Can Narendra Modi win elections using Big Data as Trump did?
How Big Data works in politics is fascinatingly relayed in the article.analysis Updated: Feb 06, 2017 15:13 IST
As the world is preoccupied with the dramatic early days of Donald Trump’s presidency there is an article circulating on social media which is shocking readers.
Originally published in Zurich-based Das Magazin, the article by Hannes Grassegger and Mikael Krogerus explains how Big Data (our digital footprint) is being used for innovative political marketing – and how it may have helped Trump win. This is important because the use of big data is profoundly affecting how election campaigns are run, and may even determine winners – including in democracies like India, where parties like the BJP, Aam Aadmi Party (AAP) and the Congress are increasingly relying on digital strategies to influence voters.
How Big Data works in politics is fascinatingly relayed in the article.
A shorter summary goes like this: Companies that record online behaviour are able to build psychological profiles of users, which are used to display ads that are suited to their audience. Michal Kosinski, now a professor at Stanford, made startling discoveries while doing his PhD at Cambridge University when he and a fellow student designed a Facebook app called MyPersonality – that used a psychometric questionnaire to evaluate humans based on five personality traits known by the acronym OCEAN: openness, conscientiousness, extraversion (sociability), agreeableness and neuroticism (degrees of sensitivity).
Millions filled out the survey; Kosinski then correlated psychometric results with other data of users including their Facebook behaviour, what they posted or liked, and was able to predict with high-levels of accuracy their sexual orientation, religious affiliation and political preferences. Over time, Kosinki found that he could evaluate a person better than friends, parents and even partners. As Grasseger and Krogerus write, not only could this technique create psychological profiles but that the data could also be used to search for specific profiles, say “all anxious fathers, all angry introverts…maybe even all undecided Democrats.” “Essentially, what Kosinski had invented was a sort of a people search engine.”
Over time a company called Cambridge Analytica started using methods similar to the Kosinski model (unbeknownst to him) and deployed targeted ads in Brexit’s Leave campaign, the US primaries and eventually in Donald Trump’s online campaign.
Cambridge Analytica maintains that it does not use Facebook data; its spokesperson said “psychographics was hardly used at all”. But in a 2016 lecture at New York the company’s CEO Alexander Nix prominently featured psychographics in his talk (see 5:38 in this video).
He said his company has profiled the personality of every adult in the US – 220 million people through a combination of demographics and attitudinal factors, revealed in consumer and lifestyle habits. This personality profiling was used for targeted political marketing through ads on social media platforms. Profiling helps with identifying relevant issues and creating specific ads for citizens.
A neurotic introvert would, for instance, see an ad on the need for guns with an image of burglar breaking in, while a more well-adjusted individual would view images of a father and son on a shooting trip. Like other marketing entities, the firm thinks running one ad for an entire demographic is pointless since customised ads for individuals were more effective.
Cambridge Analytica divided the US population into 32 personality types and focused on 17 states. They ran ads to suppress Democrat voter turnout in Florida, such as highlighting Hillary Clinton’s comment in 1996 referring to perceptions of black men as “super-predators”. Trump’s team tested 175,000 ad variations for Facebook culling arguments that he used in his debate with Clinton, fine tuning them using “different headings, colors, captions, with a photo or video.”
Nix claimed to reach villages, apartment blocks in a targeted way, “even individuals.” He has said that profiling through online data can be correlated with cable TV viewing habits, so viewers can get tailored ads during television shows.
Nix claims that his company played an integral part in Trump’s victory although this has been challenged on the grounds that there is no clear evidence of this yet. But there is little doubt that profiling and targeted content can be powerful ways to shape the political imagination of urban audiences as they can be used to lure users towards material that is in line with their own ideological preferences. This possibility is real given the phenomena of fake news sites on Facebook and the proliferation of Facebook groups.
Cambridge Analytica’s work has, not surprisingly, generated interest in political classes elsewhere, including from Switzerland, Germany and Australia. Countries may even see working with the company as useful for access to the Trump administration. Cambridge Analytica has Trump’s senior adviser Steve Bannon on its board and a key investor in the firm is Rebekah Mercer, daughter of Robert Mercer, a hedge-fund manager and supporter of Trump. The company is exploring newer markets and it is unlikely that India is not on their radar.
