Maybe I've been reading too much science fiction lately, but the next couple of decades will be dominated by AR/VR like devices and experiences. There is no doubt in my mind. Certainly, the highway will be littered with road kill to get there, but get there we will. Machine learning is integral to this future and an area that I'm keenly interested in.
I was originally going to tee this up with an example of magic metrics and data analysis to determine what your magic metric is and how to use data analysis (and machine learning) to identify (and address) customer churn. A post for later.
But then I read the Wall Street Journal this morning about the exit of Lending Club's founder and CEO, Renaud Laplanche.
In a world where every executive team says that they are data driven, then you have to ask yourself a question. If that is true, then can you automate the tactical leadership decisions through machine learning?
The CEO is certainly critical for many things, like vision, fundraising, culture (including cheerleading) and team building, but is the value overblown? Was Renaud really worth total annual compensation that exceeded $11 million? Or could he have been replaced by a machine learning algorithm? Think about it.
Agile has merits in some parts of product development. It has some serious flaws as well IMO. Two things that strike me as true with Agile are:
- Inspect and Adapt attitude
- Executive team members are not product owners (this is a strict and often violated tenet)
We're disrupting all sorts of things with machine learning (search, computer vision, robotics, Wealthfront, Betterment, HealthKit, CareKit, ...). If the executive team members aren't product owners, then can what they do be both improved and cost reduced? Is it time to disrupt the basic organization of the company?
An internal review of Lending Club found multiple instances of loans being made to an investor(s) in violation of company policies. The lack of full disclosure was unacceptable to the board. Renaud was fired. This example surely doesn't need machine learning. But, machine learning could be critical to more complex situations. What if you could instantly and correctly analyze an infinite number of variables as they pertain to the operation of your business? Automatically? Without leadership intervention?
Imagine the interactions that you have today with company leadership. How much of it might be a candidate for using machine learning? Will this reverse the widening pay gap between the top and the bottom?
In any case, expect more thoughts about machine learning here over the next few months.
Also, I can't say enough about how entertaining Season Three of HBO's Silicon Valley is! Action Jack could certainly be improved with machine learning. But thank you for his invention of the Conjoined Triangles of Success!