Thursday, March 28, 2013

Learning organization

Anyone who knows anything about modeling knows that you cant just be the best at any one modeling system. You have to be in an environment of learning, with a clear vision of not only how to improve your models via feedback from customers, but also agile to know the limits of your modeling system and move on when the problems dont match the techniques. This is hard to do.

The ECMWF has chosen to do one model and do it well, organizing a myriad of people to attack fundamental problems. They have focused resources, a clear vision, and well defined goals. In the US we have many modeling systems, a robust set of testbeds, and as a result a diversified community of modelers. We have a broader learning organization that is by definition less focused due to diversity. We have goals but our problem is the lack of agility despite this diversity. Diversity can be equally paralyzing because there are too many options and we become distracted and governed by our perceived priorities and focus heavily on our limitations. This results in quick approaches that pay off in the short term but not in the long term.

In essence we will ask our best people to always be playing catch up and not give them enough time to be forward thinking. The time spent reading papers, going to conferences, making connections and exchanging ideas is well worth the cost of travel. It provides an outlet, an entry into the creative process through collaboration, and a realization of perspective that you cant get internally (e.g. the purpose of achieving new perspectives is for growth which can quickly stagnate in low morale environments). It is effectively a way to get other people to invest in your problems without using your money. We all benefit when these folks make successful, mutually beneficial collaborations and has been a hallmark in the aerospace industry!

This is precisely about creating a learning organization that has agility and flexibility of choosing good methods without being hampered by custom or culture to implement good ideas. This is all about interaction between researchers and forecasters, to make sure that the best methods, the best techniques, and the best mutually beneficial interactions occur to get the job done. This is an iterative process, and is a hallmark of a learning organization. But learning organizations have well defined, forward looking goals that equate to a solid vision and they have tactical support to implement, refine, and iterate.

Bogging down the system prevents even good intentioned people from creating a vision, crafting a strategy and developing a tactical approach to get the job done. Hiring freezes when organizations are already short staffed, or cutting corners on travel and interactions only serve to amplify the long term impacts. And when those impacts surface they will be unbearable.