Limitations of Root Cause Analysis

Root Cause Analysis > Forum Log in

Limitations of Root Cause Analysis
Gary Wong, Consultant, Canada

The RCA model assumes:
1. Linearity
2. The existence of cause and effect relationship
3. That a system can be broken down into components and each fragment analyzed (reductionism).
In the past, this has been the prevailing thinking since the 1930s.
But in today's reality, many problems have become non-linear with many unknown and unknowable connections amongst human and technical components. For such other problems, other tools are to be preferred.

Limitations of Root Cause Analysis
A. Achour, Analyst, Germany
One point to add on RCA models if I may. RCA as known generally focuses on problems then moves on to cause and effect to get to the root cause.
I find this approach unpractical while 4. problems take different forms from various perspectives. A problem for one party isn't necessarily seen as a problem from another party and this could lead to conflict but if the influential party wins the debate (which is always so), the solution is going to be based on the problem as seen from the stronger party. We see many solutions today resolve the problem core, but give birth to new problems.
Another approach which I find a lot more practical is to focus on the effected company goals and from here ellaborate on the cause and effect relationship.
In this way, no matter how parties disagree on how they understand the problem, their perspectives on affected goals are all the same.

Limitation of Root Cause Analysis in Management
Samir Desai, Manager, India
If one hopes that RCA in management problem will also throw up the results as sharp as the RCA in technical problem (say maintenance) than one is surely in for some disappointment.
RCA will be successful in management problem only if 5. all the involved functionaries have a stake in its success. This is the biggest limitation of RCA, or this is the most crucial element management will have to incorporate in any RCA exercise.

Requirements of Root Cause Analysis
A. Achour, Analyst, Germany
@Samir Desai: the success of an RCA initiative in problem management depends first and foremost on the following elements:
1. Company's policy on risk management (processes, scope and empowerment etc...)
2. Investigator's approach and method used
3. Information/data at hand like reports, stats etc...)
If any of the above components is not fully respected, the RCA is doomed to failure at start. On the second position, if the focus is mainly on the company goals, then every stakeholder has to be concerned with the investigation because when goals are at stake, we are touching the company's strategy. As for the remaining components or requirements for RCA, they will fall into place like in a puzzle.

Linearity is Still a Reasonable Assumption Most of the Time
Bill Wilson, Analyst, Canada
@Gary Wong, you mentioned the 1930s as a date of origin for root cause... Are you referring to Heinrich's Domino Theory of accident causation, or Toyoda's 5 Whys for general problem solving? Whatever the case, I respectfully submit that for most reputable root cause practitioners, those models are interesting primarily from a historical perspective.
Strictly linear, single chain of causation models are today used only for the simplest of problems or by the least experienced/knowledgeable practitioners. The most common alternative is probably the multi-chain cause/effect model based on necessary & sufficient logic. This usually still assumes linearity, but that's okay because non-linearities are mostly confined to individual components or processes that can be treated in "black-box" fashion within the larger context of an analysis.
Most problems to be solved are still happening in the "merely complicated" world; true complexity is still relatively rare in most business contexts.

Apply Complexity Science
Gary Wong, Consultant, Canada
@Bill Wilson: Thanks for catching my typo; it should be the 1950s when Charles Kepner and Ben Tregoe introduced their KT analysis. RCA can still work in a Newtonian world where linearity and Gaussian (normal) curves exist. Deductive and inductive thinking are valid. However, in a non-linear Pareto (power law) world, abductive thinking is required.
Other preferred tools I refer to my opening comes from applying complexity science where we go in the opposite direction of Reductionism. Instead of breaking down a system, we ask how do things assemble themselves? How do patterns emerge from these interacting parts? These patterns can lead us to new solutions which are more valuable than finding a root cause.


    Do you wish to study further? You can learn more from the summary, forum, discussions, lessons, courses, training, instructions, expert tips, best practices and education sources. Register.  

Special Interest Group Leader
Bill Wilson

More on Root Cause Analysis
Best Practices

Expert Tips


About 12manage | Advertising | Link to us | Privacy | Terms of Service
Copyright 2016 12manage - The Executive Fast Track. V14.1 - Last updated: 25-10-2016. All names tm by their owners.