Root Cause Analysis

Root Cause Analysis Methods: 5 Why, Fishbone, Apollo, and Fault Tree Compared

12 min read

Root cause analysis (RCA) is a structured way to find the underlying reason a failure or problem occurred, not just the symptom. The goal is to fix the real cause so the problem does not come back. Different RCA methods suit different situations: some are fast and simple, others are rigorous and evidence-based. This guide compares the four most common methods and when to use each.

What Is Root Cause Analysis and Why Does It Matter?

Root cause analysis is a process for finding the fundamental cause of a failure, defect, or incident. Fixing symptoms gives temporary relief. Fixing the root cause prevents recurrence.

RCA matters because:

Repeat failures cost money. Unplanned downtime, scrap, rework, and safety incidents add up. One study found that unplanned downtime in manufacturing costs $50 billion annually in the U.S. alone.
Band-aid fixes waste effort. Replacing a failed bearing without fixing the misalignment that caused it leads to another failure in months.
Regulators and auditors expect it. In pharma, aerospace, and other regulated industries, RCA is often required after significant deviations or incidents.

A good RCA answers: What happened? Why did it happen? What will we do to prevent it from happening again? The output is not just a report. It is a set of actions with owners and due dates. Without follow-through, RCA is an academic exercise. With it, RCA drives real improvement in reliability, safety, and cost.

5 Why Analysis

5 Why is the simplest RCA method. You start with the problem and ask “Why?” repeatedly until you reach a root cause. The name suggests five questions, but you may need more or fewer depending on the problem.

How It Works

State the problem clearly.
Ask “Why did this happen?” and write the answer.
Ask “Why did that happen?” about the answer.
Repeat until you reach a cause that you can control or fix.
Verify that stopping at that cause would have prevented the problem.

Example

Problem: Pump failed at 3 a.m.

Why? The bearing overheated.
Why? Lubrication was insufficient.
Why? The grease fitting was clogged.
Why? Grease was not purged during the last PM.
Why? The PM checklist did not include a purge step.

Root cause: PM procedure gap. Fix: Add purge step to checklist and train technicians.

Pros and Cons

Pros	Cons
Fast: 15-30 minutes for simple problems	Can stop too early or follow wrong path
No special tools or training	Single chain can miss contributing causes
Easy for anyone to use	“Why” can become subjective or blame-focused
Good for straightforward, linear causes	Weak for complex, multi-factor failures

Best for: Simple, single-cause problems. Quick triage. Team alignment on a shared understanding. Daily production issues where speed matters more than formality.

Fishbone (Ishikawa) Diagram

The Fishbone diagram, also called an Ishikawa or cause-and-effect diagram, organizes possible causes into categories. The problem sits at the “head” of the fish; categories form the “bones.” The classic categories are the 6Ms: Man, Machine, Material, Method, Measurement, and Mother Nature (environment).

How It Works

Define the problem and write it in a box on the right (the head).
Draw a horizontal line (the spine) pointing to the problem.
Add diagonal lines (bones) for each category.
Brainstorm causes within each category.
For each cause, ask “Why?” to drill deeper.
Identify the most likely root causes and verify with data.

Example

Problem: Conveyor belt tears every 2-3 months.

Man: Operator overloads belt; training gaps on load limits
Machine: Worn idlers; misaligned pulleys; sharp edges on transfer points
Material: Belt specification wrong for load; foreign objects in material
Method: No inspection schedule; improper splice procedure
Measurement: No belt wear monitoring; load sensors out of calibration
Mother Nature: Dust, moisture, temperature affecting belt life

The team then investigates the most likely causes (e.g., misaligned pulleys, overload) with data before concluding.

Pros and Cons

Pros	Cons
Visual; good for group brainstorming	Can produce long lists without prioritization
6M structure reduces overlooked areas	Categories can overlap (Man vs Method)
Encourages broad thinking	Still needs data to validate causes
Works well in workshops	Can be time-consuming for simple problems

Best for: Complex problems with many possible causes. Cross-functional teams. When you need to avoid jumping to one cause too quickly.

