REPLY Getting a Good Problem Statement (SD6601)

[SDMAIL Jean-Jacques Laublé]

Posted by  Jean-Jacques Laublé <jean-jacques.lauble at wanadoo.fr>

Jim

You used solutions meaning I suppose that there are many level of
definitions.
For me the first level of definition should not include anything about the
solution because the problem exists independently of the solution and that there are
many solutions to the same problem depending on the time resources that
exist and that you will decide to use.

I do not pretend that one cannot step back having studied a first level of
definition and began to apply the solution. This is why there is a sort of
chicken egg process here.

The first level of definition includes the critical questions that make a
project fail or succeed :the level of the nuisance of the problem and if it is continuous or
discreet.
By continuous I mean that the nuisance can be reduced more or less and
discreet if there are only two solutions (on can imagine more than two
solutions) for example: the problem is solved or not solved.

The range of time left to solve the nuisance and the minimum reduction of
the nuisance needed.

The level of risk the client accepts to solve the problem.

Example: if the client is a multinational company and the problem is in one
of its branch of small importance, he may accept a high expectation
characterized by a very high profit and a smaller probability of success,
while a company owned by a family who earns its life from it, will prefer a
smaller expectation characterized by a lower profit and a higher probability
of success and low probability of failure and overall cost.

The cost of the overall operation must be considered too, especially if
there is a failure.

These are fundamental questions that have nothing to do with the solution.

These points will automatically have an influence on for instance the
boundaries of the model, the horizon, the quantity and precision of the date
needed  etc.

But these questions must come first.

To my opinion if these questions are well addressed, 80% of the problem is
solved.

But I think that too properly define the first level of definition, should
take at least half of the total effort in time and effort allocated to the
resolution of the problem and it is because it is not done this way that SD
is not well appreciated.

I could give you examples as you asked, but to my opinion the best thing
would be to work on concrete subjects like the one I exposed 'modeling the SD growth' although I
think that it would be difficult because of the lack of data.

There should exist a public place like this forum where organisations,
business (of course after a selection to be defined) could expose a problem
that would be openly solved and criticized (or not solved) so as to make visible the
concrete process of SD modelling.

The people involved in the process and how other people could ask questions
or even have an influence on the process should be well defined.
This would give another credibility than text books examples.
I do not say it is easy to do, but one can start with something less
ambitious.
Regards.
Jean-Jacques Laublé
Posted by  Jean-Jacques Laublé <jean-jacques.lauble at wanadoo.fr>
posting date  Sun, 9 Sep 2007 13:06:42 +0200