The 4 Types of Root Cause Analysis

A survey of 106 top executives from 91 companies across 17 countries revealed a widespread concern that organizations struggle to properly diagnose problems.?

85% of the executives surveyed admitted their companies were poor at pinpointing core issues. Additionally, 87% agreed this inability to uncover root causes resulted in substantial expenses and inefficiencies.?

The data indicates a clear need for improved root cause analysis skills at the highest levels of business leadership worldwide.

When you define the problem correctly, you’re just half the way through reaching the right solution. However, being not able to solve the problem is most probably because you’re not able to find out the underlying root causes of the problem.

The difference between solved problems and popping-up problems is that in the former, all the root causes of the symptoms have been rooted out.

But, before delving into the details. If you’re not subscribed to my newsletter Smart Models?

Subscribe here to get practical tips on problem solving, decision making, and leadership.

What is Root Cause Analysis (RCA)?

Root cause analysis involves identifying the underlying reasons behind problems in order to recommend targeted solutions.?

According to Harvard Business School professors Joshua Margolis and Anthony Mayo, leaders must scan the landscape around their organization to spot trends, threats and opportunities.?

By working with others to analyze the root causes of issues, leaders can develop effective solutions. Margolis states that leaders need to act as beacons to illuminate potential problems.

Before jumping into the four types of RCA, let's cover important principles to keep in mind when conducting root cause analysis.

Core principles for RCA

There are a few core principles that guide effective root cause analysis, some of which should already be apparent. Not only will these help in the quality of the analysis, these will also help the analyst gain trust and buy-in from stakeholders.

• Take a comprehensive, holistic approach.
• Focus on fixing root causes, not just symptoms.
• Don't neglect short-term relief of symptoms.
• Examine how and why, not who.
• Expect multiple root causes.
• Support claims with evidence.
• Consider prevention and future impact.

4 Types of Root Cause Analysis

In his book, Four Types of Problems, Art Smalley outlined 3 types of root cause analysis.

However, with the rapid movement of AI applications, especially after releasing ChatGPT by OpenAI in December 2022, it seems that the only thing that AI is not able to do till the moment is mimicking people's reasoning!

“Machines’ lack of understanding of causal relations is perhaps the biggest roadblock to giving them human-level intelligence.” Judea Pearl, Turing Award winner and AI pioneer

So, in this article, I’m going to give you a glimpse of what AI can do when it comes to root cause analysis.

Now, let's cover the four types of RCA, one type at a time.

1) Logic-Based

Logic-based RCA uses qualitative methods to find out the?cause-and-effect relationships. Tools like the 5 Whys and fishbone diagrams trace back from the problem through layers of causal factors via deductive logic.?

Asking "why" at each level reveals connections in a logical sequence. Fishbone diagrams visually map multiple contributing factors against an overall effect.?

Logic-based RCA is straightforward to apply but relies on the knowledge and analytical skills of the investigators. Subjective biases can influence the hypothesized causes.

2) One Variable at a Time (OVAT)

OVAT techniques use statistical analysis to isolate factors and quantify their impact on process performance. It looks at one factor at a time to see its effect.?

Statistical process control tracks measurements over time to find abnormal patterns. Capability analysis studies if a process is operating within expected limits.

OVAT is data-driven and provides measurable insights on relationships between inputs and outputs. However, it studies one variable at a time, unable to account for complex interactions between multiple factors.

3) Multiple Variables at a Time (MVAT)

MVAT methods conduct statistical analysis with multiple factors simultaneously.?

Multivariate analysis methods can determine correlations among many variables, while techniques like regression modeling can estimate the relative importance and statistical significance of different causal relationships.?

Design of experiments tests different combinations of factors to determine which have the biggest influence, which is by far the most reliable technique for identifying real root causes.

MVAT provides more sophisticated modeling of real-world complexity compared to OVAT. But it requires large datasets and advanced analytical expertise, and most importantly; time and money.

4) Causal AI

Causal AI represents an emerging approach to root cause analysis using causal reasoning algorithms. Causal AI is the only technology that can reason and make choices like humans do.

It infers causal relationships directly from data, constructing network diagrams that visualize causal connections. Causal AI can uncover hidden variables and mechanisms driving observed statistical patterns.?

It incorporates expert domain knowledge to guide causal discovery and quantify causal effects. As an AI-driven method, causal AI can rapidly analyze massive, complex datasets with hundreds of variables.?

Causal AI provides actionable insights by revealing the latent causal factors underlying business problems. However, performance depends on the quality of data and domain expertise.

By using the proper type of root cause analysis, you can identify the issues behind your organization’s problems, develop a plan to address them, and make impactful changes.

Smart Models just launched, and every single like, comment, and share helps. Thanks for reading!

Mohammad Elshahat