Root Cause Analysis in Manufacturing: Methods, 5 Why, Fishbone and Examples

Introduction

A hydraulic press stops working. The maintenance team replaces the burnt motor, the press restarts, and everyone goes back to work. Three weeks later, the same motor burns out again.

The team fixed the motor. Nobody asked why the motor burned out.

Had they asked, they would have found an overloaded circuit, caused by a blocked filter, caused by a missing filter cleaning schedule, caused by the maintenance plan never being updated after the machine was upgraded two years ago. Replacing the motor treats the symptom. Updating the maintenance plan treats the cause.

That difference is what root cause analysis is about, and it is one of the most valuable skills an engineering student can develop. Every manufacturing company on earth loses money to problems that keep coming back, and every one of them is looking for engineers who can stop that cycle.

What Is Root Cause Analysis in Manufacturing?

Root cause analysis, usually shortened to RCA, is a structured method of investigating a problem to identify its underlying cause, so that corrective action prevents the problem from happening again rather than just repairing the damage.

The word “root” is doing the important work in that definition. Most problems have three layers.

The symptom is what you notice. The machine stopped. The part is out of tolerance. The customer complained.

The immediate cause is the direct technical reason. The motor burned. The tool was worn. The operator used the wrong setting.

The root cause is the system level reason that allowed the immediate cause to exist. No maintenance schedule. No tool life monitoring. No work instruction at the machine.

RCA is not a single tool. It is an approach, supported by several tools such as the 5 Why method, the fishbone diagram, fault tree analysis and 8D.

Simple definition for your exam: Root cause analysis is a systematic problem solving process used to identify the fundamental cause of a defect or failure, so that permanent corrective action can be taken to prevent its recurrence.

Why Root Cause Analysis Is Important in Manufacturing

It stops the same defect from returning month after month, which is where most hidden cost sits.
It reduces scrap, rework, downtime and warranty claims permanently rather than temporarily.
It saves money, because a recurring problem costs far more over a year than one proper investigation.
It is a mandatory requirement in customer complaint responses across the automotive, aerospace, medical and pharmaceutical sectors.
It creates documented evidence for ISO 9001 and IATF 16949 audits, where corrective action without root cause analysis is treated as a finding.
It builds organisational knowledge, since each investigation teaches something about how the process actually behaves.
It shifts the culture from blaming people to improving systems.

There is a phrase used widely in quality work that captures the whole point. If the same problem happens twice, the first corrective action was not a corrective action. It was a repair.

Symptom vs Immediate Cause vs Root Cause

Students lose marks on this distinction constantly, so it is worth a table.

Layer	What it is	Example on a machining line	Action it leads to
Symptom	What is observed	Shaft diameter out of tolerance	Sort and rework the batch
Immediate cause	The direct technical reason	Cutting tool badly worn	Replace the tool
Root cause	The system reason it was allowed	No tool life monitoring or change frequency defined	Define and standardise tool change interval

Fixing only the symptom is called containment, and it is necessary but temporary. Fixing the root cause is called corrective action, and it is permanent.

Both are needed. Containment protects the customer today. Corrective action protects the company tomorrow.

The Root Cause Analysis Process Step by Step

Real RCA follows a defined sequence. Skipping steps is the main reason investigations fail.

Step 1: Define the problem clearly. Write what happened, where, when, how often and how big it is, using numbers. “Quality issue in machining” is not a problem statement. “12 percent of outer rings on grinder 3 show surface waviness above 2 microns since 14 July” is a problem statement.

Step 2: Contain the problem immediately. Protect the customer first. Sort stock, hold suspect batches, add temporary inspection. Containment is not the solution, but skipping it is unacceptable.

Step 3: Collect data and evidence. Gather measurements, defect counts, machine parameters, material batch records, shift details and photographs. Go to the gemba and look at the actual process rather than relying on reports.

Step 4: Identify possible causes. Use a fishbone diagram with a cross functional team to list every plausible cause under the 6M categories.

