Dr. Syama Prasad Mookerjee Research Foundation

Big Data and Smart Politics

By Praful Shankar

Two years after Barack Obama’s historic election to the White House, the Democrats suffered one of their worst defeats in recent history during the 2010 mid-term elections. The Republicans took over control of the House and came extremely close to flipping the Senate, puncturing the halo around President Obama and delivering a huge setback to his legislative agenda. The ‘shellacking’ (as Obama had put it himself) also washed away the perceived inevitability of his reelection in 2012. Suddenly, it seemed as though the ‘Hope and Change’ President would find himself suffering the ignominy of being the first US President since Jimmy Carter to be restricted to a single term.Consequently, Obama went into the 2012 Presidential race in a very different political environment than the one in which he swept to power 4 years earlier. The 2012 Obama campaign team appreciated early on that while the buzz around candidate Obama and the widespread anger against the Bush administration pretty much ensured a wave at the polling booths in 2008, the 2012 campaign would need to work much harder (and smarter) in order to ensure that they were able to sell President Obama’s achievements and motivate constituents to show up at the polling booth, especially in the crucial swing states.

The man managing the Obama re-election bid, Jim Messina, realized that the one distinct edge that the Obama campaign had over their rival’s was in the digital space. The 2008 campaign – one of the first and most successful ones to use the internet and social media for voter engagement- had already done a lot to bring digital technologies into the political mainstream. But this time around a rapidly evolving technology landscape (with the emergence of areas like Big Data, NoSQL and analytics) had presented Messina with the opportunity to place data and digital technologies at the center of the 2012 campaign. According to news releases, the 2012 Obama campaign employed around 100 full time data analysts who ran an average of 66,000 computer simulations every day. The goal, as Messina put it, was to ‘measure everything’.

The end result of this initiative was the domination of Obama in the swing states, particularly in key precincts. Even when most pollsters began to give Romney the edge in popular vote and certain key states, the Obama campaign appeared unruffled with key adviser David Axelrod even betting to shave off his moustache on live TV if Obama lost any of Minnesota, Michigan or Pennsylvania states. Needless to say, the advanced digital targeting and synchronized ground game of the Obama campaign ensure that Axelrod’s moustache stayed where it was.

Data driven politics has arrived

Over the past few years, more and more political players have sought to go the digital route. Both the Torries and Labour employed Obama campaign veterans during the elections of last year as advisors. In the current US Presidential race, Hillary Clinton’s campaign team has looked to build on the groundwork laid by the Obama campaign and leverage the same towards her election bid. In India, we have seen the Narendra Modi campaign make the use of social media outlets like Facebook and Twitter along with its own Mission 272+ platform to engage with voters and prospective volunteers.

As advantageous as leveraging social networks for political campaigning are, the next level of digital engagement in politics – built around fast maturing Big Data analytics technologies – will be even more pervasive and valuable.

With the advent of software and database platforms like Hadoop, NoSQL etc along with the emergence of hardware systems, like HP Vertica (which was used by the Obama campaign), which have the capability to store and process large amounts of data, a new range of opportunities have opened up for political entities, turning things which seemed impossible just a few years go into rapidly evolving realities.

While the potential applications of such technologies covers a rather varied landscape, the major areas of applicability are three.

1. Voter Registration

This is an area which is expected to be crucial in Indian elections in the coming future. A UN study conducted two years ago said that India has the largest population of young people (10-24 years) in the world. A separate study estimated that a whopping 49.91 % of the country’s population in under the age of 24.

These voters – likely first or second timers – tend to fall within the swing voter demographics. With rapidly changing economic and social structures across the country, the traditional lines of caste/religion to party mapping are constantly being diluted. If political parties are looking to form a voter coalition which is disruptive of existing social walls, targeting the youth vote has been shown to be one of the more fruitful methods. This is because young voters respond to being spoken to directly and once convinced, tend to be more resistant to social pressures which may attempt tochange their political viewpoints.

In the 2014 General Elections, the domination of the youth vote by the BJP swung numerous constituencies towards the party and also added a verve and energy to the campaign which set it apart from those of the Congress and others. Even as we speak, US Democrat Bernie Sanders is running a campaign centered on the youth vote – one which the powerful Democratic establishment is struggling to put out in favor of Hillary Clinton.

Also, Data analytics is not just about identifying unregistered voters and getting them to register. Analytics enabled analysis will be able to observe and judge the motivating factors and digital footprint profiles of unregistered voters and identify patterns pointing to which demographics among young voters lean towards which party, on what issues and how they can be engaged. It also helps identify motivational factors (more on this later) for specific groups to go out and register, and can even help in easing the voter registration process by providing constituents with timely updates, directions and contacts.

