Forget 'Big Data.' Beware 'Little Data' -- and the Horrors of TMI

How much data is too much data? I'm pretty sure I knew the answer -- you'll know when your head starts swimming -- back when I studied statistics in high school, but now I'm not so sure anymore. Because now my head always feels like it's swimming in data; that's my brain's default state. Everything is about numbers these days, even when it really shouldn't be. I'm looking at you, BuzzFeed, with your 1.6 Billion Listicles That I Really Shouldn't Waste Time Reading, But Probably Will Anyway. (Be right back. Gotta go read "21 Questions All Bros Must Answer Immediately" and "9 U.S. Presidents If They Were Pokémon.") (OK, I'm back.) And of course the very thing you may well be holding in your hand right now is recording and sharing all manner of data about you, even if you kind of think you've told it not to. (If you're a user of the secret-sharing app Whisper, I hope by now you've read The Guardian's exposé, "Revealed: How Whisper App Tracks 'Anonymous' Users," which neatly spelled the story out in three bullet points: "Some Whisper users monitored even after opting out of geolocation services," "Company shares some information with U.S. Department of Defense" and "User data collated and indefinitely stored in searchable database." Secret is a really hilarious word to use in a digital context, isn't it?) And beyond your phone, much of your other stuff is spitting out data -- and telling on you -- too. (The Internet of Things is only just beginning.) Consider what Dirk Wollschläger, general manager of IBM's global automotive operations, recently told John R. Quain of The New York Times: "Cars are generating a huge amount of data, some billion gigabytes of data every year." In many ways we're still kind of in the gee-whiz honeymoon phase of the Big Data era, so "some billion gigabytes of data" can immediately get spun, not as Too Much Information, but as an awesome thing. Quain paraphrased Wollschläger as explaining that "cars' sensors were recording data that could be useful not only in reporting traffic conditions, but also to locate, for example, potholes. Every time a car's suspension reports a sudden impact, that telemetry could be compared with data from other cars traveling the same route to pinpoint a road repair." That's cool, right? Except for how there's no money left in the budget to fix the damn pothole. And also how most of the data we're collecting, pothole-detecting aside, is actually useless. A lot of media and marketing people have bought into the general idea that consumer/personal data in aggregate is always worth mining and drawing insight from, but let's step back and examine that presumption for a moment, shall we? If you think of all of us humans as ambulatory data-generating machines, leaving trails of 1s and 0s in our wake, well, what do those numbers add up to? Sure, there are telling patterns -- some of them meaningful (humans are creatures of habit, which is a useful thing to be reminded of when, say, considering how to address brand loyalty) -- but just how "relevant" is anything that any average human being does from day to day, hour to hour, minute to minute? I don't mean to get too existentialist on you, but most data about what humans are doing is meaningless and useless because much of what we do as humans (in the grand scheme of things) is meaningless and useless. Or at least entirely routine. Not terribly surprising. Hardly revelatory. (For instance: Wake up. Check phone. Pee. Brush teeth. Check phone. Pour a bowl of cornflakes. Check phone. Get milk from fridge. Check phone again to check time. Check Twitter, click on "23 Secrets Supermarket Employees Won't Tell You," which of course takes you to BuzzFeed. Eat cornflakes….) Just because we're now doing what we do as humans with the assistance of digital devices that track everything doesn't make our mind-numbingly redundant and banal daily routines any more valuable (to marketers or the NSA or anyone!). Data mine all you want, data miners -- for the most part, all you're gonna hit is silt. Unfortunately, the data-mining racket has a certain self-perpetuating logic to it: Everybody wants more data to work with, but the more data you collect, the more complicated (and expensive) it gets to sift through it to look for those elusive golden insights. But making Big Data an even bigger deal won't necessarily, you know, embiggen your spirit (or your market intel or your profit margins). There's a point of diminishing returns for everything; there's a point at which you can break a problem down into such small pieces that you can't even see the trees (or the branches or the leaves), let alone the forest. To put that another way, most Big Data is about lots of Little Data -- and most Little Data is worthless. Just a little something to keep in mind as you go about your business, dear reader. Simon Dumenco is the "Media Guy" columnist for Advertising Age. You can follow him on Twitter @simondumenco.

Forget 'Big Data.' Beware 'Little Data' -- and the Horrors of TMI

Featured Stories