ROOT CAUSE ANALYSIS
“Nothing is at it seems.”
― Pearl Jam
Time for root cause analysis, which incorporates prioritization, hypotheses, hypothesis trees, disaggregation, deductive and inductive logic, and questions.
What is a root cause?
When there are issues such as decreasing sales, increasing costs, poor quality, inefficiencies, or poor morale you have to understand the root cause of the issue. Only by addressing the root cause can a problem be fixed. Time and again, organizations don’t put the necessary resources in understanding and addressing root causes. And in doing so, they often misdiagnose situations and apply the wrong antidote, often exacerbating the issues.
You often have to do a lot of digging to get to the root of an issue. Usually, when you examine a problem, what you typically initially see are the symptoms of a problem, and only by digging into the actual causes of the symptoms can you find the root cause.
A great example is uncovering the root cause of an issue we had at a product company on returns. Here is how the deductive inquiry went over a few days:
• Question: “Why are our margins decreasing?”
• Answer: “The main culprit is increasing return costs.”
• Question: “Why are our return costs increasing so much?”
• Answer: “Looks like a high return rate of our larger portable rechargers.”
• Question: “Why do we have a high return rate of our larger portable rechargers?”
• Answer: “The batteries only have a shelf life of 3 to 6 months, and customers aren’t recharging them.”
• Question: “Why do the batteries only have a shelf life of 3 to 6 months?”
• Answer: “Well the batteries actually have a shelf life of 1 to 2 years, but the circuitry on the rechargers is drawing down the batteries during storage.”
• Question: “Why does the circuitry on the recharges decrease the shelf life 50-80%?
• Answer: It looks like a poor design, and we can probably redesign the circuitry and the firmware to increase the shelf life to over 1 year.
Once we got down to the root cause that our margins were decreasing because of a poor design in our circuitry and firmware, we could actually fix the problem. Any other solution would only be a temporary Band-Aid to the problem. And, trust me, we had a lot of Band-Aids to the problem and a ton of downstream costs, in returns, customer defection, lost sales due to short shelf life, obsolete inventory, and costs to replace batteries.
How do you diagnose root causes?
You know when you finally get down to a root cause when there is no other explanation for the issue or situation. Many of the tools in this problem solving module are helpful for root cause analysis. Root cause analysis necessitates framing the problem correctly through problem statements, hypothesis trees, disaggregation, and then using deductive and inductive logic, prioritization and questions. Analytical root cause analysis necessitates analytics to baseline a situation, prove or disprove hypotheses, and determine causal factors through statistics.
Some of the best practices for root cause analysis include:
- Ask 5 Whys. As in the example above, one of the best tools is simply using iterative Why questions, until you get to the root issue. You can always iterate a sixth or seventh Why, but typically 5 Whys is all you need.
- Utilize the problem solving toolkit. Often, root causes aren’t singular, but instead multi-faceted. In these instances, you can use the use the problem solving tools we’ve outlined.
- Conduct a Pareto analysis. Pareto analysis is one of the analytic tools to assist root cause analysis. Pareto analysis is rank-ordering issues by their magnitude. In the case of the Goal Zero example, we found that over 80% of our return costs were confined to less than 20% of our product SKUs. You’ll see this 80/20 phenomenon all across businesses.
- Utilize lean and six sigma. Lean is a great methodology for understanding the bottlenecks and inefficiencies in a process. While six sigma, especially the DMAIC process (Define, Measure, Analyze, Improve, and Control) is a strong statistical-based methodology to drive to root causes.