REPLY Definition of root cause (SD6826)

[SDMAIL Jack Harich]

Posted by  Jack Harich <register at thwink.org>

SDMAIL John Gunkler wrote:
> Posted by  "John Gunkler" <jgunkler at sprintmail.com>
>
> Root cause analysis is a foundation of the process improvement 
> methodologies (such as Six Sigma) in business.
>
> I am not speaking authoritatively here, but it seems to me that what 
> happens in real life is that we stop looking for a root cause when ...
> (1) We have found (at least) the fifth one [there's a method called 
> the "5 Why's" that suggests you ask "Why?" at least five times, 
> iteratively, before accepting the answer.]
> (2) We suspect that delving any deeper will not be productive; and
> (3) We have arrived at something we can change that we're confident 
> will eliminate the problem(s) we're investigating.
>
> Obviously, these are "principles of practice" not theoretical 
> absolutes. Perhaps they apply, in some form, to SD modeling as well(?)
>   
Interesting how close these are to my earlier proposition:

"A root cause has three identifying characteristics:
(1) It is clearly a major cause of the symptoms
(2) It has no deeper cause.
(3) It can be resolved.
Sometimes it’s useful to include unchangeable root causes in your model 
for greater understanding. These have only the first two characteristics."

Point 3 is a perfect match. Your point two is an improvement on mine. 
I've been troubled by the lack of cost benefit as you continue to look 
deeper and broader for root causes. Changing mine to "It has no 
productive deeper cause" would seem to fix this. Thanks!

Our versions of point one differ. Yours essentially says you have dug 
deep and have what might be a root cause. Mine doesn't imply digging 
deep at all, and merely says "it" is a major contributor in the causal 
chain. My point two is the reason point one need not say how deep.

But I don't think we're there yet with a standard, applyable definition 
of root cause for SD. This is because the field still clings to phrases 
like "dynamic hypothesis" of what is causing the problem instead of root 
cause and feels that if a model can reproduce the symptoms, then it must 
contain the root cause.

I wonder if the preference for dynamic hypothesis over root cause 
reflects the fact that SD is used primarily for business system 
optimization. Episodic problems trigger an SD analysis. While the 
problem does have a root cause(s), decision makers prefer to see the 
broad behavior of the business modeled, so they can evaluate a broad set 
of scenarios. This mindset would lead to preference for no particular 
root cause, but rather a general feeling for the many causes that lead 
to many behaviors.

> [Parenthetically, I feel obligated to say that the one thing that 
> makes me most uncomfortable about the Six Sigma mindset is the 
> assumption that there is one root cause, and that changing one thing 
> will solve a problem (without unintended side effects.)  The good news 
> is that the Six Sigma community is rapidly discovering System Dynamics 
> and there is a chance that feedback loop thinking will prevail in the 
> future.]
>   
Hmmm, I don't see Six Sigma assuming one root cause. Maybe a simple 
introduction made it seem that way.

For example, a fishbone diagram shows many causes. See 
http://en.wikipedia.org/wiki/Ishikawa_diagram

The Six Sigma Handbook, by Pyzdek, 2003, page 63 says "These results 
are, in a sense, 'effects' caused by things taking place in the process. 
.. At some level we reach a 'root cause' or most basic reason behind an 
effect. Black and Green Belts learn numerous tools and techniques to 
help them identify these root causes. In Six Sigma work, results are 
known as Ys and root causes are known as Xs. ... Y = f(X1, X2)" Note the 
plural in the second and third use of the term "root cause" and that Y 
can have multiple root causes, in the form of X1, X2, X3, etc.

BTW, nice story about the chain of tortoises.

Owen Ambur wrote:
> Jack, you may also wish to check out Dietrich Dorner's book entitled 
> "The Logic of Failure."  He says we "court disaster in predictable 
> ways."  My paper on his book is available at 
> http://mysite.verizon.net/ambur/failure.htm  
Thanks. I read it in October 2001. Fantastic read. A super sleuth. This 
book was one of the reasons I developed an interest in being able to 
model problems.

I found the list of decision maker bad habits on page 18 to be very 
educational:

- Acted without prior analysis of the situation
- Failed to anticipate the side effects and long term repercussions
- Assumed that the absence of immediately obvious negative effects meant 
that correct measures had been taken
- Let overinvolvement in "projects" blind them to emerging needs and 
changes in the situation
- Were prone to cynical reactions

This applies to so many difficult dynamic problems.
> Posted by  George.McConnell at selex-comms.com
>
> Just to throw another source into the ring, you may be interested to read
> http://www.systems-thinking.org/rca/rootca.htm
>
> regards - George

A good introductory read. I liked "Research has repeatedly proven that 
unwanted situations within organizations are about 95% related to 
process problems and only 5% related to personnel problems."

I wonder if a thousand years from now, historians will write "Research 
has repeatedly proven that unwanted situations within problem solving 
paradigms are about 95% related to process problems and only 5% related 
to personnel problems." :-)

"The Plant Manger was horrified when he realized that he was the reason 
there was oil on the plant floor. Bingo!" - Nice to see twists like 
this, to make the point and delight the reader.

"Once the root cause is determined then it has to be determined whether 
it costs more to remove the root cause or simply continue to treat the 
symptoms." - This deals with the cost benefit trade off on resolving 
different underlying causes. The piece is a little off, because there is 
more to choose from than treating the symptoms and removing the root 
cause. There are all the intermediate causes between those two ends of 
the causal chain.

"Seeking the 'Root Cause' is an endless exercise because no matter how 
deep you go there's always at least one more cause you can look for." - 
This would depend on how you have defined root cause. By including "(3) 
It can be resolved" in my definition this trap can be avoided. You stop 
when you approach causes like water freezes at 0 degrees Celsius because 
that cannot be resolved. Models are full of causal chain ends that are 
natural constants and are unchangeable.

Clicking on the link to http://www.bill-wilson.net/root-cause-analysis, 
what do we find about root cause analysis? "It can be employed in almost 
any situation where there is a gap between actual and desired 
performance." Isn't this pretty much any problem that system dynamics is 
applied to?

Thanks, George
Posted by  Jack Harich <register at thwink.org>
posting date  Tue, 18 Mar 2008 20:07:59 -0400