Step 5: Narrow down and verify. Use Pareto analysis, stratification and scatter diagrams to test which suspected causes the data actually supports. A cause is not confirmed until evidence links it to the effect.

Step 6: Identify the root cause. Apply the 5 Why method to the verified cause and keep going until you reach a system level failure that you can actually control.

Step 7: Develop and implement corrective action. Choose an action that removes the root cause, not one that adds another inspection step. Prefer mistake proofing over reminders.

Step 8: Verify effectiveness. Measure the process after the change. If the defect rate has not moved, the root cause was wrong and the investigation reopens.

Step 9: Standardise and prevent recurrence. Update the standard operating procedure, control plan, FMEA, checklists and training. Then check whether the same root cause exists on other machines and lines. This step is called horizontal deployment and it is where the real value multiplies.

Step 10: Document and close. Record the investigation, the evidence and the results, so the knowledge survives after the team disperses.

The 5 Why Method Explained with Examples

The 5 Why method is the simplest and most widely used root cause tool in manufacturing. It was developed at Toyota by Sakichi Toyoda and became a core part of the Toyota Production System.

The method is exactly what the name says. Ask why the problem happened, then ask why that answer happened, and continue until you reach a cause you can actually control.

Classic Example

Problem: the machine stopped suddenly.

Why did the machine stop? The circuit overloaded and the fuse blew. Why did the circuit overload? The bearing seized and the motor drew excess current. Why did the bearing seize?

It was not receiving enough lubricant. Why was it not receiving enough lubricant? The lubrication pump was not circulating properly. Why was the pump not working properly? Metal chips had entered it because there was no filter fitted.

Root cause: no filter on the lubrication pump. Corrective action: fit a filter and add it to the preventive maintenance schedule.

Notice what happens if you stop at the second why. You replace the fuse, the machine runs, and it fails again next month.

Manufacturing Quality Example

Problem: customer rejected a batch of brackets for missing chamfer.

Why was the chamfer missing? The operation was skipped on 40 parts. Why was it skipped? The operator was not aware chamfering was required on this variant. Why was he not aware? The work instruction at the machine showed the old revision of the drawing.

Why was the old revision at the machine? The document control update after the design change was never issued to the shop floor. Why was it never issued? There is no defined process linking engineering change notices to shop floor document updates.

Root cause: no link between the engineering change process and shop floor document control. Corrective action: define and standardise that link, and audit it.

Notice how the answer moves away from the operator with each why. That is the sign of a correct 5 Why. If your chain ends at “the operator was careless”, you have stopped too early, because the next question is why the process allowed a careless moment to reach the customer.

Rules for Using 5 Why Correctly

Five is a guideline, not a rule. Some problems need three whys, others need seven.
Every answer must be supported by evidence, not assumption.
Stop when you reach a cause you can control and whose removal prevents recurrence.
Never accept human error as a final root cause. Ask why the system permitted it.
Test the chain backwards using “therefore”. Read from the root cause upward and check that each step logically produces the next one.
A problem may have more than one branch, so run separate chains where the causes split.

Fishbone Diagram in Root Cause Analysis

The cause and effect diagram, also called the fishbone or Ishikawa diagram, is used to organise all possible causes before narrowing down.

The problem is written at the head of the fish, a spine runs to it, and diagonal bones carry cause categories. In manufacturing the standard categories are the 6M.

Man, meaning skill, training, awareness and fatigue. Machine, meaning equipment condition, settings, wear and calibration. Material, meaning raw material properties, supplier variation and storage. Method, meaning the process, sequence, parameters and work instructions. Measurement, meaning gauges, calibration, inspection method and gauge repeatability. Mother Nature, meaning environment such as temperature, humidity, dust and vibration.

Fishbone or 5 Why, which one first? Use the fishbone to open the investigation wide and capture every possibility from the whole team. Then use 5 Why to drill deep into the specific causes the data supports. Fishbone is for breadth, 5 Why is for depth. Most strong investigations use both, in that order.

