Ventana software support forum

Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr>

[ Was Feedback and Behavior (SD5408) ]

John Voyer writes

> I think the reason is that managers so poorly understand the integration
> that goes on in stocks and flows that the resulting dynamics are revealing
> to them.

May be it is 'not revealing' that was meant?
I have always been surprised by the fact that academics and consultants have
a very poor opinion of business people, and the reverse is true too.
A recent survey of business people opinion reported that 70% of them were
generally dissatisfied by their consultants.
Now about feed back loops that business people do not understand.
They do not understand the feed back loops or they do not understand the
utility of studying them?
The concept of feed back loop is very simple to understand, and I cannot
imagine a business man, accustomed to deal with all sorts of difficulties
not being able to understand so a simple concept. To my opinion there is a
confusion between understanding feed back loops and believing that making a
model that
takes in consideration feed back loops is useful for them.
For example I have been making SD models for now more then 3 years, and I
understand what a feed back loop is, and I understood it at the beginning
too, and I am more and more sure that in the majority of cases studying feed
back loops, in most cases does not help very much.
Why?
Because most of the problems encountered by business people are solved by
experimentation, which is possible because the consequences of their
policies
can be seen relatively quickly.
Where SD would help, is when the consequences are not visible soon, like
those you find in ecology or when you have only one shot, and there is no
second chance. The second case is difficult to study with SD, (no reference
modes for instance) and in the first case, people are less interested in the
long range consequences of their decision. This is the main reason of the
'not understanding of feed back loops' as explanation of the disinterest
towards SD.
Regards to everybody.
J.J. Laublé
Allocar, rent a car business
Strasbourg, France

Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr>
posting date Tue, 9 Aug 2005 14:28:58 +0200

Posted by ""John Gunkler"" <jgunkler@sprintmail.com>
I'm shocked by the last two messages from Jean-Jacques Laublé and hardly
know where to start to respond. I suspect others out there feel the same
way. In the past, Jean-Jacques has written, I thought, insightfully about
SD modeling -- but now he seems to be saying that the entire enterprise (SD,
I mean) is worthless!?!

His recent statements violate some of the most basic principles upon which
SD is founded.


1. He writes: ""...the consequences of their policies can be seen relatively
quickly."" But we understand that the immediate consequences of policy
changes are not the only consequences, nor (often) the most important
consequences. In fact, we see this so often that we use an aphorism to
describe it: ""Worse before better."" That is, the policies that work out
best (in the middle and long term) often make things worse in the short
term. And, by the way, policies that create immediate benefit often snap
back and make things worse than before after a little while.

So, I would say, business people (and most others) only fool themselves by
believing they can do ""experiments"" and immediately see if they're doing the
right thing.

2. He writes: ""The concept of feed back loop is very simple to understand,
and I cannot imagine a business man, accustomed to deal with all sorts of
difficulties not being able to understand so a simple concept."" But we know
that, in spite of the relative simplicity of the concept of feedback loop,
people in fact do not understand how they work in practice. There are
several well-conducted experiments, by Dr. Sterman among others, that
demonstrate that people (even highly intelligent people) are very poor at
predicting what will happen when a feedback loop is at work -- especially if
there is any delay built into the loop!! So, while the ""concept"" of the
feedback loop is simple enough, the consequences of having them around are
anything but simple and easy to understand.

3. He writes: ""...I am more and more sure that in the majority of cases
studying feed back loops, in most cases does not help very much."" But we
understand that the unwillingness to take feedback into account explains
much of the failure of businesses to achieve and manage growth. The entire
literature of SD, and its history of successful modeling, militates against
Jean-Jacques' statement. And the failure of econometric models to help
people truly understand and exercise control (or, at least, influence) over
economic results, and the continual failures of management ""flavor of the
month"" fads (simple, quick-fix anodynes), are testimony to what happens when
one does not take feedback into account.

