Elections: the role of Big Data in political campaigns

22 February 2018

Charisma, leadership, but also Big Data and the analytical skills to exploit them. These seem to be the qualities that make up the profile of the winning candidate, according to what emerges from the latest electoral experiences faced by citizens around the world. In June 2013, The Washington Post addressed Obama as The Big Data President, bringing to the fore the prelude to an era in which elections and referendums are won by ad personam propaganda.


The campaign for the re-election of Obama has gone well beyond the usual use of database of voters, profiled according to the classic subdivision in socio-demographic categories. Thanks to an acute use of the information obtained from Big Data, his staff was able to evaluate, on the basis of preferences and habits expressed online, every single voter. The application of data mining techniques made it possible to become aware of the propensity to vote of individuals, towards which a meticulous microtargeting campaign was then directed.


Perhaps the most striking case that saw Big Data as the undisputed protagonists in the construction of a communication strategy for political use, is that of the 2016 presidential elections in the States. Thanks to the use of sophisticated analytical techniques and a surgical approach to digital marketing, Donald Trump was able to convey his electoral messages in the ways, occasions, and in the most appropriate places to reach each elector. Taking for example the favorable position of Trump on the issue of the possession of weapons, it has been proposed as a ‘defense’ to the frightened citizens who live in dangerous neighborhoods, ‘constitutional’ to the purists of the Ten Amendments and as ‘tradition’ to the lovers of sport hunting. In short, the same message, a substantially different value for the recipient. And for the candidate, considering a victory still today unexplainable by many.


The company behind Trump’s online campaign, which specializes in data analysis and strategic communication of electoral processes, is the same one that had worked for the Leave.EU party in the very early stages of the Brexit campaign, Cambridge Analytica. Founded in 2013, it is an American extension of SCL, a British company experienced in human behavior predictions through data collection, analysis and use. Among the main customers of SCL we find the Ministry of Defense of the United Kingdom, the US State Department and NATO, which has used its services to understand how to neutralize online Islamic recruitment.


Will Big Data play a fundamental role in the upcoming elections in Italy? It would seem so, according to statements made by Fabio Martini, a journalist of La Stampa who during the presentation of his latest book claimed that one of the political parties running for the government of the country has turned to Cambridge Analytica for a consultancy. Nothing else is known for the moment, but on the website of the company it emerges that there has already been, in the past, its involvement in Italian political life. In 2012, in fact, Cambridge Analytica has worked or a political party that had its latest successes in the ’80s, bringing it to a result far above expectations.


“We find your voters and put them into action,” reads the slogan on the Cambridge Analytica website. In short, Big Data inserted in a political context are transformed into a weapon of consensus. But how does the company, captained by the financial analyst Alexander Nix, manage to offer its clients electoral campaigns on the edge of tailoring? The method used combines psychometric analysis, the field of psychology that specializes in measuring the psychological characteristics of the individual, to Big Data. From this intersection generates OCEAN, the model based on the analysis of the Big Five: openness, conscientiousness, extroversion, collaborative approach and emotional stability. The analysis of these five basic variables made it possible to profile 220 million Americans during the Trump campaign, understanding their needs, fears and likely behaviors, and then translating them into personalized dark posts that were only visible to selected audiences. The company is so convinced of its method that on its site promises to be able to reach about half of American homes with “a precision that does not include waste”.


Where do the data needed to accomplish this new miracle of digital technology come from? Big Data differs from traditional databases for three characteristics that must necessarily be present at the same time: large volumes, speed and variety of sources and nature. The Internet, with its 450 million daily B2C and B2B transactions expected by 2020, is obviously the primary source of this extraordinary resource.


Amazon, Facebook, Google and others have long used the profiling generated by the analysis of Big Data to create effective marketing strategies, to which we are already partially accustomed. The data collection strategies put in place by politics are much less known. For example, for the Trump campaign, a hundred people were paid to complete a questionnaire through the Mechanical Turk website. Citizens, in order to receive their payment, had to download an app that was able to access, in a completely legal way, their Facebook profiles and those of their friends. The data provided by third-party companies were also added to the huge database offered by Facebook. Cambridge Analytica has thus succeeded in identifying Trump’s probable voters and creating targeted electoral messages, to the point that the current US President has spent less than half of the $ 521 million of his adversary Hillary Clinton to win.


An uncontrolled use of Big Data hides numerous dangers, among which the risk of becoming accustomed to pre-built information and aimed at specific purposes, where the citizen finds himself faced with a continuous, manipulated confirmation of his beliefs about the world (Confirmation Bias). The protection of privacy, freedom of choice, and equal opportunities, are issues of fundamental importance, which are increasingly protected in national and international legislation. An example is the GDPR, the European Regulation 679/2016 aimed at the protection of the Privacy, which will become directly applicable by all member states starting from 25 May 2018.


The warning to a more controlled use of Big Data also comes from the academic world. In 2008, Professor Michael Kosinsky, back then a researcher at Cambridge University, used the OCEAN model to produce a questionnaire published through the My Personality app. The accuracy of the forecasts emerged from the analysis of the participants’ Facebook profiles was so surprising, that Kosinsky himself decided in 2014 to publish a research on the possible risks these tools entail in terms of control, manipulation and privacy.


“Our children will not be able to understand the concept of mass communication,” says Alexander Nix, CEO of Cambridge Analytica. It is to wonder if they will be able to understand the difference between their own thought and the one suggested by an algorithm. Especially when it comes to voting.




Bologna Business School offers to those who want to approach the challenges of the future with the right skills, programs designed to train the specialists of today’s and tomorrow’s technologies.



Back To Top