French entrepreneur and political consultant Igor Lys gives us a short yet pertinent glance at what data in politics is all about.
Data brings change to much more than just the commercial side of our lives. The amount of the information that companies have about who we are and what we are as social units is so huge, that this data reshapes the very fabric of our societies. And, obviously, the politics, as the very structure of the societal governance, is heavily impacted. But is data in politics just a better way to target political ads? Not really.
In Western democracies, we tend to associate politics with the process of rising to, and exercising power. As NBC News’ Chuck Todd and Carrie Dann speculate, big data has “broken” politics, because knowing so much about those whose votes put you into the office — basically, the customers you are selling your services to — allow you, just as it is the case with the businesses, to focus (“target”) your political campaign so well, that you don’t even need to speak to a broader audience. According to them, you just have to motivate those who are already ready to cast their vote in your favor and give a small boost to those who hesitate.
Todd and Dann may be wrong, and big data in politics is probably more than just a better way to buy ads on Facebook. But still, knowing where we spend our time, what series we watch, what books we read, what food we prefer and what words are we most likely to use in our tweets makes the difference. But what is it that makes the politicians “addicted to big data like it’s campaign cash”, as Slate titles in one of their articles? Let’s try to figure it out.
What is data in politics used for?
Sherlock Holmes once said (in A Scandal in Bohemia): “it is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts”. This quote dates from 1891, and yet it is exactly the approach to data, big or small, that people tend to apply today. As Emily Kumler reminds us, “any man in ancient Rome who failed to show up for the census risked losing his citizenship. The census was, after all, a justifiable means of determining taxes and establishing the social hierarchy of the population”. This information was used to ensure better decision making.
Unfortunately, this “addiction” to data has induced politicians and their campaign managers into the same illusion that businesses are struggling with right now: big data allows reliable prediction. Unfortunately, for many of us, the teaching of the mathematical theory of chaos that states, amongst others, that there are elements of the world so sensitive to initial conditions, that we are technologically incapable on integrating them all into our calculations, — this teaching is correct for the political use of data. The election of Donald Trump was surprising for data analysts, but not for the common folk casting votes and getting “the feel of the field”. Similar processes, a kind of return to sources, can be seen in intelligence agencies, who were investing heavily for years into Echelon-type data analysis systems only to discover that what they call HUMINT, human-based intelligence, is sometimes much more reliable and important.
But all this doesn’t mean that collecting and processing big data in politics is useless. It is not. Clearly, the smartest people don’t use it the way lots of journalists and social activists think they do.
Most people believe — because of huge public buzz scandals like the one of Cambridge Analytica — that big data in politics serves the goals of a better manipulation. That evil agencies, allied with cynical campaign leaders and why not the Russian government, analyze your behavior so they can better manipulate you into loving the right candidate and hating the wrong one. This is partially true: indeed, the phenomenon of putting you into an “opinion bubble” by better targeting political ads and thus motivating you to actually go and vote, and become (unknowingly) an ambassador for the power that has you in its aim, exists. My experience of working with executive political figures in several countries shows that the real use these people make of massive data collection and analysis concerns less the prediction and the manipulation of the future result, than the better analysis of the already existing ones.
In the media sphere, the data collection and processing when it comes to political and social phenomena are almost exclusively associated with vote manipulation and Russian interference. The very famous scenes of Mark Zuckerberg being questioned in the U.S. Senate have traveled around the globe. The fact that Facebook and other social networks collect data on us is presented as something outrageous — for example, here is how the Investigation into the use of data analytics in political campaigns report of the UK Information Commissioner’s Office, published in November’18, states it:
“We have concluded that there are risks in relation to the processing of personal data by many political parties. Particular concerns include the purchasing of marketing lists and lifestyle information from data brokers without sufficient due diligence, a lack of fair processing and the use of third-party data analytics companies, with insufficient checks around consent.”
In reality, the same problematics exists since years, huge amounts of personal data being sold to corporate clients. And yet, we only start panicking when we see how the illegal, or barely legal trade of our life patterns collected by social networks impacts our political choices. It is, indeed, a problem.
But I believe that it would be irresponsible to limit our understanding of the influence of the data on our socio-political behavior. The issue of data collection in the interest of the political actors must not be reduced to just cynical Frank Underwood-style power brokers buying data on where we eat and what we watch on Netflix and who our friends are to better sell us their quotes about how they are gonna make our lives better.
Data in politics is not just a way to target ads. Because we are talking about more complicated transactions than just selling a certain type of product with a unique value proposition, we have to acknowledge that our data has much more than just “one-shot” value.
Berlin-based organization Tactical Tech defines three main types of data interpretation in politics, which we will see in details:
- Data as a Political Asset: valuable stores of existing data on potential voters exchanged between political candidates, acquired from national repositories or sold or exposed to those who want to leverage them
- Data as Political Intelligence: data that is accumulated and interpreted by political campaigns to learn about voters’ political preferences and to inform campaign strategies and priorities, including creating voter profiles and testing campaign messaging.
- Data as Political Influence: data that is collected, analyzed and used to target and reach potential voters with the aim of influencing or manipulating their views or votes.
