Root Cause

["Robert Crow" <jr1crow@mindspring.com>]

Hello Group:

I have decided to wade in on the Root cause discussion.  There is a definite difference between special cause and root cause.  When you are tracking something with a control chart or simply a run chart most of the variation occurring will be random or common cause variation, but from time to time a trend, points outside the control limits etc., will develop.  These are called special cause variation and are indicators that something has happen outside the normal system.  In other words, a deviation from the norm.  This can also be a definition of a problem, a deviation from what is normal.

This special cause variation is not a root cause, but a symptom of the problem, an indicator that a problem exists.  When you have a special cause of variation there will be a need to do some problem analysis to determine the "probable cause" of the problem.

One of the chief barriers to effective problem solving is the "leap to judgement".  Before the problem has even been clearly defined someone says, do this, that will fix it.  The problem with this is that it may temporarily fix the problem, but it will reoccur the next day, next week, or next month.  To correct the problem permanently you must get to the "root Cause".  For this you need a method, or a problem solving process.

1.  Recognize that a problem exists - What is actually happening vs what should be happening.  Having control charts set up to track what is happening are very helpful here.
2. State the problem in writing
3. Review the problem statement.  If you have a team working on this make sure the team is in agreement that the problem statement is an accurate description of what is happening.
4. Describing the problem - what, where, when, and extent.  What is and what is not happening.
5. Determining the probable cause of the problem
6. Verifying that the probable cause is the true cause.
This is not unlike the Plan, Do, Study, Act of continual improvement.

If you are only experiencing random variation there will not be a "root cause", but you may still have a problem.  An example could be that your customers consider on-time delivery to be very important, but when you gather data and construct a control chart of on- time deliveries you determine that you are only able to deliver products to the customer on-time 78% of the time.  Further more the variation around that figure may be exhibiting random variation with no trends or points outside the control limits.  In this situation there will not be a "root cause", but possible a series to events in the order fulfillment process that are contributing to the problem, and the solution will be more involved.  You will need a cross organizational group of all people/departments involved in the process to meet and come to an agreement of what the problem is and is not.  Flow charts of the total process and then sub charts of each step in the process.  Over time and with dedicated commitment by all involved the problem can be corrected.  On going control charts are essential to maintain the groups interest and commitment and to track how each improvement to the process is resulting in an improvement to the companies ability to provide product to the customer on-time.  A starting point here is to make sure that the company and the customer have the same definition of on-time.


Robert Crow
jr1crow@mindspring.com