Every year, hundreds of millions are spent in predicting election outcomes the world over. There's a plethora of psephologists and experts who have spent decades trying to master this art. But, of late, many of their predictions have proved wrong.
The recent US Elections, when almost nobody predicted a Trump victory, is a prime example. Back home in India, just one pollster was close enough to predicting the thumping Modi win in 2014. The predictions for Brexit were also no different.
What if there was a relatively simple way to predict these outcomes and is more accurate? When I discovered it, I felt like I had been searching for a treasure trove which was lying in my own backyard. So, without much ado, here’s presenting the new age election predictor – Google Trends.
An empirical analysis shows that the search term which was leading in the run-up to the elections typically ended up winning. The way Google Trends work is that it shows up the relative search frequency of a particular search term for a selected region over a period of time.
Google Trends Backed Trump
Let’s begin with the big one – the US presidential elections of 2016. Going by Google Trends, Trump clearly led Clinton throughout the year in the United States before the elections.
The thumping majority for Narendra Modi in the 2014 elections is clearly reflected in the huge interest he enjoyed in comparison to Manmohan Singh, the then Prime Minister.
For those of you who think that Manmohan Singh was a relatively quiet Prime Minister, here is some data from the run-up to 2009, when he led over his competitor LK Advani and ended up winning the elections.
Looking at State Elections
Let’s shift focus to state elections held recently. In the Delhi polls, BJP was tipped to win by psephologists. But AAP emerged winner, and Google Trends data had already predicted this.
The next big election was in Bihar. Google Trends data shows a neck-and-neck fight between Nitish Kumar and Narendra Modi. But when one adds the interest in Lalu Yadav, who was an alliance partner of Nitish Kumar, it clearly takes him over the top.
What Google Predicts for UP
Now, it is worth theorising on why this is happening. Marketing guru Philip Kotler wrote that prior to capturing market share, companies need to capture mind share. Google Trends data in a way shows who has captured mind share and to what extent – and has been proven right many times.
Now for some Google Trends-based predictions for the elections in UP, Punjab, Goa and Uttarakhand.
Looking at the data for the UP elections, BJP clearly seems to be leading. Even after adding the interest for SP and Congress, they have largely lagged behind BJP and only recently have taken a 1-2 point lead. BSP is nowhere close.
This means that BJP will emerge as the single-largest party, but it is clearly a close contest. It could end up being a hung House in UP.
As for Punjab, an interesting story has been unfolding. The interest in AAP has waned recently and the Congress just might surge ahead. Based on the data, it is likely that the Congress will win, with AAP taking second position.
The data for Goa, meanwhile, points to a BJP victory, with the Congress bagging the second spot.
In Uttarakhand, BJP has a clear lead over its rival and is likely to win.
Though I am no pollster, I hope this provides some insight and creates a basis for further analysis, all the way down to city-level data.
I also hope that this places an interesting tool in the hands of those with an interest in predicting election results. For those of you who want to play around with this tool, here is the link: https://trends.google.co.in/trends/
(Rahul Gupta is an IIT Kanpur, IIM Bangalore and Cornell University alumnus and has worked at Goldman Sachs, Hong Kong. He was part of the 2014 Modi Campaign and led the ‘Chai pe Charcha’ campaign. He currently runs a startup in Gurgaon, www.inoneapp.in, and is an advisor to www.getzyme.com. He can be reached @rahulngupta. This is an opinion piece and the views expressed above are the author’s own. The Quint neither endorses nor is responsible for the same.)