Apollo Root Cause Analysis

Apollo Root Cause Analysis is an evidence-based method that builds a causal tree. Every cause must be linked to the effect above it with a clear “because” statement. It is more rigorous than 5 Why and less graphical than Fishbone.

How It Works

Define the problem (the top of the tree).
For each effect, ask “What causes this?” and “What evidence supports that?”
Write causes as complete sentences: “X happened because Y.”
Continue until you reach root causes (causes with no further practical causes).
Identify actions that would prevent each root cause.
Assign owners and track implementation.

The key rule: no cause without evidence. If you cannot point to data, a record, or an observation, the cause stays as a hypothesis until verified. This discipline prevents the “we think it was probably X” conclusions that fail under auditor scrutiny. Apollo also distinguishes between root causes (things you can fix) and root cause statements (the full causal chain in one sentence). Both matter for communication and action tracking.

Pros and Cons

Pros	Cons
Evidence-based; reduces speculation	More time and discipline than 5 Why
Clear causal logic; good for audits	Requires training to do well
Handles multiple contributing causes	Can feel bureaucratic for simple issues
Action-oriented; links causes to fixes	Not as visual as Fishbone

Best for: Significant incidents. Regulated industries. When you need a defensible, auditable analysis. Problems with multiple contributing causes.

Fault Tree Analysis

Fault Tree Analysis (FTA) is a top-down, deductive method. You start with an undesired event (the “top event”) and work down using logic gates (AND, OR) to show how combinations of failures lead to that event. It is common in safety-critical industries like aerospace, nuclear, and defense.

How It Works

Define the top event (e.g., “Pump fails to start”).
Identify immediate causes. Use OR gates if any single cause can produce the event; use AND gates if all causes must occur together.
Break each cause into sub-causes.
Continue until you reach basic events (component failures, human errors) or limits of analysis.
Calculate probability if data is available.
Identify critical paths and high-impact basic events.

Pros and Cons

Pros	Cons
Rigorous; supports quantitative risk	Steep learning curve
Good for safety and reliability modeling	Time-consuming to build
Handles complex failure combinations	Requires failure rate data for quantification
Standard in high-reliability industries	Overkill for simple, one-off failures

Best for: Safety-critical systems. Design-phase risk analysis. When you need to model combinations of failures (e.g., redundancy, common cause). Industries like aerospace, nuclear, defense. FTA is also used in reliability block diagram (RBD) analysis to understand system availability and identify single points of failure. The output can feed into quantitative risk assessment and support design decisions about redundancy and maintenance strategy.

Comparison Table: Which Method When?

Method	Complexity	Time Required	Best For	Typical Industries	Tools Needed
5 Why	Low	15-30 min	Simple, linear problems	Any	None
Fishbone	Medium	1-2 hours	Multi-cause brainstorming	Manufacturing, quality, operations	Whiteboard or diagram tool
Apollo RCA	Medium-High	2-8 hours	Significant incidents, audits	Pharma, chemicals, regulated	Template or software
Fault Tree	High	Days to weeks	Safety-critical, design risk	Aerospace, nuclear, defense	FTA software

How to Choose the Right RCA Method

Use these criteria to pick a method:

Problem complexity: Simple, single-cause issues fit 5 Why. Multi-factor or systemic issues fit Fishbone or Apollo.

Stakes: Low-stakes, internal problems can use 5 Why. Regulated or high-consequence incidents need Apollo or FTA.

Time available: Quick triage favors 5 Why. Deep dives need Apollo or FTA.

Audience: Executives and operators often prefer 5 Why or Fishbone. Auditors and regulators expect Apollo-style evidence and logic.

Industry norms: Aerospace and defense use FTA. Pharma and chemicals often use Apollo or similar evidence-based methods. General manufacturing uses all four depending on the situation.

When in doubt, start with 5 Why. If the answer feels incomplete or the problem recurs, escalate to Fishbone or Apollo. Some organizations define escalation rules: for example, any incident above a certain cost or safety threshold automatically triggers Apollo. That ensures consistency and prevents under-investigation of serious events.

