Improving '5 Whys' root cause analysis with the Many Whys

Abstract

Root Cause Analysis (RCA) is a tool used in problem-solving and decision-making processes. This paper examines an evolutionary approach to RCA, referred to as “The Many Whys,". We focus on the methodology of this new approach, and its advantages over an existing method, “The 5 Whys”

Introduction

In today's complex and fast-paced world, effectively identifying and addressing the root causes of problems is a crucial aspect of decision-making processes. Root Cause Analysis (RCA) has traditionally been a valuable tool in this regard. However, conventional RCA methods often fall short when dealing with multifaceted issues, leaving room for improved problem-solving approaches.

Background on root causes analysis

A root cause analysis attempts to find the underlying cause of a problem, that if resolved would be the most effective solution the problem or issue. It avoids immediately attempting to solve the problem and digs deeper.

Effective problem-solving tends to follow a lorentzian distribution, such that resolving proximiate causes tend to be ineffective at solving the problem. Whereas, accurate identification of the root cause tends to be very effective at resolving the problem. Finally, distant causes whilst they objectively may be the ultimate cause as to why a problem exists - such as the fact that the big-bang occured - are ineffective in resolving that problem.

Problem

The underlying problem is the limitation of traditional RCA methods, such as the "Five Whys", in addressing complex issues with multiple causes and diverse branching problems. These methods tend to oversimplify problems and often assume linear and singular causality, leading to incomplete understanding and ineffective solutions. This problem impacts a wide range of industries and sectors, affecting decision-makers who rely on RCA to solve critical issues.

The problem with '5 Whys'

The five whys is a technique that asks the question “why” 5 or more times.

    graph TB
    A[Problem: Our client is refusing to pay for leaflets we printed for him] --> B{Why?}
    B -->|The delivery was late, so the leaflets couldn't be used| C{Why?}
    C -->|The job took longer than expected| D{Why?}
    D -->|We ran out of printer ink| E{Why?}
    E -->|The ink was all used on a large last-minute order| F{Why?}
    F -->|Root cause: We didn't have enough ink in stock, and couldn't order new supplies in time| G[Root Cause]

   

Whilst the 5 Whys is easy to implement without supervision it has many cons

• 5 Whys may oversimplify complex problems by often only considering a single root cause
• Assumes linear causality: The 5 Whys method assumes that problems have linear and singular causes, whereas many issues may have multiple contributing factors and complex relationships between them, that may prevent one from reaching the final cause.
• Inadequate exploration of branching-related problems: The 5 Whys method does not inherently accommodate the exploration of branching questions that may arise from initial answers. As a result, critical aspects of the problem may remain hidden, as problems are skipped over, leading to an incomplete understanding of the issue
• The 5 Whys can often stop short of a comprehensive analysis, as stopping at the fifth "why" may not always reveal the true root cause. Even though you don’t need to stop at the fifth why, the naming of the method often constraints. Some problems may require more or fewer "whys" to reach the underlying cause (Dekker, S. (2006). The Field Guide to Understanding Human Error. CRC Press).
• Limitations due to simplistic questioning: The 5 Whys technique relies on asking the simple question "why," which may not always be sufficient to explore complex problems. The technique may require clarification or additional probing questions to yield meaningful insights (Kepner, C. H., & Tregoe, B. B. (1965). The rational manager: A systematic approach to problem solving and decision making. New York: McGraw-Hill).

Proposed solution

This paper introduces the "Many Whys" approach, a evolutionary RCA method developed through consulting practices. This approach enhances the traditional "Five Whys" method by incorporating additional questions into the process.

The "Many Whys" approach allows for a more comprehensive exploration of problem branches and ensures clarity within the team. The process includes asking "Why?" to attempt to identify the root cause. However, the addition of additional questions uncovers potentially undiscovered answers, enabling multiple avenues of investigation.

Evidence from the application of this approach suggests that it facilitates a more profound and broader understanding of problems, especially those with multiple causes. It also promotes team engagement and understanding, thus enhancing the overall effectiveness of the problem-solving process.

How it works

The process

1) Ask “why?” - General why

2) Ask “why does {cause} create {effect}?” (a why chain) - to chain the whys in a cause an effect chain

3) Ask “what else?” zero or more times - to branch the solution

The why chain

The why chain, directly links cause and effect together. It asks why was the cause created in reference to the previous cause. e.g. Our code base keeps getting rewritten (Cp) → People keep leaving the team (C) why does “People leaving the team” cause “Our code base to keep getting rewritten”?

    graph TB
    AP[Apparent Problem]
    P1[Problem 1]
    P1_1[Problem 1.1]
    P1_2[Problem 1.2]
    P1_2_1[Problem 1.2.1]
    P2[Problem 2]
    P2_1[Problem 2.1]
    P2_2[Problem 2.2]
    P2_1_1[Problem 2.1.1]
    RC[Root cause]
    P2_3[Problem 2.3]
    P3[Problem 3]
    P3_1[Problem 3.1]
    P3_2[Problem 3.2]

    AP-->|"Question: Why?"|P1
    P1-->|"Question: Why?"|P1_1
    P1-->|"Question: Why does problem 1 cause apparent problem?"|P1_2
    P1_2-->|"Question: Why does problem 1 cause problem 1.2?"|P1_2_1

    AP-->|"Question: What else?"|P2
    P2-->|"Question: Why?"|P2_1
    P2-->|"Question: Why does problem 2 cause apparent problem?"|P2_2
    P2_2-->|"Question: Why?"|P2_1_1
    P2_2-->|"Question: Why does problem 2.2 cause problem 2?"|RC
    P2-->|"Question: What else?"|P2_3

    AP-->|"Question: What else?"|P3
    P3-->|"Question: Why?"|P3_1
    P3-->|"Question: Why does problem 3 cause apparent problem?"|P3_2

Vodel

Visual model of the Many Whys process + domain