portion of the site. If a voter emails with a policy question, you can collect information and tailor an appropriate response. When people visit a candidate’s site, there are several drop-down menus about various topics, from healthcare to education. The user selects a subject and is then routed to the best staff member based on demographic and geographical information.

2. Data helps campaigns make voter and election-related predictions.

More than 44 percent of the world’s population votes. This year, more than 50 countries worldwide — from Australia and South Africa to Germany and New Zealand — will hold elections, including presidential, legislative and general elections.

For any election, the question is, how many of these voters will turn up at the polling place to cast a ballot? And of the many voters registered in a specific region, how many will cast a ballot for a specific candidate?

For example, in the 2008 U.S. presidential election, Barack Obama’s team assigned every voter in the country a pair of scores based on the probability that the person would both cast a ballot and support Obama. For each battleground state every week, the campaign’s call centers conducted 5,000 to 10,000 short-form interviews that quickly gauged a voter’s preferences. Also, 1,000 long-form interviews were also conducted. Algorithms searched for data patterns between respondent opinions and the data points the campaign had assembled for every voter. This helped the campaign derive individual-level predictions, as it included 1,000 variables drawn from voter registration records, consumer data warehouses and more.

Data collection and algorithms go hand in hand with data analytics, or looking for patterns related to voter party, how a person will vote and whether he or she will turn up to vote at all. Prediction is both an art and a science. For presidential primaries, you could rank voters on a scale from 0 to 100 on “likelihood to vote” and 0 to 100 on “likelihood to vote for your candidate.” However, once a person gets into the voting booth, he may react more on an emotional level than a logical level, thereby invalidating your data-based prediction.

3.Data is used to allocate/focus resources and achieve budget “bang” for the buck, reaching the right supporters at the right time via the right channel.

Data helps campaigns determine where best to spend time and money, ensuring the focus is on people most likely to come out and vote. For the average political race, the campaign works with a pollster. The pollster goes into the field and asks questions and then, based on the 700 or so people he talked to, assigns points to each voter. For simplicity, let’s assume points (“10” is the highest) are assigned based on responses to weighted questions, and these weighted questions, in turn, are based on specific responses the campaign is looking for. A “1” means zero of the questions were answered favorably toward the campaign (so the person is obviously not a supporter). Any person with more than “6” points is someone the campaign should target. Campaigns start by focusing on people with higher points; we start with “10s” and work down. This point system helps determine whose door is worth a knock and from there, you decide how to contact the person – live call, robocall, mailers, ad buys, etc.

Ultimately, the goal of the data-driven campaign should be to understand the humans behind the data, using the data to read which issues matter to people, determine which candidates they support and communicate with them effectively. All political campaigns are about selling a product – and that product is the candidate. That’s why the human element is so important in all aspects of a campaign. Despite the use of technology and data, every campaign can be distilled down to humans trying to connect with and influence other humans.

image credit: Justin Grimes

Like this article? Subscribe to our weekly newsletter to never miss out!

Previous post

What Recruiters and Hiring Managers Are Looking for in a Data Scientist

Next post

Big Data Breaches Shouldn't be a Concern for Small Business (But They Still Are)