Nate Silver, Life, The Universe and Enterprise Social Networks

By Richard Hughes on November 09, 2012

On Wednesday morning, a clear winner emerged from the US election. No, not him – I mean Nate Silver, the New York Times blogger and statistical wizard. In an election universally described as “too close to call”, Silver correctly predicted the result in every single state. Writing in The Guardian, Bob O’Hara asks:

But how did Mr Silver predict the presidential race so accurately? What was this dark magic that he used?

The answer, of course, is that darkest of all dark magics, “mathematics”. Prior to polling day, another article in The Guardian had suggested:

The question of who will emerge victorious depends on whether you ask the priests or the mathematicians

Whatever your religious persuasions, there was really only ever going to be one winner there.

In an unashamedly blatant piece of newsjacking, I can’t help but see parallels between this and the way organisations measure the success of their internal-facing social business initiatives. Social networks, like opinion polls, generate a massive amount of data, yet people are all too happy to ignore that data and rely on their instinct. They use subjective measures such as “does it feel like we are engaging with employees better?”,  rather than analysing the data they have and answering the question objectively. Why is this?

I believe there are two main reasons. Firstly, fear of getting an answer they don’t like – a kind of social business “don’t ask, don’t tell”. This is a natural instinct, especially when some organisations are still struggling to make a business case for social networking inside their company. Understandable, but really not justifiable.

Dan Lyons, writing on ReadWrite about Nate Silver, said:

This is about the triumph of machines and software over gut instinct. 

This is only partially true. Yes, “gut instinct” was the clear loser, but the real triumph was the way Silver described the problem to the machines, and represented his statistical model in software. The machine was simply doing what it was told. (Incidentally, this is why I refuse to use the term “smartphone” – I was always taught that computers are fundamentally stupid, but are very good at doing precisely what you tell them. Today’s phones are not “smart” at all – they are just more powerful, and better-programmed).

Many years ago, a customer said to me “I have a 6 million line activity log from our eCommerce site, what shall I do with it?”. I said, “what do you want to know?”. “I don’t know”, he replied. “Well delete it then” was my, rather flippant, advice. If you can’t describe the problem effectively, then all the Big Data in the world is not going to help you. (Big Data has always struck me as “running before you can walk” because so few people can handle Little Data effectively).

Of course, all this was presciently satirised by Douglas Adams in The Hitchhiker’s Guide To The Galaxy, with the colossal  processing power of Deep Thought reaching an answer that merely highlighted how poorly the question was phrased. (Deep Thought was, however, clearly a much better consultant than I ever was, as it successfully sold a massive follow-on consulting engagement to define the question).

Getting back to social business, the point is that many projects start without really understanding which questions they should be asking, which measurements were most important. It is tempting to reel off a list of important things to measure (like this one I wrote), but the truth is that the most important metrics differ from project to project, depending on what the objectives are. The questions we ask need to be described scientifically and aligned to business goals in order to separate (to borrow the title of Nate Silver’s book) the signal from the noise. And that’s not a job for Big Data-chomping machines; that’s a job for humans. Hopefully Nate Silver’s success will inspire a few more humans to embrace this “dark magic” of mathematics.