If you have ever been on Facebook and seen a survey that some of your friends have done, like a personality test, or an intelligence test and you completed it – thinking it would be fun to see how you score – you are among many millions of others. But you might be more than a little surprised how accessible that information is – and how that very same data was used to influence the outcome of the presidential election. No laws were broken – no Russian spies or anything along those lines – that’s quite frankly what makes it all so scary.
But I am getting ahead of myself.
The term Big Data is often used as one of the most positive things to come along since penicillin. Finally, with enough data, and the right software, computers can expose patterns the human eye cannot see at a speed few could have even imagined just a few short years ago. This TED talk highlights some of what’s possible and what’s coming. Incredible power in unlocking these insights.
But with any power worth having, there are good uses, and bad uses. And lately, some bad, and scary uses have emerged – and they are likely to create some severe backlash in how people think about Big Data. More importantly – it is going to lead to some big shifts in what we think is harmless in terms of what we share about ourselves.
This is one of the most amazing articles I have read in recent memory. It explains how simple surveys on Facebook enabled Donald Trump to win the election using the OCEAN method of classifying personality types so that they could be very targeted in their marketing and spending. Trump spent a tiny amount of money compared with Clinton, and won the states he needed to win the election.
In short, OCEAN is an acronym for five different personality traits of openness, conscientiousness, extraversion, agreeableness, and neuroticism, and simple surveys completed by millions on Facebook allowed those people to be “bucketed” into one of 32 different categories. From that, anyone trying to influence one of those people to do anything different, whether it’s buy a Big Mac or vote for Donald Trump could be tailored to their specific fears, motivators, and weaknesses.
It makes perfect sense – but it gets pretty creepy pretty fast.
Almost ten years ago – Louise Story wrote of the coming shift from massive spending on untargeted marketing, to much smarter spending on much more targeted marketing. Untargeted marketing was all there was until there was better data about the buyer, and a better vehicle to connect in a more intimate way with that buyer. That day is here and Cambridge Analytica is the tip of the spear in wielding this information – this new superpower. The fact that Clinton won the popular election by over three million votes, but Trump got the votes where he needed them most only amplifies the message of just how successful Cambridge Analytica was in their efforts using the OCEAN method. Talk about winning the battle and losing the war.
Last week this article appeared in The New York Times – getting to how a single grade in high school can be an accurate predictor for both success, and likelihood to graduate. Forget grade point averages and SAT scores – one grade in one class is the #1 most accurate predictor with high percentages on large volumes of data.
So sure, there are good and evil ways to wield this new power. But that’s not the only problem. The other problem, one that could be even bigger, is that it can make us lazy. If one grade in high school is a more accurate predictor than grades and SAT scores – are we really going to hide behind the percentages and use them to draw conclusions about whether my own son, George, is going to graduate from college in four years? Whether he is going to succeed in his chosen field of work?
Not so fast Big Data.
If you haven’t seen the movie Sully, with Tom Hanks, you should. It’s one of the best movies I have seen in a while – and I see a lot. This movie is a true story about a pilot who landed a plane on a river in New York City in 2009 and everyone survived. But the computers at the National Transportation Safety Board, NTSB, looked at all of the flight data, and through repeated simulations, found that he had plenty of time to actually land safely in not one but two nearby airports – which meant his employer could fire him without a pension – instead of giving him a ticker-tape parade for saving the lives of all 155 on board. Sully came very close to criminal prosecution (the movie implies) until he asked a couple of questions – questions only a pilot of 40 years would have had the confidence to ask. He asked if the big data and their computer simulations had taken the “human element” into consideration. It turned out the data of the NTSB made a couple of big assumptions about how much time was needed to determine whether or not it was safe to try to make it back to an airport – and once they understood the flaws in their assumptions – it was revealed that the plane would have crashed if the pilot had actually tried to make it back to an airport. The pilot in Sully had 40 years of experience and made the right decision. Let’s be really careful about what assumptions we roll into Big Data and not overlook the human element. After all – my son isn’t going to have 40 years of experience in high school when it’s time for him to apply to college – so he won’t be able to defend himself if the school of his choice makes a wrong assumption about him – just because of one grade.
Another, shorter way to emphasize the importance of the human element is something my father says “it’s not the size of the dog in the fight – it’s the size of the fight in the dog.”
Big Data is incredibly powerful – let’s be smart about how we use it – and let’s not make too many assumptions about what the data means.