REPLY Getting a Good Problem Statement (SD6613)

[SDMAIL Kim Warren]

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

Seems to me that SD offers a particular kind of solution to a particular kind 
of problem, but one that is near-universal - something is changing over time 
in a way we would prefer to be different from what is actually occurring. This 
may seem a trivial distinction to our community, but problems seem more often 
to be stated as "the current *situation* is not what we want, and we want its 
*state* to be different", [e.g. return on investment is only 5% and we need it 
to be 8%]. We in SD would worry that this risks creating a better-before-worse 
outcome [e.g. hitting 8% next quarter, but damaging future profitability].

If so, the reference mode is absolutely critical to SD's contribution. If the 
owner of the problem [or opportunity] cannot with our help state the problem 
in this form, then I don't see how an SD process is going to make meaningful 
progress. I also don't see how it is possible to trace out the likely causality 
in the situation without linking it back to this performance-over-time of the 
reference factor, e.g. X is falling at this rate, because Y is rising at that 
rate, in spite of Z changing like this, which is happening because W is ....... 
Otherwise, discussion of causality disconnected from the reference mode risks 
being little more than idle speculation.

The only complication I can see is that there may be more than one element to 
the dynamic specification of the problem, e.g. "sales are falling *and* we are 
spending more on marketing". But since these multiple elements are part of the 
same overall system, they should come together in any decent SD architecture 
depicting that situation. 
 
None of this need assume that people have already decided to use SD, merely 
that the performance-over-time nature of the problem means that only SD  will 
be suited to addressing it. We do not need to be defensive against the 
accusation of hitting every problem 'nail' with the SD 'hammer' [assuming we 
know when to defer to another appropriate time-based tool, such as discrete-
event or agent-based]. The really dangerous hammer that is used to hit every 
problem as if it is a nail is the spreadsheet!

Kim Warren
Posted by  "Kim Warren" <Kim at strategydynamics.com>
posting date  Tue, 11 Sep 2007 19:12:15 +0100