Good question. I think I have an answer, and it stems from the key observation that it’s pretty easy to do mediocre analytics. Even worse, it’s scarily easy to “do analytics” without any meaningful value.
The barrier to entry is small, and the excitement around analytics in most enterprises makes it very tempting to “just get going” without thinking through some fundamental questions and understanding and planning for barriers you’re likely to encounter.
Ideally, you’d do this before investing significant time and money – but we’ve seen many examples where this wasn’t the case. You may very well have seen this too.
You may be in a company without an analytics capability, wondering if your data and/or your business supports investing in one. You may have made that investment, but months or years later are wondering why your analytics program hasn’t produced meaningful value you can proudly point to. Or (almost worst), your analytics team may actually have produced valuable insights, but your company doesn’t understand how to use the insights you’ve worked hard to uncover.
In my previous post, I said “I’m as starry-eyed as the next person about the incredible progress and possibilities” (of Machine Learning and AI), and that’s still very much true. But, there’s also reason to be pragmatic about the promise of analytics. There are several pitfalls you’re likely to encounter when attempting to get significant, ongoing value from data.
There are technological and process barriers.
You need a data model, a data governance program, hardware, and software in place that is up to the task of meaningfully interpreting an extremely high volume of data.
There are cultural barriers.
To effectively use data, you must have a data-driven culture. Or rather, “data informed”. Kirk Borne has said that enterprises should be “data-informed, technology-powered, and outcome-driven” (not data-driven), and I agree with that distinction. Data itself doesn’t drive an enterprise; data and technology (and people and processes) should all be in the service of producing the right outcomes.
There are capability and knowledge barriers.
You don’t have the right people doing analytics, and/or they’re without appropriate “guard rails” to keep them moving along a path that can produce tangible value. More often than not, companies don’t know as clearly as they should what outcomes they want that data can inform, and thus don’t know the critical questions they need to ask and are easily lost in a pile of data.
There are cognitive barriers.
Leaders with decades of experience often function on an intuitive level, and many have flourished. This is fine, of course, but we are increasingly living in an age in which gut feelings can actively be substantiated (or disproved) with data. Scarily enough, many people who are used to operating without being “data informed” do not recognize this, and if they do, they simply believe that the data is wrong. (By the way, for the semantic purists, I know that the word “data” is plural; it’s just that most people feel the phrase “data are” sounds strange. I’ll just keep using it as if it were singular.)
Avoiding “mediocre analytics” by balancing excitement about the possibilities with a pragmatic approach isn’t easy to get right, but it can certainly be done, and that’s what keeps me going and made me want to write up some thoughts in this blog. Trexin has decades of collective experience thinking about what went right and what went wrong with dozens of analytics projects. We can help you avoid common mistakes and help you understand the best way through the maze of options.
My next post is called “Six Reasons You’re Not Ready for Analytics, Part 1”. The title is deliberately a bit provocative, but I’ll cover some common barriers to meaningful analytics in more detail and ways to get past them. In the meantime, thanks for reading, and please let me know what you think in the comments.