REPLY Why don't organizations function better? (SD6588)

[SDMAIL Schuette, Wade]

Posted by  "Schuette, Wade" <wschuett at jhsph.edu>

Bill, the SPE pyramid doesn't explain where the Structure came
from in the first place or why it persists. 
 
Probably earlier structure generates processes that 
generates more elaborate later structure.  To paraphrase John Gall,
complex systems that work have invariably evolved from simple
systems that worked, and can't be designed directly. (Air Traffic
Control system designers take note.)
 
Still, the fascinating difference between a social system and 
a computer-feedback-controlled aircraft is that none of the parts 
of social systems are guaranteed to stay where you put them.    
So every question of stabilty of what a SD system does has 
a corresponding question of stability of what the system IS.
 
Still that might be a feature, not a bug.  Maybe, as in theory
Y of management,  the parts know more than the designer and 
we should let them move slowly to the positions they want to be in.
Is the right question how to design an organization, or how to stop
getting in the way of an organization reinventing itself continually
successfully?
 
Near the Theory X point (top down control),  variations that
way are highly unstable.  near the theory X point (bottom up
emergent design) variations that way are the whole point.
 
So, that would imply, I think, that the way to get a large complex
organization that works is to start with a smaller one, down to 
one person, who has the two basic cybernetic loops already 
working -- they are in touch with themself and in touch with
the customer's outcomes.   Then, add people to increase scale
and depth, but at each step attend to restoring the loop invariant 
and getting those core loops re-established and always at the 
most stable point.   That's the design formula. 
 
The theory I see as to why that works better than doing some
large engineering design is that life is fractally complicated, and 
people aren't very good perceivers to begin with, so there are 
very low odds you can design, from scratch, a system that will 
take into account everything it needs to.   But, if the system 
evolves and remains near the agile and responsive point,  those
things will show up and be adapted to as you go. 
 
One way to break that evolution is if the people "at the top" begin to 
think they can see everything and know everything, and start
over-riding everyone else and doing planning in secret.  Another is
if the system attempts to grow faster than it can digest change and
so it cannot re-establish the loop invariants.  (Which would imply
that venture capitalists who demand 40% growth per year are 
largely shooting themselves in both feet.)
 
Then both loops break and successful adaption to reality stops. 
 
It's a simple model that suggests a few good things to look for.
If those loops aren't in place, survival would seem to be unlikely.
They may not be sufficient, but they strike me as necessary, 
and something that can be assessed before getting lost in detail. 
 
Wade
Posted by  "Schuette, Wade" <wschuett at jhsph.edu>
posting date  Thu, 6 Sep 2007 15:46:40 -0400