Root cause is defined as the factor responsible for any anomaly or inconsistency that leads to deviation from specified standards. It is the core of the underlying issue. A root cause can also set in motion several cause-and-effect reactions that may or may not be directly related to it but could ultimately impact the end result in production.
Root cause analysis (RCA) collectively denotes the techniques and methodologies used to identify the cause of any deviation or irregularity in a given system. There are a wide range of tools, techniques, and approaches that enable these analyses. While some of these methodologies are adopted to identify and rectify the core of the issue, some are used to offer leverage and support to other RCA activities.
RCA is generally viewed as a reactive measure. However, comprehensive application of RCA techniques can also aid in creating a proactive system that can predict problems before they occur.
Before evolving into the form as we know it today, RCA first made its appearance in the field of engineering. Much like lean manufacturing, the beginnings of RCA can also be attributed to the management philosophy instituted by the Toyota Production System (TPS). One of the key contributions of Sakchi Toyoda along these lines was the formulation of the “5 Whys”. The basic underlying premise here is asking “why” five times consecutively to arrive at the core of the problem.
For instance, let’s assume the internet is down. These would be the 5 Whys you would ask:
So, typically the 5th why would lead to the core of the issue.
This example is one of the more simplistic forms of RCA, but differing models can vary in their complexity.
In 1986, Motorola developed a new strategy called Six Sigma. This was primarily built around enhancing risk management. It employed specific methods and statistics to outline RCA. Six Sigma also refers to the highest quality attainable which roughly translates to about 3.4 errors or defective products per million.
Tools used to perform root cause analysis largely depend on the type and complexity of the issue. The tool used is as critical as the problem definition itself. There is no one-size-fits-all solution when it comes to choosing the tool or technique to be employed. For instance, the “5 Whys” technique is most effective when used to tackle simple to moderate problems that have a limited number of causes. This is partly because the “5 Whys” is a linear approach in one direction, which is not likely to suit problems with multiple causes.
We have put together a brief summary of some of the most commonly used RCA tools.
A Pareto chart is most commonly used when analyzing data about the frequency of problems. It can also be used in prioritizing and identifying the most significant problem when a number of issues exist.
A Pareto chart is a bar chart or a histogram with a line graph that cuts across depicting the frequency of various problems to portray their relative significance. The bars denote the frequency in descending order and the line displays the cumulative total moving from left to right.
The Fishbone or Ishikawa method is mostly used to analyze a problem statement or brainstorm the cause of a problem. It can also be used to scrutinize process and quality improvement.
The fishbone technique portrays a visual method to diagnose the problem. It allows focus on the underlying problem rather than the symptoms that lead to it. Fishbones offer a great way to brainstorm within a well-defined structure.
The Scatter diagram method is ideally a follow up to brainstorming sessions of causes and effects using the fishbone method. It is done to determine objectively whether a particular cause and effect are related. It is primarily used while trying to determine if two variables are related or not.
The scatter plot or scatter diagram technique employs the use of pairs of data pointers to reveal relationships between different variables. It is a quantitative methodology that helps determine the correlation between two variables. Normally, a scatter diagram is created by plotting the independent variable along the x-axis, which is the cause and the dependent variable along the y-axis, which would be the effect. If the pattern depicts a clear line to curve, it signifies that the variables are correlated.