Common Mistakes in Root Cause Analysis

These mistakes undermine RCA effectiveness:

Stopping too early: The first plausible cause is often a symptom. Keep asking “Why?” until you reach something you can change. If the cause is “human error,” ask why the human made that error (procedure? training? design?).

Confusing symptoms with causes: “The pump failed” is a symptom. “Bearing failed due to misalignment due to inadequate base inspection” gets closer to cause. Symptoms describe what happened; causes explain why.

Blame culture: RCA works when the goal is learning, not punishment. If people fear blame, they hide information. Frame RCA as “What can we fix?” not “Who messed up?”

No verification: A root cause is only valid if fixing it would have prevented the problem. Test your logic: if we had done X, would this have been avoided?

Skipping action tracking: Finding the root cause means nothing without actions, owners, and due dates. Close the loop. Many teams produce a thorough RCA report and then file it away. The real value comes from implementing corrective and preventive actions, verifying they work, and updating procedures, training, or design accordingly. Track actions in a simple register and review progress in regular reliability meetings.

How AI Enhances Root Cause Analysis

AI does not replace RCA methods. It augments them.

Pattern detection across incidents: AI can scan hundreds or thousands of work orders and incident reports. It finds recurring failure patterns, common causes, and bad actors that a single RCA might miss. “This pump has failed 12 times in 18 months with similar symptoms” becomes visible.

Automated causal chain suggestion: Given a failure description, AI can propose a causal chain based on similar historical incidents. Engineers validate and refine. This speeds up the first draft of a 5 Why or Apollo analysis.

Linking to FMEA and maintenance strategy: AI can map RCA findings to FMEA failure modes and suggest updates to PM tasks, inspection frequencies, or design changes. That closes the loop from incident to prevention.

For more on AI in RCA, see our post on AI-powered root cause analysis.

How Tacit AI Approaches This

Tacit AI’s Intelligent RCA capability helps reliability teams move from isolated incident investigations to pattern-based learning. The platform ingests work orders, maintenance records, and incident data, then surfaces recurring failure patterns and suggests causal chains.

When you run an RCA, Tacit AI can:

Show similar past incidents so you are not starting from scratch
Propose causal chains based on what has happened before on similar equipment
Link findings to your FMEA so failure modes and controls stay current
Track bad actors across sites so you prioritize which assets need deeper analysis

The goal is to make each RCA faster and to connect individual investigations into a broader picture. One pump failure might be a fluke. Ten similar failures across three sites point to a systemic cause that a single 5 Why could miss.

Tacit AI works with your chosen RCA method. You can use 5 Why, Fishbone, Apollo, or FTA in your process; the platform provides the data and pattern detection to support it.

Frequently Asked Questions

When should I use 5 Why vs Fishbone?

Use 5 Why when the problem seems straightforward and you want a quick answer. Use Fishbone when there are many possible causes, you need to involve a cross-functional team, or you want to avoid jumping to one cause too soon.

Is Apollo RCA worth the extra effort?

For significant incidents, regulatory scrutiny, or when you need an auditable record, yes. Apollo forces evidence-based thinking and clear causal logic. For minor, internal issues, 5 Why or Fishbone may be enough.

Can I combine RCA methods?

Yes. A common approach: use Fishbone to brainstorm causes, then use 5 Why or Apollo to drill into the most likely ones. Or use 5 Why for a quick pass, then Apollo for a formal write-up if the incident is serious.

How does AI change RCA?

AI does not replace the method. It speeds up data gathering, finds patterns across many incidents, and suggests causal chains. Engineers still validate, judge, and own the conclusion. AI makes each RCA faster and connects it to the bigger picture.

What if we cannot find the root cause?

Sometimes the cause is unknown or unverifiable. In that case, document what you know, what you tested, and what actions you will take anyway (e.g., add inspections, change procedures). Avoid declaring a root cause without evidence. “Inconclusive” with preventive actions is better than a wrong cause.

Get the Right RCA Method for Your Problem

Choosing the wrong RCA method wastes time. Choosing the right one gets you to the real cause faster.

Book a working session to see how Tacit AI’s Intelligent RCA capability supports pattern detection and causal chain suggestion across your maintenance data.