Other Root Cause Analysis Methods

Pareto analysis ranks defect types by frequency so the investigation starts with the problem that matters most, rather than the one that shouts loudest.

Fault Tree Analysis (FTA) works top down from an undesired event, using logic gates to map how combinations of failures could produce it. It is used heavily in safety critical industries such as aerospace and nuclear.

FMEA, meaning Failure Mode and Effects Analysis, is preventive rather than reactive. It predicts how a process or product could fail, scores each failure mode by severity, occurrence and detection, and prioritises using the risk priority number.

8D, meaning the Eight Disciplines, is the standard structured problem solving format used in automotive customer complaint responses.

The disciplines are: form the team, describe the problem, implement interim containment, define and verify root cause, choose permanent corrective action, implement and validate it, prevent recurrence, and recognise the team. If you plan to work in the automotive supply chain, learn 8D by name, because you will fill out these reports.

A3 problem solving presents the whole investigation on a single A3 sheet, moving from background and current state through analysis, countermeasures and follow up. It is a Toyota format valued for forcing clarity.

Change analysis compares the situation before and after the problem appeared, asking what changed in material, machine, method, people or environment. It is extremely effective when a stable process suddenly starts producing defects.

Scatter diagrams and regression verify whether a suspected cause statistically relates to the effect.

Comparison of Root Cause Analysis Tools

Tool	Best used for	Complexity	Typical users
5 Why	Going deep quickly on a single clear problem	Very low	Operators and engineers
Fishbone diagram	Listing all possible causes with a team	Low	Cross functional teams
Pareto analysis	Deciding which problem to investigate first	Low	Quality engineers
Change analysis	Sudden problems in a previously stable process	Low	Production and maintenance
Fault tree analysis	Complex or safety critical failures	High	Design and reliability engineers
FMEA	Preventing failures before they occur	Medium	Design and process engineers
8D report	Formal customer complaint response	Medium	Supplier quality teams
A3 problem solving	Structured summary of a full investigation	Medium	Lean and improvement teams

A Complete Root Cause Analysis Example from a Factory

A bearing manufacturer received a customer complaint for noisy bearings, with a rejection rate of 4 percent on one product code.

Define. The problem was stated with numbers: 4 percent noise rejection on a single product code, starting in the second week of March, all from one assembly line.

Contain. All stock of that code was held and 100 percent noise tested before dispatch, protecting the customer while the investigation ran.

Collect data. Records showed the defect appeared only on parts assembled on the night shift, and only after a raw material batch change in early March.

Identify possible causes. The team built a fishbone diagram. Under Material they listed the new steel batch. Under Machine they listed grinding wheel condition and spindle vibration. Under Method they listed the wheel dressing interval. Under Measurement they listed the noise tester calibration. Under Man they listed the night shift setup practice.

Narrow down. Stratification by shift and by material batch showed the defect correlated far more strongly with the wheel dressing interval than with the material batch. A scatter diagram of dressing interval against noise level confirmed a clear positive relationship.

Find the root cause. Applying 5 Why: bearings were noisy because raceway surface waviness was high, because the grinding wheel was glazed, because dressing was done less frequently on night shift, because the dressing interval was left to operator judgement, because the control plan specified no fixed dressing frequency.

Root cause: no defined wheel dressing frequency in the control plan.

Corrective action. A fixed dressing interval was defined, added to the control plan and standard work, and an automatic dressing counter with an alarm was fitted to the machine.

Verify. Over the following four weeks, noise rejection fell from 4 percent to 0.2 percent.

Standardise and deploy. The control plan, work instruction and process FMEA were updated, operators on all shifts were trained, and the same check was applied to the other three grinding lines, where two were found to have the same gap.

Notice that the final action was not “instruct the operator to dress more often”. It was a system change plus a device that makes the correct behaviour automatic.

