Posted by Bill Harris <
bill_harris@facilitatedsystems.com>
> ""SDMAIL Jean-Jacques Laublé"" <
jean-jacques.lauble@wanadoo.fr> writes:
>> The only explanation I can find about this fear of naming SD, is that
>> there must be, or must have been a lot of studies pretending to use
>> SD that failed to deliver the expected results.
Hi Jean-Jacques,
I think there may be another explanation. Say you as a businessperson meet two consultants. One says, ""Hello, my name is Jane, and I do system dynamics,"" while the other says, ""Hello, my name is Mary, and I help you solve your problems."" There's an impression among some (many?) consultants that Jane cares about himself and his technology, while Mary cares about the client, and, as a result, people will be more likely to hire Mary.
I really don't know which is true. You seem to be making the claim that you'd often rather hire Jane, for at least you can identify quickly what she does. Have I understood you correctly? Interestingly, I received that advice from one other system dynamics practitioner years ago. Most of the advice I've received favor's Mary's approach.
I did write
http://www.iseesystems.com/(pwmxldaxedx ... arris.html
that gives my thoughts of a few years ago. It's written from the point of view of a Mary.
>> Having practised now the field for about 5 years and having made
>> standard programming for 40 years with all sorts of languages, I can
>> tell that SD is really a field whose difficulty stands at least two
>> points above current programming difficulty (with SQL,C or any language) .
There's a level at which SD seems easy. After all, it's merely hooking up a few components in a stock and flow diagram, adding some equations and parameters, and pressing ""Run.""
As you may be suggesting, though, there are multiple levels of complexity hidden in that description:
programming (for the model is a program, but this is likely the
easiest)
feedback control theory (an approach which can help make sense of a
model and its output; most likely nonlinear control theory will be
needed, involving as much classical or modern control theory as one
wishes or needs)
numerical analysis (to ensure the calculations work as intended,
although this is often just getting DT or TIMESTEP right if you don't
have an automatic integration routine)
statistics (to determine parameter values to use and possibly to
analyze simulation results)
modeling (to transform a problem into a model)
>> In France, saying that one practices SD, will have no effect, the
>> field being completely unknown. Of course not living in the USA and
>> not selling SD, I may be wrong and other reasons may exist.
I know some people in other parts of the world from where I live, and they see the USA as having a particularly restrictive view of the concept of ""systems thinking"" when compared to Europe, Australia, and elsewhere. To them, we in the USA limit systems thinking either to quantitative, simulation-based system dynamics or to CLDs and system archetypes.
They make the claim that those in Europe and Australia, among other places, tend to have a much broader conception of systems thinking (including SSM and a number of other approaches).
>> You mention that I know when to use or not use SD.
>> I have a better knowledge about it, but it is far from sufficient,
>> even with my problems that I know very well. I can then still be wrong about it.
That's true of all of us, I think. I've gravitated to people who tend to have multiple approaches at hand, because the first tool any of us picks (or the only tool we might use) might not be sufficient to make sense of the problem we're facing. I'm probably better at SD than at many other systemic approaches, but I feel more comfortable being able to switch methodologies if the task seems to demand it. I'd rather not do an SD model for someone simply because I know the tool, not because their problem needs it.
Some of those people have been associated with ""Systems Concepts in
Evaluation: An Expert Anthology""
(
http://www.wkkf.org/default.aspx?tabid= ... nguageID=0).
It can be also ordered in book form:
,----
| Williams B. Imam I. (2006) Systems Concepts in Evaluation - An Expert
| Anthology EdgePress/AEA Point Reyes CA.
|
| ISBN 978-0-918528-22-3 softbound, and 978-0-918528-21-6 hardbound
|
| Copies can be ordered from EdgePress PO Box69 Point Reyes, CA 94986
| USA or via Amazon
`----
>> To my opinion instead of looking towards wrong directions, the
>> profession should ask the right questions.
>>
>> 1. What are the characteristics of the business or any organisation
>> susceptible to be interested by SD?
>> The characteristics being their job, their size, their length of
>> service, the type of problems they currently encounter that can be
>> solved by SD, the expected profit and the expected cost (money,
>> time), the delay to deliver the study, the delay to notice the expected profits etc..
>> and once the characteristics are defined, how to practically find and
>> get in relation with such businesses and organisations.
You raise some good and hard questions. Part of the answer involves disclosing business models, I think, and I wonder if we're sometimes shy about sharing those.
For example, I could say that you can only address problems that are worth 100,000 Euros or more with SD. For certain approaches, that might be true, but then I'd also risk being embarrassed by other SDers who might claim to be very successful at working with small models to solve all sorts of problems. Or I could do the opposite, saying I could do a lot with small models, and risk being embarrassed by those who might point out the inaccuracies in my models when compared to their bigger ones.
I'm not sure that's the case, but I wonder if it is.
As you can see at
http://www.artsjournal.com/adaptistrati ... nt_22.html,
I'm trying a bit different and lower-key approach to see how it works.
So I think it's heavily dependent on the business model of the individual practitioner. I will say that I had a very easy time early on with managers who had been practicing control theory engineers. They understood feedback loops, and so they were very receptive to SD once they saw that feedback ideas could be applied to their organizational problems.
>> 2. How to verify that one prospect fulfils the above
>> characteristics, and that there are reasonable chances that the
>> prospect has real problems susceptible to be solved by SD.
The essence is still determining if the current problem can best be understood and fixed by assuming it is controlled by information feedback.
>> 3. How to verify that one has the necessary resources in expertise,
>> time, people to model that problem? To what point is it necessary to
>> already know the profession of the prospect. Sders who work with the
>> health profession seem to have better results.
I've heard it said that, in many cases, knowing the field is important.
I've also been told that I tend to be a generalist, able to go against that trend and work in a number of areas. I'm not sure if that makes it easier or harder for me to sell my services.
In any case, I think it's very important to listen carefully to the client and to work with, not independently of, the client.
>> 4. When one is sure to have a good prospect, how to convince him to
>> start a relation and what should be proposed at first?
>> Is it mandatory to start with SD, or is it possible or better to
>> start with something easier?
How would you like to be approached? I'm guessing most people would prefer to have someone inquire into their needs and get to know their problems, not talk about what they can do. That notion is supported by ideas such as Huthwaite's SPIN® Selling. Then again, that might be my blind spot.
>> 5. How to make sure of the quality of the service offered after the
>> prospect has been convinced?
That's a very important part. Assuming the SDer has a certain minimum competence, the process of delivering the service can be straightforward when things go well. When things don't, then it can take a lot of listening and a lot of skill to make the situation right again. So I think one side is technical competence and the other is consulting competence. For the latter, Peter Block's _Flawless Consulting_ is a classic text.
>> 6. At least at what stage are the real difficulties?
>> Finding a prospect, convincing him or doing the job?
Yes, yes, and yes.
Thoughts?
Bill
- --
Bill Harris
Facilitated Systems
Posted by Bill Harris <
bill_harris@facilitatedsystems.com>
posting date Mon, 16 Apr 2007 15:55:02 -0700 _______________________________________________