There seems to be an underlying assumption here that business people are
successful in running their businesses these days. That's a joke! Look at
the failure rate of businesses!! Look at what happened, just a short few
years later, to the companies chosen as exemplars of great management in the
Peters and Waterman book ""In Search of Excellence"" -- and what you find
when you look at current business practices from a ""Lean"" perspective (most
manufacturing operations are only at a level of 3% efficiency, or less!!) --
and when you watch the continuing struggles of U.S. automakers. What does a
business manager do when revenues fall short these days? Lay people off!!
What are the consequences of laying people off? Immediate positive hit on
the bottom line, and then (next year, or even next quarter) even less
ability to do what needs to be done to be successful, and then (if the
business somehow manages to cling to survival, mostly because its
competitors made the same stupid decisions) the expenditure of massive
amounts of capital to try to rebuild the intellectual capacity they threw
away. SD could help such managers make much better decisions -- better for
the long-term survival and health of their businesses, and certainly better
for their employees!

People repeatedly make the mistake of believing that (even short-term)
success indicates good management. Thus, in the '50's and '60's, when the
U.S. was the only fully functioning economy in the world and GM became the
largest human institution ever built, people made the mistake of thinking
the U.S. had discovered the Holy Grail of business management. Robert
McNamara, of GM, took his supposed expertise and tried to use it to manage
the war in Vietnam with disastrous results. In truth, a monkey could have
been running an American business in that era and done just as well as the
so-called expert managers did. But, in part because of feedback loops which
those managers ignored, the decisions of the '50's and '60's eventually came
back to haunt U.S. businesses which dropped like flies when better-managed
Japanese and European companies began to compete.

And I haven't even touched on the most salient points!
Posted by ""John Gunkler"" <jgunkler@sprintmail.com>
posting date Wed, 10 Aug 2005 09:57:44 -0400

Posted by ""Zagonel-Santos, Aldo A"" <aazagon@sandia.gov>
>I'm shocked by the last two messages from Jean-Jacques Laublé and hardly
>know where to start to respond. I suspect others out there feel the same

I find this reply to be too excessively judgmental to be constructive and appropriate
for this list serve. I am saddened by the use of patronizing statement ""in the past,
so-and-so has written, I thought, insightfully about SD modeling -- but now...""
There is plenty of room for disagreement around the issues, but to judge someone's
statement, perspective, approach to be ""invalid"" is presumptuous and precludes good
people from engaging in the field of SD. I am also disturbed by the use of the third
person since we're all in the ""room"", including the author of the statements and
opinions.

Aldo A. Zagonel
Albuquerque, NM
zagonel@aol.com
Posted by ""Zagonel-Santos, Aldo A"" <aazagon@sandia.gov>
posting date Thu, 11 Aug 2005 12:06:36 -0600

Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr>
Hi everybody

At least having not convinced any body I brought some animation to this
list, which due to the character of the field is not an easy matter!

Now about what I wrote.

John Gunkler writes about what I wrote:

<but now he seems to be saying that the entire enterprise (SD,
<I mean) is worthless!?!

Where on earth did I write such a thing?

By the way if I was convinced of this, I would have better to do then loose
time participating to this list.

<So, I would say, business people (and most others) only fool themselves by
<believing they can do ""experiments"" and immediately see if they're doing
the
<right thing.

Most businesses managers are faced by short term decisions. These short term
decisions are governed by experience that comes from experimentation.

I never said that this practice is good or bad, but it is a reality, and is
on top of that sometimes good and sometimes bad.

<But we know
<that, in spite of the relative simplicity of the concept of feedback loop,
<people in fact do not understand how they work in practice.

I said that business people understood the concept of feed back, but never
pretended that they understood their implication, particularly when there
are many of them.

On the contrary I agree completely. This is why being convinced of the
difficulty to understand the working of feed back loops, I prefer to use
them with parsimony, and in the most useful manner.

<He writes: ""...I am more and more sure that in the majority of cases
<studying feed back loops, in most cases does not help very much.""

I confirm this. Most cases mean more then 50%. This leaves plenty of cases
to study.

I think that in fact it is much more then 50%, because people are mainly
concerned by short term problems, and not enough concerned by long term
decisions.

