Feeds:
Posts
Comments

Posts Tagged ‘modeling’

old-man

In the new normal, data has become the Holy Grail for making business decisions.  Everywhere you look the importance of data comes to the surface.  Marketing programs, websites, hiring practices, social media, finance, sourcing, and manufacturing, to name just a few, all rely heavily on data as a feedstock for decision-making.  Throw this data in to a computer and ‘presto,’ you have your answer.

As the world moves to the age of data scientists, data engineers, data analysts and data architects, I reflect on something I experienced many years ago that I believe remains true today.

I took a course at Massachusetts Institute of Technology in computer simulation of business strategies.  Part of the course was a three-day forecasting competition using the data of a real business.  We set up four groups.  Each group developed their own business strategies and entered the data into the computer system each night.  The next day we reviewed the results and decided on new input for the next run.  It was a financial simulation model that involved using random number generators and distribution functions that mirrored historic company data and performance.  At that time, this type of simulation modeling was very sophisticated.  It was like the data analytics of today.

One group who corralled at the back end of the class room, included an older gentleman (old man).  He didn’t say much but when he did, his questions and comments were measured and well thought-out.  On the final day, when we were comparing our results, he spoke up.  Out of the back of the room came, “It won’t work.”  What?  “Your models won’t work.  I just cut off your raw material supply.”  Silence.  Game over.  Using his 40 plus years of business experience, intuition, and knowledge of the industry, he made a human decision that trumped our models.  It didn’t matter what the data was saying.  I never forgot this experience.

I have built many financial simulation models in my business career and realize that in every fancy algorithm, there is an old man.  I believe he also exists in today’s data analytics.  –  Beware of the old man.

Read Full Post »

%d bloggers like this: