Analytics and Magical Thinking

magic461875119Any disruptive technology brings with it a disruption in thinking. The Industrial Revolution led for a brief while to a belief in almost infinite human mastery over the physical world. The information technology (IT) revolution gave us new awareness of our physical and metaphorical interconnectedness with others around the world. So, given the disruptive nature of data analytics and its associated technologies, what disruptions in thinking can we expect from it?

To avoid overstating my case and blowing my credibility up front, let me clarify: I don’t think that data analytics is going to radically reshape human geography the way that the Industrial Revolution did or fundamentally alter the job market in the way that the IT revolution has. But, it is greatly changing our ability to process and act on information as a species. And at the moment, it’s leading to a resurgence in magical thinking.

By magical thinking, I mean that like any new technology, data analytics is a thoroughly black box to the uninitiated. This opaqueness can breed confusion, not only about how the technology works, but also about what it can do.

The black box phenomenon was a big part of IT’s growing pains in the 70s, 80s, and 90s, and part of the source of the “Business-IT Divide.” You had a brand new technological toolset that could automate, improve efficiency of, and integrate your organization in a way that nothing could before, but the means of doing it was highly unfamiliar to anyone who hadn’t grown up with their hands on a keyboard. Over time, due to increasing maturity among business users and higher technical literacy among those joining the workforce, this gap has reduced. Business users have increasingly gained at least a layperson’s sense of what their technology can and can’t do for them.

Data analytics is at the early, exciting, and mysterious stage right now. Some great results have been delivered, with even greater ones promised. But few outside of the analytic teams have an adequate appreciation of how it works or what’s feasible. Laypeople just know that you feed massive amounts of data in, the “machine” uses some sort of logic to look for patterns, and then it spits out insights that can help better guide the business.

A hotel chain can envision what it wants: clear definitions of customer segments, dashboards showing occupancy rates in comparison to top competitors, updates on the locations getting the best and worst buzz on social media. There are tons of data out there—internal and external, structured and unstructured. Someone just needs to make sense of it. That’s what data analytics is there for, right?

But how does the machine do its magic? That’s less clear. And, without a solid understanding of the relationship between the inputs and the workings of the machine, expectations for unrealistic outputs can grow quickly. A new gap can be the result, and the last thing that most enterprises need at this point is additional gaps in understanding.

What to do about this? There are two basic options:

  1. Use this confusion to your own advantage. Build a data kingdom within your organization, and use jargon to terrify or bore those who come too close. For reasons of ethics and creating a sustainable career path, this option is not recommended.
  2. Build data literacy in your organization. More about this much better approach in a future post.

Related Courses
Updating your Database and Business Intelligence Skills to Microsoft SQL Server 2012 (M40008, M40009)
Cloudera Introduction to Data Science: Building Recommender Systems
Essentials for IBM Cognos BI (V10.2) (B5270G)

In this article

Join the Conversation