If it is a good or bad thing depends on the problem, and generally people do
what they think is best and it is always easy to say after what they should
have done.

I will give an example, of a seemingly bad decision that is in fact a good
one.



Our field is in the worst situation even known, reported by our syndicate
since second world war.

This is due to a decreasing demand, due to a small growth, and to the fact
that many businesses in the automobile sector, seeing their activities
declining, started new activities to increase their margin. Some chose
renting because it looks an easy job. It is in fact an easy job, but
difficult to earn money from.

So presently, more then 50% of rental companies in France, struggle to
maintain their cash, whatever the long term consequences. A company always
dies from lack of cash.

So prices are going down, often at the cost of the stock of vehicle getting
older very quickly which biases the value of the stock of vehicles in the
accounting and brings less cash when sold. But at least it keeps them alive,
waiting for the competitor to die before them and reduce the level of
competition and restore profits.

What it the best solution? Die in good health having preserved the future
because you did not sufficiently lower your price and lost too many
customers, or survive reporting the problem to the future in bad health but
still having the chance that next year may be better and recover?

<And the failure of econometric models to help
<people truly understand and exercise control (or, at least, influence) over
<economic results, and the continual failures of management ""flavor of the
<month"" fads (simple, quick-fix anodynes), are testimony to what happens
when
<one does not take feedback into account.

I always look at reality as it is, and anybody will recognize that if you go
to a science book shop in France, you see dozens of econometric books, and
generally not one on SD.

I do not mean that one is better than the other, but it is a fact, that has
some explanation, which does not lie necessary on the quality of the
methods.

On thing is sure, if SD was working exceptionally well on most subjects, you
would after nearly 50 years, find at least one or two books on the subject.

<There seems to be an underlying assumption here that business people are
<successful in running their businesses these days.

<That's a joke! Look at
<the failure rate of businesses!! Look at what happened, just a short few
<years later, to the companies chosen as exemplars of great management in
the
<Peters and Waterman book ""In Search of Excellence""

Papers are more ready to report bad stories then happy one. It conveys often
a bad view of reality. The businesses are doomed to fail one time or
another, to leave the place to other more competitive or more ready to
struggle to get something they never had.

Well I tried to explain better then I thought I had done it, my ideas.

I think that SD will still grow and probably be one day a well known and used
methodology, but not in the near future and principally applied to long
range decisions.

I apologize if I shocked some of the members of the list. This was not my
intention.

Regards to everybody.

J.J. Laublé, Allocar
Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr>
posting date Thu, 11 Aug 2005 15:49:35 +0200

Posted by ""Bill Braun"" <bbraun@hlthsys.com>

>> I said that business people understood the concept of feed back,
>> but never pretended that they understood their implication,
>> particularly when there are many of them.


We rely extensively on micro worlds in our graduate management program and
it takes students three games on average to understand the concept. These
are all working managers with 4-5 years of experience, on average. (Shayne
Gary's paper at the conference explored sustained learning and changes in
mental models. His work seems relevant here.)

I ""reverse engineered"" the microworld into a large CLD model which I
presented to several groups of students before they began the game. I
would say that their conceptual understanding was far better than the
average manager. Notwithstanding, their ability to leverage that into an
above average performance in the microworld never materialized.

Statistically, n=2 is meaningless, but it does keep in the forefront of my
thinking the question of how much managers understand about dynamic
behavior in their organizations.

Bill Braun
Posted by ""Bill Braun"" <bbraun@hlthsys.com>
posting date Fri, 12 Aug 2005 10:43:12 -0400 (EDT)

Posted by Mark D Neff <mneff@csc.com>
Very interesting.

I read this message from Aldo A Zagonel-Santos and was immediately
interested to see what could have evoked such a response. I then read the
previous message from Jean-Jacques Laublé. After sitting for a minute, I
decided that I should take a minute to provide my feedback. While I
understand the conviction in both messages, I think we may be missing a
great opportunity.

I believe that we can use SD to model short time frames. The feedback loop
does not have to be in terms of months or years - although I agree we
frequently focus more on long term projects. We could be studying the
propagation of fruit flies or bacteria and see entire generations come and
go in days or even hours. So the unit of analysis does not have to be long
time.

