First Among Equals

 In AI, ML and Data

Data driven Politics

The US Presidential election has the highest of stakes. And it is no wonder that technology plays an integral part in the game. Social data was billed as the winner of the 2008 and 2012 presidential elections when Barack Obama’s campaign team used social media data analytics to influence the voters. In 2016 the question has changed from, “Will Big data disrupt the elections?” to “How much will Big data disrupt the elections”.

In the 2008 Democratic campaign, Dan Wagner, an expert in econometric forecasting, handled data analysis for the State voter files. Throughout the primaries he gathered State voter data and used statistical modeling to sort the electorate. This MIT study details how each voter in the country was assigned a pair of scores, one based on the probability that the voter would cast a vote and two, that the vote would support Obama. The data to generate the score was gathered through traditional survey interviews, past voter registration records, consumer data warehouses, and past campaign contacts. There were as many as 1000 data variables for each voter. The data was used to create micro target segments where volunteers could have scripted conversations with targeted set of voters. Obama’s campaign was also extremely agile. They quickly fed back into the system their learning to update the model  and have even more meaningful conversations with the next set of voters. In comparison McCain’s campaign ran the statistical model only once and the changing dynamics of the voters were not considered for course correction. The Democratic campaign data team put Obama in the White House by knowing at an atomic level each of the 69,456,897 Americans that voted for him.

Just 2 years later in the 2010 midterm elections the Democratic party suffered some of the biggest losses since the great depression leading to one of the largest seat changes and control of the U.S. House of Representatives passing to the Republican party. Interestingly Dan Wagner had predicted these results 5 months before the midterm polls. He achieved this by using technology to track public opinion, and identifying the sample that could be extrapolated to represent the whole.

Smarting from the 2010 losses, the 2012 reelection campaign had a dual agenda. Get the voters that voted in 2008 to vote for Obama again and mobilize new voters for any voters that had defected. They experimented with the messaging and the targeted audience to find what message resonated with which micro-sector. Way before the term was associated with data and analytics, the campaign team leveraged ‘extreme personalization’. Individual behavioral analytics based on interests, passions, affiliations, opinions, and demographics. As an example consider this ad advocating equal-pay for women. This was used to win over Republican women voters who were at odds with their party on gender concerns.

Fun fact – Campaign donors that described their occupation as data and analytics were sent invitations to apply for work in Wagner’s analytics department for the 2012 Democratic campaign.

Cut to 2016, the election cycle is heating up. For each campaign, digital is the secret sauce, and each campaign is keeping their data strategy a closely guarded secret. However I see the same challenges and opportunities that are being faced by traditional Enterprises presenting themselves to politics. Just replace customers with voters.

The primary goal of every campaign is to identify, connect with and engage voters. Campaign movers and shakers are focused on finding reliable and consumable data from multiple sources, unifying it, and gathering real-time and actionable information to influence voters. In 2008 this voter engagement model was ‘reactive communication’. 2012 it evolved to being ‘proactive and targeted’. 2016 the voters are empowered and the engagement model demands ‘collaboration’. Till 2012 the campaign messages were organized and controlled. Powerful people owned the message and controlled it through TV and radio. Now, Internet and social media have transferred control to the streets. Anyone can post a message and that can influence millions. The establishment has limited control and is trying to use digital technologies to get insights on what happened, why and how to prevent or facilitate it.

They also have to contend with restricted campaign budgets. Even the campaigns with deep pockets seem to be using more COTS (commercial off the shelf) software and cost effective cloud computing based software to get more bang for the buck. The Ben Carson campaign, that recently ended, used HubSpot to maximize their outreach and web presence while leveraging existing APIs to lower costs. Similarly Hillary Clinton leveraged to acquire voters in the same way that businesses acquire customers.

Early on in the buildup to the campaign, most teams tied up with third party voter database companies like L2, Aristotle, or GOP Data Trust. And started leveraging targeting products like Targeted Victory, Cambridge Analytica, or Deep Root Analytics to analyze the data. As per this FastCompany article the Cruz campaign has the most sophisticated data strategy. In addition to the Data team, the analytics solutions and the micro-segmentation they even have a gamification app to gamify support for the #cruzmissile.

And just when all the campaigns seemed equally tech savvy one candidate is bucking the trend or trumping the bend. The Trump campaign did not include data till as late as October of 2015 and even today is way behind in leveraging data as compared to the competition. Does the campaign believe that his brand value and product are so compelling that they can afford to under-invest in analytics?

Time and data will tell!

Image Credits:  

Interested in learning more about our offerings and solutions?
Let's connect
Recommended Posts

Start typing and press Enter to search

Contact Us
close slider
Contact Us

Yes, I would like Persistent to contact me on the information provided above. Click Here to read our full Privacy Notice.