Last weekend, I watched an engaging Netflix original Bollywood suspense thriller. One of the characters in the movie is a math teacher who is frustrated with how his students perceive the questions in his exam papers. Rather than testing if a student can solve a geometry problem, he embeds algebra in the problems. Overall, the questions are easy, but the thought process to solve them is not. Hence, his students keep complaining about the difficulty level of his problems.

He calls his approach “MathLogic”. He insists that the traditional approach of solving math problems will only lead to a workforce who are only good at solving what they are told, what they need to solve for, and what to expect as a result.

When we see a geometrical figure in an exam question, our mind immediately constrains us. We then dive into the problem with the “geometry” approach and mindset. This means we would struggle if we had a teacher like the one in that Netflix original movie.

The education system circumvents this problem by framing questions that do not transcend boundaries. Even though a geometry question is difficult, in terms of the number of angels involved (a triangle in a circle), our brain will still find it easier to solve. Because we have practiced similar problems, we do not have to force our minds to span multiple sub-areas in mathematics (like algebra, trigonometry, etc.). This means that we end up carrying this constrained thinking in the post-education world as well, into corporate.

The education system can circumvent this problem. The corporate world can not. The corporate world has been partly oblivious to the very existence of this problem for a long time. Siloed thinking in analytics professionals, which I like to term “Analytics Fragmentation”, is a much more significant challenge in the corporate world.

In the last five years, you may have seen a spike in discussion around data fragmentation and its challenges. One of the reasons is that as companies embark on building a data architecture (lake, mesh, fabric, and all those buzzwords), they realize that the current state is broken. The same applies to analytics fragmentation.

Analytics fragmentation is relatively more challenging to visualize. You can build an architecture to showcase how data is fragmented. To showcase analytics fragmentation will take much more effort. Analytics fragmentation is not about analytics being done in silos. It is about whether those analytics pockets incorporate enterprise-wide impact and thought process in those analytics initiatives.

A centralized enterprise analytics capability that can serve the entire organization is a bit far on the horizon for large organizations. A significant amount of work needs to be done on the data side first, to make that vision a reality. But even if you had such a capability, the gist of the tactic to eliminate analytics fragmentation is the thought process vs. IT infrastructure.

You must cross-train your analytics pros (and, in the long-term, your citizen analysts and data scientists) to develop the MathLogic thinking.

As analytics professionals, we are so excited with the numbers and our quest to solve a problem that the aspect of framing it and pondering over that framework becomes secondary. Every analytics problem is like the MathLogic problem mentioned in the beginning. Customer foot traffic in the store going down? Finished goods area in a manufacturing plant overflowing? Raw material costs for a specific product line have gone up? Solving these correctly and permanently requires you to think beyond the math. You must embrace the logic that helps connect the drivers to build the complete picture. The math will work only when we decipher and start problem-solving with the right logic, as highlighted by the teacher in the movie.

While analytics fragmentation can still deliver value for now, every such analysis leads to some missed value that could have been realized if the approach had been more holistic. We will discuss this in more detail in this week’s episode of “Think About It.”