Obviously, the use of data analytics need not be restricted to just youth voters. Analytics algorithms allow users to slice and dice aggregate information, look at it from different angles and obtain insights which might not have been visible at first look. With inter-state migrations becoming more and more common across the country, analytics driven analysis of voter profiles and registration rates can even help parties create supporter bases in constituencies where none have existed even in the recent past.

2. Voter Persuasion

Among all the advantages that data analytics presents to political entities, voter persuasion is arguably the biggest.

Analytics allows parties to study voter behavior against a multitude of influencing factors – from the more traditional ones like caste, religious beliefs and economic status to completely new ones like internet behavior, emotive issues, celebrity preferences, work profile, reading patterns, social circle etc. Further data aggregation and regression analysis will allow parties to spot specific characteristics of their existing supporter bases and also understand patterns using which they can spot independents who have a high probability of being converted into party supporters. The advantages here are that:

• This provides the ground level party volunteers with valuable information during activities like door to door campaigns, voter mobilization on polling day and volunteer recruitments

• It also allows party campaigns to understand voter habits and preferences, better, enabling parties to better manage their contact points with prospective voters –directing reach-outs through mediums that the voter is more comfortable with and during periods of time when he/she is more receptive to being engaged

• Most importantly, it allows parties to profile voters based on the likelihood of support. Parties will be able to understand the levels of support different sets of voters have towards their candidates and issues, the likelihood that the individuals will show up on voting day and most critically what influencing factors would be most effective in nudging a fence-sitter over the line

For instance, during the Obama campaign, analysts helped the campaign staff identify that a large number of young voters were more likely to turn up at the booth for Obama if someone from their social circle prodded them to do so. Using information collected from social networks, the campaign was able to identify committed Obama supporters within the circles of such independents and get them to push their friends to the voting booth.

Additionally, identifying prospective voters at a micro and macro level along with gaining insights about their preferences will help parties spend campaign funds in a more focused and judicial manner. For example, rather than wasting money buying TV ad time in a channel which is rarely watched by key demographics, campaigns can redirect those funds towards another channel or program whose viewers are more likely to vote favorably or even design a micro-targeted message that is relayed – without intermediaries – directly to the voter’s cell phone.

3. Increasing Turnout

Ratlam in Madhya Pradesh is a constituency which has a history of being a Congress bastion which the BJP had never been able to win. In the 2014 election, Congress candidate for Ratlam Kantilal Buria polled almost 1 lakh votes more than what he got in 2009 when he won. He was still defeated by over another lakh votes by the BJP’s (late) Dileepsingh Bhuria, who managed to mobilize close to 5.5 lakh supporters on voting day – such is the effect that increased voter turnouts can deliver to political parties.

Technology has the capability to be one of the great enablers of this process. Analytics can help in creating an insight-led ecosystem of engagement with the voter, from the initial stages of the electoral process to the last and most critical action of actually voting. Using this process, parties will be able to identify and understand their voter bases and engage with them on a sustained and periodic basis, and create momentum towards the polling day.

This plays an even more vital role in constituencies which are generally close or facing multipolar contest where even a swing of under 1% from one candidate to another can make the difference between victory and defeat. What is more, the possible levels of data penetration and segmentation can even allow parties to track voter mobilization at booth levels, even linking voters to booth-level party workers in order to ensure that the turnout rates remain favorable.

Another advantage is that it also allows parties to dispel rumors and innuendo which inevitably pops up at a local level as the election eve approaches – especially in India where the dissipation of false news on reservations, post-poll alignments and even news regarding candidate’s health have been known to have swayed voters at the last minute in the past.

Beyond Elections

The prospects that data analytics present are not just restricted to the electoral process. It can also bridge the gap between governments and its constituents. Analytics can open the doors to instant feedback of government schemes, tracking of public sentiments on key issues and can even help parties come out with manifestoes specifically catered to key demographics.

Even more importantly, data and digital technologies can make both politics and governance personal. If political entities can create an ecosystem of constant engagement of voters by effectively merging the dynamism and agility of digital technologies with the larger party and governance apparatus, it will allow leaders to be in a state of constant dialogue with the masses.

Such a system of engagement would break the so-called ‘power bubble’ and inform leaders of what the public really thinks and feels. Conversely, it will also allow leaders to cut through the noise of political spin and convey effective messages on their achievements, schemes and larger governance agendas.

In an era where various power structures linked to politics, like the media, bureaucracy and power brokers, at times work to create a wall between the elected and the people, data sciences and digital outreach has the potential to be the ‘great leap forward’, which can redefine the terms of direct political engagement like never before.

(The author is a Chennai based IT Strategy Consultant & a Columnist analyzing national politics)