Common Mistakes in Root Cause Analysis

Stopping at human error. “Operator mistake” is almost never a root cause. It is the point where the real investigation begins.
Blaming people instead of examining the process. The moment blame enters the room, honest information stops flowing.
Confusing containment with corrective action. Sorting the batch protects the customer but changes nothing about the process.
Adding inspection as the corrective action. More inspection detects the defect. It does not prevent it, and it adds permanent cost.
Guessing without evidence. A cause is a hypothesis until data confirms it.
Investigating alone. The person who knows why a machine behaves oddly is usually the operator, not the engineer.
Ignoring the possibility of multiple root causes. Some problems have two or three contributing causes acting together.
Skipping verification. If you never measure after the fix, you never know whether the fix worked.
Forgetting horizontal deployment. The same root cause usually exists somewhere else in the plant.
Closing the report before the evidence supports it, which is the most common cause of the same complaint returning three months later.

Where RCA Fits with Six Sigma, TQM and Lean

Root cause analysis is not a competing system. It is the analytical core shared by all of them.

In Six Sigma, RCA is essentially the Analyse phase of DMAIC, supported by the same tools of fishbone, Pareto, scatter and hypothesis testing.

In Kaizen and lean, the 5 Why method is the everyday habit that turns small problems into small permanent improvements.

In TQM, RCA supplies the fact based decision making principle, and provides the evidence trail for corrective and preventive action.

In ISO 9001 and IATF 16949, documented root cause analysis is a formal requirement of the corrective action process, and auditors specifically check whether the recorded root cause actually explains the nonconformity.

Frequently Asked Questions (FAQs) on Root Cause Analysis

1. What is root cause analysis in manufacturing?

It is a structured method of finding the underlying reason behind a defect or failure.

The aim is to stop the problem from happening again, not just to repair the immediate damage.

2. What is the 5 Why method?

It is a technique of asking why repeatedly, usually about five times, until the underlying system level cause is reached.

It was developed by Sakichi Toyoda and used widely in the Toyota Production System.

3. Do you always have to ask exactly five whys?

No.

Five is a guideline. Some problems reach the root cause in three whys and others need seven or more.

4. What is the difference between a symptom and a root cause?

A symptom is what you observe, such as an out of tolerance part.

A root cause is the system level reason that allowed it, such as a missing tool change standard.

5. What is the difference between containment and corrective action?

Containment is a temporary action to protect the customer, such as sorting or extra inspection.

Corrective action is a permanent change that removes the root cause.

6. Which is better, 5 Why or the fishbone diagram?

They serve different purposes.

The fishbone opens the investigation wide and lists all possible causes, while 5 Why drills deep into a specific one.

Most good investigations use the fishbone first and 5 Why afterwards.

7. What are the 6M categories in a fishbone diagram?

Man, Machine, Material, Method, Measurement and Mother Nature, meaning environment.

8. What is an 8D report?

It is a structured eight step problem solving format widely used in the automotive industry for responding to customer complaints.

It covers team formation, problem description, containment, root cause, corrective action, implementation, prevention of recurrence and team recognition.

9. Can operator error be a root cause?

Almost never.

If the chain ends at operator error, the next question is why the process allowed that error to occur and to reach the customer.

10. How do you know the root cause is correct?

By verification.

If removing the identified cause makes the problem stop, and reinstating the condition would bring it back, the root cause is confirmed.

Conclusion

Root cause analysis is the difference between a factory that solves a problem once and a factory that keeps solving the same problem forever.

For your exams, hold three anchors. The distinction between symptom, immediate cause and root cause. The 5 Why method with a worked example you can reproduce from memory. And the 6M categories of the fishbone diagram, along with when to use breadth versus depth.

For your interviews, prepare one real investigation you can walk through. State the problem with a number, describe the data you collected, show the why chain, name the corrective action and give the result. Interviewers in manufacturing listen specifically for whether your chain ends at a system fix or at blaming a person, because that single detail tells them how you will behave on their shop floor.

For your career, build the habit early. The next time something goes wrong in a lab, a project or an internship, resist the urge to fix it and move on. Ask why once more than feels necessary. That one extra question is the entire skill.