posted 26 March 2000 10:49 PM               

> Marc,
>
> How do you know you have true root cause in a corrective action process?

Like any investigation, you form a hypothesis and then design tests to prove or disprove it. You know you have the root cause when your experiments tell you so. Now, I say that with some hesitation, as it has been know to happen that an incidental cause is identified and corrected and the problem lessens to a point that it is not significant when compared to the magnitude prior to the fix. Same case in the event of multiple root causes (it happens).

> work in company were people answer 8Ds and corrective actions without
> identifying the real cause of the problem.
> People ask why but not always the
> right whys to reach the reason the problem occurred.

Not unusual. In fact, pretty common.

There is no way to tell someone when to stop asking WHY. There is no way to tell someone what WHYs to ask. There is training and experience. With consideration to Columbo, Perry Mason, Matlock, and all the detectives and cop shows - it should be evident that some folks 'have the gift' of asking the right questions and some quite simply do not. Training works for some folks. Some folks simply do not grasp the concept. Some people simply don't care (typically they believe they have a job to do and 8D investigations are not what they were hired to do).

Determination of root cause is not always straight forward and sometimes, in fact, a fix corrects the problem yet the fix does not address the 'true' root cause.

8D is supposed to be a cross-functional exercise. When this is truly the case often the differing ideas uncover the root cause. However, many times it becomes a matter of someone leading the 8D who has 'mastered' the art. This is the reason for structured investigations with tools such as fishbone diagrams and such.

> What is your answer to this predicament

Training - or pull a system into place where you appoint a 'master' to review and participate in appropriate meetings.

> and do you know of any web information that addresses the...

posted 04 April 2000 08:24 AM               

quote:

---

Originally posted by Marloun:
Same question, during the why-why analysis, when do we know that we still have a symptom and not a root cause? ....
My question will be, for inexperienced people, how do we know that we have the root cause?

---

Was your corrective action effective in eliminating a reoccurrence? If not, did you treat the root cause or a symptom? Probably a masking symptom.

Asking "Why?" as traditionally written, five times, is not a fast rule. You may ask it twice to determine the true root cause, you may ask it eight times. The individual(s) doing the problem solving must agree that they have uncovered the true root cause and treat that. Experience, with the theory of '5-Whys?' will help to guide the process.

Regards,

posted 05 April 2000 05:39 PM               

You make a very good point! Here is the problem. Asking the 'where' and 'when' can be introduced to the process at 3-5-8 'why's.

How do you know when to stop asking 'why' and begin to get in to the 'where' and 'when'? Any ideas? (PDSA)

Regards,

posted 05 April 2000 06:14 PM               

If the first three responses agree it should be time to move on to a what/where to change issue...

If all three resulting answers are completley at odds then I would think it's time to repeat the Why question a few more times, until a similar viewpoint (or appropriate solution starting point) was understood & agreeded to by all involved.

Then its time to plot the what/where PHYSICAL change - put the change in place and find out if your Why result was correct in the first place...

posted 07 April 2000 08:40 AM               

Plan: create a theory of what went wrong. How many 'whys' you ask, either as a group or independently, is up to you. The 5-why technique is meant to drive the thinking beyond the superficial thinking. 3-5-8 or more, it only matters to the extent of how far you drive the thinking process.

An example

Problem: fuse on the press keeps blowing.

Q: Why did the fuse blow?
A: The press motor overheated.
Q: Why did the motor overheat?
A: The strain on the motor was excessive.
Q: Why was the strain excessive?
A: The material thickness used to make the part was too thick.
Q: Why was the material too thick?
A: The incoming material was over specification
Q: Why was it accepted over specification.
A: Gauge used to accept the material was out of calibration
Q: Why was it out of calibration?
A: We do not have a calibration program.

If we react to the 'problem' with superficial thinking, we end up replacing fuses, never addressing the root cause. We instead address a symptom. If we react to the line of questioning, we create a calibration program (which extends beyond the singular event) and perhaps react to some of the smaller whys as appropriate. Some problems have multiple root causes.

Do: train on and deploy your theory to eliminate the problem.

Study: Did you eliminate the problem? How did your theory work?

Act: Standardize or revise your theory.

As we are all aware, standardizing our practice results in a formalization of a process. We create procedures to address the 'when' and 'where' as Jim pointed out. Many times the process does drive you to gaps in the program or areas of fuzziness.

Just some more for the mix. Back to the group....

Regards,

posted 07 April 2000 12:04 PM               

Would not the following viewpoint draw us to the same conclusion?.

Q Why did the machine blow a fuse?
A Materials were oversized
Q Why are materials oversized?
A Gages used to verify incomming materials were not calibrated
Q Why were gages not calibrated
A Inadequate or non existant cailbration program

Same conclusion as in your example - less why's asked... the end result remains the same.
Do: Apply the theroy
Study: Was the problem corrected?
Act: Standardize the methods

IF the STUDY result of "Was the problem corrected" in either case proves that the problem was not corrected then and only then could you discover if you had stopped asking why too soon.

If you asked the exact same question, recieved opposing expierience information then more why questions would be needed.

Q Why did the fuse blow?
A1 Bad fuse
A2 old/worn machine
A3 Incomming power surge
A4 Materials too thick

Opposing answers would lead us to conclude that each branch answer would need investigation prior to starting the plan for application stage.

The number of why's to ask is directly dependant on the information received, and we won't know if enough whys were asked until action to apply the theroy has been made.

posted 07 April 2000 04:19 PM               

You are correct in pointing out that the information supplied about the problem needs to be detailed (as much as is possible). My example is rather simple and could have been shaved down to fewer 'why' questions being asked.

An incorrectly stated problem, or lack of detail, could lead an investigation off track. In my CA training, I stress the importance of giving as much detail as possible. It can only help.

Regards,