Sunday, September 12, 2010

Root cause for root cause analysis failures

When we were at the Marcus Evans conference last week in Melbourne the topic of Root Cause came up.  Several people, including me, commented about the common failure of root cause.  The most common outcome of root cause analysis is "no root cause identified" or "multiple potential root causes".

I raised at the meeting that in my opinion, the root cause of this failure of root cause analysis to find a root cause (this root cause run on is intentional!) is usually because we do them badly and we do them wrong.  Wrong time, wrong way, wrong indication.

How many incident forms do we create that say "what was the root cause of this incident?"  How many accreditation reports say  "in the event of a interlaboratory (PT or EQA) comparison failure, a root cause shall be performed to detect the cause".  In how many group discussions about a problem do you hear someone ask "so what was the root cause?"

Sorry, but with the best of intentions, this is all "nonsense".  I think these questions come from a common source: folks who have never done a root cause analysis, and have no idea what a root cause analysis entails. I understand they know about "Five Whys?" but that is about as helpful as the Six Sigma "5S".

The reality of root cause analysis (RCA) is that is takes a lot of time, and a lot of people's effort.  RCA requires  information gathering, and charting, and group discussion and consensus.  They can take days to organize and weeks to complete.

Any if anyone thinks that it  is  productive to spend time for every incident that gets reported, or every PT error to have a full RCA, then they have too much time on their hands, or they have lost touch with reality.

In most situations, a best guess, or  informed appraisal is pretty good (and in most situations, good enough) , and will likely get you moving in a reasonable direction.  Obviously, a supervisor sitting in their office, can make a reasonable guess, but not necessarily one that you would want to "bet the farm" upon. It will take  more investigation than that to raise it to the level of being an "informed" appraisal. 

To male it worth your while, it probably makes sense to postpone doing an RCA until you have seen the same error come up several times, or reserve RCAs for when a seriously major BAD THING has happened.  Really bad things have a way of galvanizing attention, and repetition makes pattern recognition easier.  If you have made a few "informed appraisals" but the same situation keeps on coming back, maybe it is time to do a full and formal RCA.

So can we agree that expecting an RCA for every event every time is a waste of time, and speaks more about the requester than it does the event reporter?

For those interested you can go to the Program Office Website [ ]and find the presentation "Bad Things in Small Packages".  I put together two flow charts.  One is the "Don't Ask - Don't Tell" and the other is the appropriate flow.
For those unsure, the Don't Ask - Don't Tell chart is a JOKE.

"Bad Things ..."  (and other files from the Marcus Evans Conference) will be available, I anticipate, on Tuesday September 14, 2010.

There are tons of good books available.  In our course we use Root Cause Analysis: Simplified Tools and Techniques, Second Edition by Bjørn Andersen, but there are many others at at comparable price.  
 Simple is good.

PS: modified for spelling and grammar - September 5. 2013.


  1. I absolutely agree with your assessment of root cause analysis. In addition, I would like to mention the difficulty that can sometimes be faced in a union environment - whereas it is not only time consuming to find the real 'root cause' by performing an in-depth analysis, but at times also impossible to fight against if it involves, for example, undesirable personality traits and/or behaviours that have resulted in an incident. The union rules make it difficult (if not impossible) to reprimand an employee for bad behaviour. Perhaps a supervisor could assume to know the 'root cause' and perhaps save the time of performing an analysis if knowing that the root cause cannot be addressed sufficiently? In this case, perhaps time is better spend by silencing those that continue to see and voice that deficiency?

    I continue to marvel at your blog. It fascinates me to see the other side of the coin. Whereas you make constant references from a "Quality" stand-point from the top down, I continue to see things from the bottom up. Dare I say, the view from down here is not as rosy as I would like to see it. All the great intentions loose their meaning if not all of those affected are on-board.

    I wish there was something I could do to makes things better from this end.

    (Perhaps leading by example can do something - although it has given me nothing but grief so far, I will continue to try.)

  2. Thanks for the comment.

    Top-down and bottom-up. Can we agree that both are important, and that both efforts do not have to meet in the middle. From my perspective, management has a greater obligation, and has to pull the greater weight. Its kind of like George Orwell in Animal Farm. All animal are equal; but some animals "have to be" more equal than others. I know that is not what Orwell said, but you get the point.

    Leading by example is a good thing. And I understand that we are not always sucessfull (I could mention about taking horses to water..., but I won't) Sometimes it works, sometimes it doesn't.

    But ultimately it is about personal vision, and personal commitment, and the personal belief that what we do can make a difference.

  3. Thank you for posting the presentations. I've enjoyed reading them very much. I made a point of paying particular attention to the "Don't ask - don't tell" chart and almost fell off my office chair as I was laughing too hard. I believe you captured it all!

    I also enjoyed the 'Transform your culture for quality' presentation. I wish I could present it at our laboratory - I would HIGHLIGHT the "AT RISK BEHAVIOR" slides in neon pink!

    Thank you,


  4. Thanks N
    Files are made available for folks who want to use them. We do not have copyright lawyers who are monitoring use, and we don't have a download fee.
    We do request is that people acknowledge the author and the blog site (


  5. This comment has been removed by a blog administrator.


Comments, thoughts...