Why relying on big data strategies is risky for political parties

16th May `18, 11:52 AM in Government

Elections are like writing an exam! There are only two kinds of people who succeed in the exams…

Jijo Thomas Contributor

Elections are like writing an exam! There are only two kinds of people who succeed in the exams – the people who work hard and the people who work smart. Elections are no different.

To win the elections, you either have to work hard or be smart. In the past few years, the big data analytics have helped various political parties across the globe make smart decisions during their electoral campaigns.

Do you remember 2010’s US Presidential Election where Barack Obama won it over Mitt Romney? Well, it was the first time people started talking about the potential of big data analytics in helping politicians in their electoral campaigns. And guess what, now every other party all across the globe is trying their luck with big data analytics.

Big data analytics in the field of politics

For years, marketers have used big data analysis to create smarter strategies to engage customers and generate new leads in the market, and have succeeded. Even in politics, some strategies influenced by the big data analysis (like US presidential election campaign or the Brexit referendum) seem to work.

For political parties, it was not possible to go from door to door and reach every voter before planning their campaign strategies. With the help of big data analysis, they could easily read the behaviors, mindsets and the preferences of the citizens at large. This is what most of the political consultant companies are doing to help their clients in setting up a successful election campaign.

However, it was recently revealed that the British political consulting company, Cambridge Analytica, gained access to the data of 50 million+ Facebook users. This data was later “misused” for political advertisements during the presidential electoral campaign in the US and the Brexit referendum campaign, as reported in the Economic Times.

When the security of private data is violated on such a large scale, it is meant to stir controversy, which it did. However, one cannot deny the fact that well-made political ads, which address contemporary issues, can have a significant impact on the voters if they are properly circulated on Facebook. So, yes, there is a possibility that big data analytics can be helpful for the political parties to some extent.

Then what is this risk people are talking about?

Experts have expressed their doubts about the efficacy of the big data analytics that uses social media data. It is not clear what kind of strategic decisions can be made using big data analytics? Popular social media websites like Facebook, Twitter can be used very well to gather data, but the collected data can be quite complicated to handle.

Social media data is often influenced, which can lead to flawed results in the analytics. Besides, finding the necessary information on such vast network can be as difficult as finding a needle in the haystack. However, the problems do not end here.

While analyzing the data, a lot of “nonsense correlations” among different variables can be identified. And as the number of variables rises, the rate of likewise correlations also surges. As a result, it often becomes too hard to recognize the various forms of causes and effects. So one can clearly see how unreliable the resulting data can be.

Nick Heudecker, who is an analyst at Gartner (an American research and advisory company), told Tech Republic that the failure rate of big data projects is close to 85 percent. Even though he was talking about the field of business, it clearly suggests where the big data analytics stand in today’s world. The experts are now using several models and algorithms from various fields of study (like statistics, mathematics and engineering) which, however, haven’t done anything new as of now. In fact, it displays more “nonsense correlations” than its previous approach.


The instances of the failure of big data analytics are way too many. The biggest one, however, remains to be the Google Flu Trend experiment (2008) that tried to predict the prevalence of flu just by analyzing the search histories before the government authorities did. In elections, where the stakes are incredibly high, relying on big data analytics can cause a major blunder.

Reportedly, big data itself is not very useful to people. It can only be effective if the analytics are applied properly to the data. It won’t be wrong to suffice that relying on big data analytics to frame the strategy is nothing less than a gamble, where the chances of succeeding are very low.

However, most political parties are putting their trust on various political consulting firms these days which use big data analytics to help them strategize the campaign. And since the political parties don’t have a better choice at this moment, they may just stick with the big data analytics for longer period, even if it is full of uncertainty.