If we were to model decision making today and financial results we may
feel that we are constrained by current financial accounting standards and
quarterly results. It does not mean that we can't go to a shorter time
frame for our analysis. Nor does it preclude us from putting in mechanisms
to get instant feedback. What do you think RFIDs do? We can put them on
our supply and receive instantaneous feedback on the outflow of our stock
from a store. We can then instantiate order systems to replace stock on a
minute by minute, basis if we felt that is what was needed to get the
maximum sales of a particulat product. I would argue that the diversion of
resources would not warrant the additional sales but it would be
interesting to look at.

There are a number of other ways to apply this as well. Instead of looking
at one post and then denegrating it or dismissing it, we need to be
conscientious that each of us has a voice or an opinion and to respect it
for what it is while we apply what we know about SD to determine what we
can do with this observation to revise our model to more appropriately
reflect what we see happening in real life.

There are several directions we can take this conversation, I would like
to see what others think and how they would progress developing a model
with more short term feedback loops especially regarding decision-making
which does seem to be done spuriously from an observers standpoint. What
we may in fact be missing is that the decision maker is actually making a
decision from a position of ""deep smarts."" They have been studying
patterns of behavior for decades and based on that experience are able to
make a decision in seconds from data that we would need to study for weeks
before we could make an appropriate decision - and even then we might
still make the wrong decision because we just don't have the receptors in
place that allow us to make the right decision.

What mechanisms are in place today to simulate short time frames that are
more realistic in our business processes. Each model has its relevance, It
is up to us to help find the right model for the situation that we are
facing.

Mark Neff
(856) 354-1654
1014 Pinebrook Road
Cherry Hill, NJ 08034
Posted by Mark D Neff <mneff@csc.com>
posting date Fri, 12 Aug 2005 08:48:25 -0400

Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr>
Hi Bill

You wrote on my remark

>> I said that business people understood the concept of feed back,
>> but never pretended that they understood their implication,
>> particularly when there are many of them.

>> We rely extensively on micro worlds in our graduate management program and
>> it takes students three games on average to understand the concept. These
>> are all working managers with 4-5 years of experience, on average. (Shayne
>> Gary's paper at the conference explored sustained learning and changes in
>> mental models. His work seems relevant here.)



The problem of misunderstanding lies in the definition of 'Concept'.

To my point of view, as soon as students begin to play the games, one

leaves the concept to its implication.

For me the primary notion of feed back is vey simple to understand.

Every body has some economies, and understands that the more they have money
in the bank, the more they will get interests and the more they will have
interest the more they will get monney in the bank etc.. This is for a
positive feed back. For a negative feed back, you can take the very simple
example of level of fatigue.

If the level of fatigue is low, it will tend to increase, because you go for
a walk. When the level is increasing it will after a while decrease because
you will stop walking, same for hungriness etc.

Regards.

J.J. Laublé
Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr>
posting date Sat, 13 Aug 2005 15:05:01 +0200

Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr>
Hi Mark



It is just after having worked hard on trying to simulate short term price
setting on a model based on days, that I decided to leave it for the moment
aside, and explore the alternative of setting up the right indicators and
trying by trial and error to improve the price setting process.

With the difference that it is made in a methodical way, where I try to
formalize the price setting modifying process, and eventually to program it
in an ordinary programming language.

This method is possible because of the short term implications of the
decision made and to its recurring process.

Regards.

J.J. Laublé
Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr>
posting date Sat, 13 Aug 2005 15:22:20 +0200

Posted by ""Weaver, Elise A"" <eweaver@WPI.EDU>
> see what others think and how they would progress developing a model with
> more short term feedback loops especially regarding decision-making which
> does seem to be done spuriously from an observers standpoint. What we may in


You might enjoy ""Blink"" by Malcolm Gladwell about fast decision making.

Elise
Posted by ""Weaver, Elise A"" <eweaver@WPI.EDU>
posting date Sun, 14 Aug 2005 09:02:08 -0400