For most people (anywhere), getting a good handle on descriptive analytics is the dimension of data that is going to be the most useful the most readily available in their day jobs.
What descriptive analytics means: having access to data that describes the world and helps you zero on actual, demonstrated behaviours at scale.
It’s not necessarily the most showy use of data ever — that would be prescriptive analytics, when analytical data is churned to suggest what may happen in the future and a set course of action is proposed to get there. But you gotta learn to walk before you can run, and any business, or person, who is starting on a journey with data will necessarily have a first step that will involve understanding more about how their business works or what their users do with their products. And there’s tons of value in this.
There are several paths you could take, of course, to augmenting your understanding of analytics. But today, I want to suggest a path that is available to almost anyone because it doesn’t rely on the particular circumstances of your employment: deciding to take a self-guided course in analytics.
Now you may say — well, it is perhaps a bit extreme to make this your first suggestion because this seems like a pretty significant commitment. But I think when folks who may call themselves data curious (many of you who receive this newsletter in fact!) are at a stage where they want to clarify and significantly grow their knowledge, it’s hard to do this without a bit of more formal learning. Because more casual learning (a blog post here or there) may just never give you the specific missing blocks that would be transformational for you.
I’ve come to this conclusion after a few months leading the Smart Data Initiative for INMA because there has been a recurring feature in many of the conversations I’ve had with folks with a demonstrated interest in our topic here but who come from non-data backgrounds.
At one point in our conversation, we’re talking about some data subject or another, and this person will ask a question that actually rests on a fundamental of analytics or data engineering — like, “But why doesn’t X happen” or “But how do we know this.” And I’ve realised these are precisely the things that blog posts, no matter how insightful, don’t usually cover.
That’s because while these aren’t necessarily super complex things, they don’t neatly fit into 1,000 words — a standard blog post. And because identifying these missing blocks is going to be pretty hit-or-miss.
Think of it this way: All our non-formal knowledge (about anything) is a kind of Swiss cheese. Even for two people with the same amount of knowledge they have organically picked up here and there — from conferences, to Webinars, or reading relevant blog posts — the pieces of knowledge each person is missing are not going to be the same.
So, as someone who does ask myself — every other week in fact — what could be an interesting topic for a news media person who works with data, one thing that’s become more and more obvious to me is that there aren’t a few obvious holes in the Swiss cheese where everyone is needing the same piece. For the most part, it’s a case of “We don’t know what we don’t know” — very hard to guide someone to their missing piece.
I also remembered something that James Robinson, who is a director for data products at The New York Times, said to me about trying to bring data closer to non-data folks at The New York Times.
James was detailing how the NYT’s vast collection of analytics tools was creating a new problem for users — figuring out what tool was useful to learn about what problem. So, James told me, they wrote a guide to explain the various tools and their use. Except this didn’t prove to solve the issue. Instead, James said, “The problem is not that we don’t have a good map of the jungle. It’s that we need the jungle to be a well-attended garden.”
Self-training is learning to orient ourselves through the jungle in the first place — whether we’re given a guide or not — and data and analytics aren’t particularly easy jungles (to James’ point). So while we can collectively hope the data product managers and tool builders of the world may work at making the jungles more like gardens, until that happens, we have to figure out how to better walk through the jungle on our own.
If you’d like to subscribe to my bi-weekly newsletter, INMA members can do so here.