Data as a Political Asset
Early June 2017, former U.S. presidential candidate Hillary Clinton gave a speech at the CODECON conference, blasting the Democratic Party for its poor data gathering skills. “I inherit nothing from the Democratic Party. I mean, it was bankrupt. It was on the verge of insolvency. Its data was mediocre to poor, nonexistent, wrong. I had to inject money into it.” Clinton said.
For a lot of people, this comment was a revelation: political parties have data pools! Previously quoted American entrepreneur Emily Kumler talks about the “assumption that we’ve accepted the use of scientific methods expressly designed to mislead us, to reinforce fears and alliances we already have, without any accountability, as a condition of our current political process.” Because, yes, data pools exist and are used as an asset in political transactions. They are being collected by third parties and sold to entities that participate in elections. In Europe, where the political systems in most countries allow a more diverse offer than the American bipartisan convention, new parties and movements are sometimes created for the sole and unique reason of collecting data of a certain type of voters.
Here is how Tactical Tech presents this type of data:
Data as Political Intelligence
Some types of data, for example, your psychological profile, is not used to target ads at all. It rather helps decision makers in their daily job of adjusting ongoing or planned activities. In an electoral campaign, Data as Intelligence allows efficient clustering of target audiences. For a politician in office, having actualized information on different layers of the population allows to better envision the possible response to a certain public decision. No need to go far for a great example: the current turmoil with Donald Trump in a deadlock with the Congress over his wall on the southern American border is a schoolbook case of how data reinforces decision making. The White House team which is, since the Cambridge Analytica story, in love with data, would never decide an Oval Office public message without disposing of very deep insight in the numerous cohorts of the U.S. population. The message to deliver to this population is designed to reinforce the core Trump electorate, to plant the seeds of doubt into the minds of those who have no strong position, and to exalt the Democrats so they do or say something stupid.
It is impossible to do so without having intelligence on your audience.
But this type of data is also used in the inner power games. When making, for example, staff-related decisions, one can tend to appeal to the data to justify firing — or hiring — a certain individual. In tough negotiations with political opponents, having data on his or her voters can provide a crucial argument. This works also on an international level.
Data as Influence
This is the part that everybody speaks about. Data that is used to directly influence your decision. And we are not just talking “Vote for me” banners appearing in your Instagram timeline. Based on your profile and the goals set up by the campaign team, you may, for example, find yourself seeing a message about long, horrible lines and the ballot station in your district that will discourage you from voting. And those messages will seem like a legit news outlet production, with no connection to any candidate whatsoever: in the difference with the TV-related legal guidelines, online advertising has basically no obligation whatsoever as per disclosure of the paying entity. And while in the E.U. the dynamics shifts towards stronger, deeper regulation, the U.S. and for example African online ads markets stay extremely liberal (for example, see here about how big data was used at the elections in Kenya).
Here is what TT gives as examples of Data as Influence :
The main question
After having read this simplifying resumé of different digital practices in political data management, one can become anxious and join the legions of online commentators panicking about the overwhelming power of the big brother that tracks our every step. The reality, though, is less alarming than that. Most of the campaigning still happens offline. There is serious academic evidence that old-fashioned political debates and rallies have a major influence on the voters’ decision. Some fractions of the population are completely insensible to the issue, as they produce little date (like the elderly generation in some countries).
So the main question about big data in politics is: does it actually work?
And the answer is actually quite surprising: we don’t know! I mean, on some level, we know that it works. Cases like Barack Obama or Bernie Sanders in the U.S., Jean-Luc Mélenchon and his movement La France Insoumise in France, or the aggressive Internet campaigns of the far-right German party AfD, show that data and digital campaigning brings results. However, there is no deep study of how much of an actual impact this type of campaigns gives. The ROI of such actions are yet to be calculated. In America, all the governmental services could not provide an accurate estimate on what exactly was the quantitive impact of the Russian meddling.
One thing we know for sure is that the clear trend of getting more and more data involved in political campaigning and decision making is there. We also know that the legal framework has not evolved as fast as have the practices of the data brokers. As Dr. Kate Dommett, director of the Crick Centre at the University of Sheffield, states, “although digital campaigning tools may help parties to win elections, they also have significant side effects for the way democracy occurs. If societies value equality of information, open debate, and transparency, these trends should be of concern. But if the emphasis is instead on engaging voters on the issues and ideas that matter to them (targeted campaigning at its most basic), new practices may be of less concern”.
Long story short: we have to be very attentive to how the practices of data collection evolve. If they are organized in a transparent, comprehensible and responsible way, this data — these assets, intelligence, and capital of influence — may actually increase the efficiency of our democracies. For the first time in recent history, France is launching in 2019 a national debate on the new ways of governance, hoping to involve most of its population into a discussion on the new ways of exercising power. It is a perfect occasion to test all the good that data can provide to the people, not by enclosing individuals into their opinion bubbles, but rather giving them more information on topics that interest them, and on political figures who represent their visions of society.
As Dr. Dommett rightfully says, “without considering these questions, there is a danger that any response may have unintended consequences and fail to advance the principles we want to uphold”.