When data gets dumber

In a world where code is king, and data is queen, data has become an incredibly powerful commodity. With each passing day, it's traded, weaponized, lifted, shared, stolen, regulated, distorted, abused and quite often misunderstood. But before we get to that, something has been gnawing at me for quite some time; the idea that data has become more important than insight–where the two are mutually exclusive. It's become the chicken and egg thing.

My formal design education came at the dawn of the digital age, where we not only learned about the formal and emotional craft of design, but also the intuitive, insightful aspect of observation and teasing out meaning. Analog research is elegant, because a human is doing the crunching; constantly tuning and refining the approach in an organic way. I was taught that before we even touched an interface, we had to know why we were beginning to make something, before determining what it is we were setting out to make. We had to learn how to learn, how to ask questions that tease out meaning, and then how to put it all together.

Over the past 20 years, the science and technology of data capture and analysis has progressed at a staggering pace. But here's the rub. I don't believe the seams that drive and surround that technology have kept pace with that exponential advancement. That's on people. Us. I can't tell you how many 100+ page decks I've slogged through that present mountains of data that reveal little more than a painfully detailed measurement of a poorly framed question or assumption. Hence the moniker "dumb data," where the lateral edges of a problem are overlooked or even ignored.

It's easy to forget that data is nothing more than a very accurate measuring tape. It can tell you how long, short and wide something is, but what it can't tell you is its size in relationship to where it sits in a room, or why people want to measure stuff in the first place. It can only do what you ask it to do. Given that data has become a commodity, it's fair to say that it's sold by the pound as a potent validator for decision making, good or bad. Just like averages, data can be easily distorted and manipulated by extremes. But to be clear, data as a measurement of specific input factors is infallible; people aren't.

Conversely, the opposite of dumb data is "elegant data," where human insight and curiosity is a critical driver of input. Humans have the capacity to consider behaviors, context, analogous factors, and seemingly discordant components of any problem. And many times the differences between human and technological analysis are subtle, but immensely important. I am arguing that we can do a much better job of using the power of computing to serve better, more meaningful results. Data without insightful input is just data; revealing nothing more than a presumptuous measurement doesn't give us much more than a dumb answer to a dumb question.

I fully admit I'm making sweeping generalizations here, and I make no claims to sporting the Data Analytics Club T-shirt. I'm just saying that we can do better by looking deeper into how we use technology. We can ask why a lot earlier, and perhaps use more elegant data to better inform meaningful hypotheses, not dumb data to simply justify or validate bad ones. One can only hope.