Saturday, April 25, 2009

Distrusting Emergent Behavior

Yesterday my TCG partner Peter Hirshberg and I were talking about why it is that people are trustful of hierarchical organizational dynamics and distrustful of emergent organizational behavior. In the past week we had both independently encountered resistance at a client with a proposed course of action which would unleash a potential group dynamic.

I related to Peter the story in Gary Hamel's Future of Management about how the father of "modern" management techniques Frederick Winslow Taylor
in 1912... appeared in front of a congressional committee and argued that scientific management required nothing less than a mental revolution...
Hamel goes on to illustrate how different the world of the latter half of the twentieth century is from the world that Taylor was encountering as he introduced his revolutionary ideas:
Consider: in 1890 the average company in the United States had four employees, and few had more than a couple of hundred workers. Had you been alive at the time, it would have been hard to imagine that a company could ever grow to the scale of U.S. Steel...
Hamel invokes the language of Thomas Kuhn who, in his groundbreaking work on scientific method, developed the concept of "paradigm shifts" in science. Hamel sees similar mechanisms at work in management theory and practice, describing current management practitioners as "partisans of the old paradigm... members of the bureaucratic class."

While I believe this to be true, and believe that it is in part our familiarity with "the old paradigm" that makes it hard for us to accept the new, I had another thought as well this morning while listening to an insightful TED talk by Bruce Bueno de Mesquita called "Three predictions on the future of Iran, and the math to back it up." I have linked the video below and it is definitely worth watching. But the applicable insight is in the first few minutes when he talks about the complexity of human interactions.

The observation is a simple one but with some really interesting implications. Take a problem in which there are 5 decision makers. There are 120 possible interactions between those 5 people (five factorial). Any reasonably smart person can hold 120 connections in their head and can thus feel some comfort in understanding how the dynamics of five people will result in decision making.

But if you simply double the number of people to 10, you increase the number of connections to 3,628,800 (ten factorial). And 100 people? We are talking about really large numbers... There is no way for the human brain to compute the complexity of decision making within these groups, although, Bueno de Mesquita points out, this is a very good task for computers.

The implications for management methodology (and human sociology more generally) are quite interesting -- we develop hierarchical organizations in order to simplify decision making in groups, creating a sufficient efficiency for a given task or activity by limiting the number of inputs into the system or limiting the roles for inputs in order to reduce complexity.

And we grow up in our societies intensely internalizing this sense about the limits of complexity because everything we want to do requires either this notion of small groups or hierarchy which compartmentalizes the type or ability for individuals to contribute to those decisions.

But computers introduced into these systems fundamentally change what is possible in a way which is alien to our internalized understanding of how these systems work. Suddenly with the mediation of computers, significantly larger groups of peers can co-produce information, decisions, insights, etc than we have ever experienced in our non-computer-mediated past.

Understanding how the computer changes who we are as human beings and how its use changes what our societies are capable of will be an essential part of overcoming this reactionary distrust of the emergent properties of these systems.



1 comment:

Zach Tumin said...

I appreciate "Huh? Wow!" stuff as much as the next dude. But looks as if someone's dining out here on over-complexifying and making mystical what history shows time and again as perfectly manageable. Can't keep track of 100 decision makers without a computer? Please. Take a successful manager of Congress -- LBJ, fer instance. He mastered the decision-making flow of 400+ members of Congress, as a Rep., Sen., and Pres. There was no mystery, and no computer. Every master pol does this. Does this analysis not seem too precious by half?