The Echo of India, a newspaper published from Kolkata and Port Blair, interestingly reported that Nix was due to visit India in January and that he has an interest in elections in Uttar Pradesh, Gujarat and Punjab. There is also other reporting that Nix has advised on ways to harvest the Indian American vote for Trump.
It is not clear if Nix visited India or if he is working on state elections but – if used – the effects of his company’s methods could be far-reaching for India’s democracy for several reasons:
One, neither politicians nor political analysts have a firm handle on what people think about politics. For instance, the commentariat is unsure about popular opinion on demonetisation and the outcome of UP elections.
In this scenario, a political party’s best bet is in the constant effort to shape the media narrative. The BJP has a fairly favourable mainstream print and TV media environment but their reach is uneven and difficult to quantify. The contest has, hence, shifted to social media where the potential for parties to reach audiences directly is impressive. India had about 350 million smartphone users in 2015; that number is expected to go up to 702 million in 2020. India reportedly had about 195 million users of Facebook in 2016.
As noted, the BJP, AAP and others, therefore take social media seriously. As is well-known online campaigns played a major role in brand building Modi en route to power. Arvind Gupta, head of the BJP’s IT cell during the 2014 elections, said that for a period of 18 months, particularly between December 2012 and February 2014, the party’s “primary campaign was on digital and social media.” The planning included 3-4 years of meticulous data collection.
“We had data on each of the 543 constituencies. We knew how many mobile and Internet users were present in each constituency… we used analytics to understand which polling booths had voted for the BJP in the previous elections…for each polling booth data analytics was used to segregate voters into blocks to determine who were pro, undecided or against the BJP…We first used deep analytics to understand group communication behaviour and then used appropriate technology to communicate with them.”
Modi’s online influence has only grown with time. He now has 26.9 million followers on Twitter (up from 8.5 million in 2014), more than 39 million subscribe to his Facebook page and he has a loyal fan based that aggressively trolls critics. Digital is playing a big part in BJP’s campaign in UP, with a team dedicated to “data crunching, seat analysis, digital media and operations.”
This includes handling a Facebook page that reaches more than 10 million people, and sending out four WhatsApp messages daily to reach 500,000 users in the state. AAP is also a formidable force online. It claims that its Facebook page reaches 150 million users, “the biggest social media page in the world” in its view. Arvind Kejriwal has seven million followers on Twitter and Facebook and a 2016 speech by him in the Delhi Assembly was reportedly watched by seven million on its various Facebook pages.
Winning the influence game online will depend on the quality of data and the amount of resources a party has to create and market tailored content. On both these counts the BJP has advantages. It obviously outspends every other political party in India and is best placed to take advantage of a “people search engine”, as it develops.
Data is in any case widely available – and thanks to demonetisation there is now an explosion of privately-controlled data in India owing to the government’s nudge to people to use apps for payments. Moreover, India does not have a privacy law, there is in fact a default opt-in principle in place when it comes to data sharing and, on average, each person has 32 apps in a smartphone. All this amounts to a very conducive environment for personality-profiling and ad targeting.
We do not know what form online political marketing will take hereon but it will no doubt be innovative and relentless, as it is the one shot political parties have for defining the political debate. This will likely end up polarising society further rather than nudging it in constructive directions since users will be consistently exposed to content that confirms their beliefs. Yes, there will be independent voices (including from mainstream media) that will challenge partisan views but on the whole, citizens will be primed by the content they are most exposed to. The BJP is likely to do better than the others in in this fractured scenario as it has both the polarising messages to keep audiences interested and the resources to reach newer markets.
There is a lesson for citizens in all this. If Indians do not take privacy as a right seriously then they will have less of a choice about who influences them. This is because online browsing habits reduce the possibility of one being exposed to knowledge that is not recommended by algorithms.
The use of Big Data promises us relevant material but paradoxically limits our vision. It constrains our contact with intellectual and social diversity – the building block for empathy on which democratic instincts rest. To put it differently, technology as we use it poses an unyielding threat to the possibility of having an informed citizenry, which is vital for a democracy to thrive. India thus needs a strong legal framework for privacy as a starting point for regulation that addresses these issues. Else content marketing firms and parties using them will be defining the future of India’s democracy.
The writer tweets as @SushilAaron