In recent election cycles, some observers have been prone to underestimating or undervaluing the power that data has come to wield in political elections. Campaign strategy that was once driven mainly by geography and demographics has undergone a transformation with the evolution of data science and advancements in computing technology. Campaign strategy now often includes “micro-targeting,” where candidates use quantitative systems and tools to evaluate the efficacy of a range of political actions on target constituents. President Obama’s highly data-driven re-election campaign demonstrated the power of this trend, as a 100-person data science team churned through terabytes of data to gain a competitive edge.
Particularly after witnessing the success the Obama campaign experienced in leveraging big data, candidates have increased their focus on and use of data to capture votes. Ted Cruz and his team leveraged personality-based data modeling and psychographic testing on his journey to becoming the first Hispanic to win a Presidential caucus. Hillary Clinton’s campaign is applying predictive analytics alongside demographic statistics to build personalized emails and targeted ads. On the flip side, Donald Trump has been mostly dismissive of the power of data and analytics in campaigns, instead focusing more on his overall image and brand to drive votes.
Micro-targeting is a concept that has been around in the corporate world for some time: companies often use data to segment consumers and identify offerings specific to their needs. Candidates are now using this same technique to segment potential voters into distinct groups who value different things so that they can fine tune their messaging in a way that will resonate most. Candidates can group potential voters into segments using data such as age, ethnicity, religion, career, geography, home ownership and even more unique data points such as magazine subscription, mobile app usage, pet ownership, and Facebook likes. Each variable acts like a puzzle piece that when combined with other variables can help inform candidates of the overall picture that is most important to that individual or segment. All of this data is collected through a variety of channels, including historical voting records, campaign-specific surveys, exit polls, and third party vendors The data from these channels are then organized into data sets within a data warehouse and structured for mining, aggregation, analysis, and quick utilization.
After candidates have captured and organized all of this voter data, they can leverage a number of statistical and analytic methodologies to identify specific sub-segments of the voter population and, more importantly, learn what messages resonate most powerfully for them. For example, a candidate may learn that Latinos in Colorado who own a home and have more than three magazine subscriptions respond strongly to messages about higher education – a correlation that may not have been self-evident without the use of data. This gives candidates a strategic advantage as they now have the ability to effectively fine-tune messages in hopes to capture enough votes to win a contested district or state.
Not only can candidates use big data to target their messages, but they can also adjust when their messages are received. Media consumption data can be leveraged to learn that, for example, parents tend to listen to radio much more frequently between 3-4pm each day as they are picking up children from school. Campaigns can target their media and advertisement spend on specific avenues as opposed to a blanket approach which can win votes and save dollars, generating better return on investment.
The bottom line is that political messages you come in contact with are most likely a result of big data. Your past online behavior may have altered the campaign ad you came into contact with as you scrolled through your Facebook feed. The political mailer you received last week may have been sent to you based on the car you drive and your monthly auto payment. Ultimately, the deciding factor in this upcoming election may not be who has the most appealing platform, but who can distribute it in the most targeted and efficient way through the use of big data.