Monthly Archives: August 2011

Thoughts on root cause analysis -part 1

I find myself thinking quite a bit about problem solving and root cause analysis because I enjoy the topic so much. One topic that always stirs much discussion is the topic and flawed belief that there is always “ONE ROOT CAUSE” to a problem. The quality toolbox is filled with many tools for conducting root cause analysis such as 5 Whys, fishbone diagram, Tree Diagram, etc…The 5 Why is a perfect example of this linear thinking approach that leads one to believe that there is one single cause, or that if you ask “Why” 5 times you will always arrive at a root cause. I am not knocking the 5 Whys approach as I believe it is necessary to ask why so as to not jump at the first “symptom”. I do believe however, that you have to use these tools with some common sense. Here is an example:

Problem – Car won’t start when leaving for work in the a.m.

Possible causes – dead battery, alternator, fuel filter plugged, fuel pump  not working, clogged fuel injector, etc.. (note I am not a mechanic)

You test the battery and it is dead (do you replace the battery and move on?). Let’s assume you do that and a day later it is dead again. Uh-oh, not enough root cause done. You find that the alternator is bad. The car has 90.000 miles on it, but why did it fail? Do you replace it and move on or do you dig even deeper?

Possible next level of causes of failed alternator – OEM suggests replacing it after xxxx miles (which you did not do). Is that the cause? Do you dig even further to find out exactly what failed in the alternator? “Why”? Most would not. Why? – it is beyond your scope of controllable variables typically. So by not taking it as far as humanly possible for example to the design level are you truly finding the “ONE ROOT CAUSE”, or are you stopping at a convenient spot on the “Cause and Effect Chain”? You can control changing the alternator out at the recommended interval, but not necessarily the design (sure you can complain, but it did meet the warranty most likely).

This same things happens in the manufacturing world when there is a problem. We stop at one of the following Cause and Effects points:

1) A cause that has come up in the past (familiarity and a ready made fix exists)

2) A cause that is within the sphere of our influence (maybe our dept., but what if it goes beyond that)

3) Only looking at the immediate issue or defect that occurred (Process Root Cause) but not continue and look at why the defect escaped our detection and was shipped to a customer (Detection Root Cause). Also – there are many times “Systemic Root Causes”. I will discuss Detection Root Cause and Systemic Root Cause in my next blog. 

Sorry about the length of this one but I love thie topic.

Happy Problem Solving!