November 9th, the results from the 2016 US Presidential Election were in. Trump won. Communism won. Racism won. Sexism won. Hate won. That was the narrative we were force fed by the mainstream media, Facebook, and Twitter. In an age of getting pretty much any information we want with a quick google, how can such unsubstantiated claims even be made? Then clicking through my phone I saw a tweet about Big Data failing America. Say what?
Big Data is all about collecting massive amounts of concise data to support analysis and planning. Big Data also helps us test key assumptions made during that analysis and planning. Big Data is beyond traditional analytics, often acting as a foundation for artificial intelligence.
To gain context on why Big Data failed America, I dug through all the news sources. There sure was a lot of talk about Big Data. Sure, they used the layman terms, terms like demographics and polls, but clearly they were talking about Big Data. And in each of the articles and TV discussions, media was blaming Big Data for Clinton losing.
Let’s take a step back. Clinton did not lose because of Big Data. More accurately, one of the factors for her losing was how her campaign used Big Data. Putting it bluntly, Clinton’s team and her alliances ignored standard practices when using Big Data. The following is just a few of those practices.
Data driven. Clinton’s narrative was not driven by facts, nor data, but on emotional rhetoric. To be fair, we need rhetoric to band people together and mobilize them to act on data, to act on facts. Instead voters received highly intangible information, variants of Clinton being a woman, an experienced politician, and that she wanted to be president. For Big Data to work there must be data-driven rhetoric. Clinton’s campaign failed to do this.
Tested assumptions. Clinton’s campaign operated under the assumption that Clinton was going to be president. This is good for the rhetoric she needed to build confidence in the voters. But, this assumption prevented her team to do their due diligence for planning and analysis. Did anyone on Clinton’s team ask what is the likelihood of their opponents winning a electoral vote or even winning the Whitehouse? Big Data is often used to answer these sorts of questions. Also, Big Data helps us test and challenge the assumptions made during planning and analysis. Clinton’s campaign did not test, did not challenged the assumption that Clinton will be president.
Unified model. Clinton’s campaign had access to a lot of deep data; data that segregated voters according to the traditional demographics. Rather than using this data to build a model, a 360 view of what the American voter looks like today, the data for each demographic was analyzed only on its own. Big Data identifies key patterns and forecasts key behaviors when everything is integrated into a single model. Clinton’s campaign had no such model. This resulted in her campaign being driven by demographics and not by the American voters.
Trump won. Does this mean Big Data failed America? No. Clinton’s campaign failed to use Big Data in the standard way. We can’t blame Big Data. So, what should be blamed? We blame rhetoric. We blame rhetoric that’s driven without facts, without data. To be clear, I’m not opposed to rhetoric. We need rhetoric to rally us up together and do something. We need rhetoric so people act on Big Data. But, to use rhetoric alone, to assume you have all the answers, to assume you know more than anyone else, to assume you know more than the information that any data source can provide…
My friend, what you’re doing just doesn’t work anymore.
Great article Chris, another thing that I noticed is that the Clinton team was watching the comments of her followers and then using their words and phrases to cater to them. The name calling that showed up later in her speeches showed up in the comments by her followers quite a bit of time before she used the same phrases.
Chris, I appreciate your view of Big Data and how in the example of the election it was not used for the results desired. It seems with today’s social media, our emotional intelligence is being tested. With the ability to sway people one way or another, it is a tricky endeavor; especially when emotionally charged. Information is misunderstood, twisted and turned, and miscalculated.
The key is discernment, what is truth, half-truth or false truth. It is not about convincing someone to follow you, rather it is through positive persuasion. A mass consciousness is at work in the background, creating a powerful energy for either cooperation or resistance.
I agree with the point you make about Big Data identifies key patterns and forecasts key behaviors when everything is integrated into a single model. By understanding this model the masses can be guided in the direction desired. Otherwise, the process will implode on itself and cause destruction, as we witnessed in the election.
Great article!
The problem with BIG DATA is that it also might have a biased sample. Many fields of study are interested in crowdsourcing. One of such BIG DATA crowdsourcing project is e-Birds (http://ebird.org/content/ebird/). If you explore the maps, and look at the data, you will be able to notice that the maps are showing where people are walking, not where birds are flying…
Though we can learn from our mistakes, we tend to learn more from what we did right, not what we did wrong.
Thank You Chris Pehura for putting the dilemma into such beautiful words!
While you are absolutely right about ‘what we’re doing just des not work any more’, we can always learn what works. The only difference lies in the fact whether we learn it before the catastrophe or after.
Too bad, in this particular case, the fact has become evident when it is already too late.
Better luck next time!
Too often the tool, not the tool user is blamed. A fool with a tool is still a fool.