REPLY Models, Problems and Systems (SD7021)

[SDMAIL Kim Warren]

Posted by  "Kim Warren" <Kim at strategydynamics.com>

Thanks for all the responses on this. Maybe my thinking is too
simplistic, but I thought the implied distinction concerned [a] trying
to solve a specific, if complicated challenge at a point in time [though
it may recur] vs. [b] trying to guide performance of a system through a
potentially wide variety of circumstances in which an equally wide
variety of specific problems may arise. 

I have heard it said that SD models should be treated like Kleenex -
used once and thrown away, because the next problem will always be
different. Yet we have models like People Express, and many others,
where multiple decisions need taking, or policies developed, to deal
with all kinds of possible circumstances. Surely sustaining profitable
growth for a large, successful airline is a quite different 'problem'
from rescuing a small, declining, and about-to-be-bankrupt one? Yet the
fundamental structure of the system is the same, and the model of that
system is certainly useful. 

What I guess many of us try to do is leave people with a model - whether
a fully whirring and churning simulation or just a robust, quantified
diagrammatic picture - that continues to help them avoid problems and
sustain performance for long continuous periods, and also helps them
respond when specific problems turn up. I would certainly concede that
such a general-purpose model may need updating, developing or extending,
but at its heart has a continuing life. I would also concede [and as
others suggest, this may be where the guidance originated] the danger in
trying to model the entire system in all its detail.

If we define 'good management of the system' to be 'the problem' then
all we are left with is semantics and as other posts have noted, surely
the guiding rule is meaningless?

Kim Warren 
Posted by  "Kim Warren" <Kim at strategydynamics.com>
posting date  Fri, 9 May 2008 14:23:15 +0100