A lot of companies make a great deal of noise clamoring about how they’re data driven or data centric or data [insert buzzword here]. But what does any of that mean? Are they really? Data can be infinitely complex, and could easily be discussed ad nauseum, which makes those claims both fascinating and vomit-inducing.
It’s highly likely that it is a marketing ploy, but most people never bother to look under the hood. I would bet that most companies barely understand what it is that they do, and somehow their business models run on magic fairy dust and unicorn power. (That guy is the chair of the US Finance Committee, FYI).
Data driven? I think for most companies, it’s all smoke and mirrors.
But, for those of us who are interested—or, since you’re reading this, bored—the sweet, sweet science of data opens a world of revenue opportunities or can land you a decent job. According to Glassdoor (who, ironically, is terrible with data,) the average data scientist makes about $120K a year, and since most people are so data-deficient, you might be able to talk a big game and not actually know anything. But first, let’s simplify stuff to the point of hyperbole and take the infinitely complex and reduce it down to three categories.
Was this Your Card?
When most people (and companies) talk about data, what they really mean is descriptive data – it’s a picture of some event in the past. Make a sale? Cool, log that in the double ledger so we can come back later and reference it for some obscure reason. This is where most people are comfortable and where most businesses play. Even the venerable data-houses like Google and Facebook spend a lot of time here, loading your browser up with cookies and tracking your movement.
Descriptive data is the foundation for anything data science-y, and it turns out there’s oodles and oodles of it. In fact, if you happen to be working for a questionable news source, you can even make it up—bad data is still data—and most people won’t bother to assess the validity, so it’s an effective business model.
But descriptive data not really that novel, and it certainly doesn’t qualify a company as being data driven—if the only function of driving was knowing where you’ve been, your chances of survival would quickly approach zero.
Getting a Good Reading
For the smaller subset of companies that are not spewing utter bullshit, there’s the second category: interpretive data (I’m sure there’s a better descriptor, but I’ve got a syllabic schema to keep up). I’m going to lump together a few concepts together, because in my warped mind, they’re different sides of the same coin—real-time, ratios, and to a small extent, probabilities. This is all data that can be used to figure out what’s happening right now.
As an individual, it seems readily apparent: you’re fighting to stay awake. But for companies, it’s incredibly complex and the clear majority of them do not do it very well. Think I’m crazy? It’s possible, but how many billions are spent on real-time analytics? On advertising and marketing tech? On consulting services? On almost anything finance related? I can promise you, it’s a lot.
And why would a firm supposedly waste such huge sums of cash? They might know where they’ve been, but they don’t know where they are, and that knowledge is critical for survival. Perversely, startups have it easy here—they’re small and relatively uncomplex. Globe-spanning mega-corporations have it a bit harder, and sometimes things fall through the cracks. Oops.
Despite my usual FB bashing, most executives would punch a baby in the face to have the Zuck’s data capabilities; despite being a zillion dollar behemoth, Facebook can pivot and react with surprising speed and agility. I bet the other CEOs look at him with a mixture of jealousy and contempt when he shows up at Bilderberg wearing a hoodie.
As individuals, we’re moderately good at predicting things a few seconds in the future—we generally don’t slam head-long into each other while on the sidewalk, even without previous data. But it takes a large amount of mental energy, and I’m not sure I’d trust myself if we were barreling at each other at a high rate of speed. Predictive models, on the other hand, generally don’t have such computational issues, but they do require vast amounts of valid data. This is what makes AI—and it’s slightly less capable sibling, machine learning—such an intriguing and dangerous opportunity.
As our time horizon moves further into the future, things get a little bit weird, because we’re not talking about data in the classical sense anymore. But, if your models are perfect, and your data is immaculate, and your horoscope says the stars are aligned, it might be possible to generate accurate predictions. Financiers have been chasing that dragon for ages, but it turns out a less-than-reputable company did it well: Cambridge Analytica.
Is your phone actively listening to you, so it can predict your next search or serve up better ads? Probably not, it’s technically complicated, and the payoff is only a few seconds in the future. Do data centers connect some disparate dots and make an educated guess about something a few weeks out? They’ve been doing it for years. Can you outfox your competitors by running your business in Tomorrowland? Potentially. And if you could accurately predict sales next fiscal year, you could streamline the supply chain, leverage to the hilt, and fire most of the finance department. Win-win-win.
Predictive data can be really cool, but it also can be really scary. Like Cathy O’Neil points out in her book, the implications are far reaching. And while a company could be so talented with data that you might have gotten duped into voting for a reality TV star, there’s also companies out there that are trying to predict when you’re going to have a heart attack and treat you preemptively. It’s magic that can be used for good or for evil.
I’m not sure we’ll ever get to the point where we’re metaphorically arbitraging life anytime soon—I’ve trivialized some extreme complexities for the sake of discussion—though it certainly warrants consideration. With the potential rewards in the hundreds of billions if not trillions, it is easy to understand why so many firms are touting their ‘data driven’ mantra: they want a piece of the action though they’re stuck in the Dark Ages.
Fun to ponder the magic and wonder, though.