What Are the 5 Whys?
Written by Jason Westland · Edited by Johanna Barnett
The 5 Whys technique is a focused cause-analysis method in which a team starts with a defined problem, asks why it happened, and tests each answer with evidence before using it as the next step in the chain.
Five is a guide, not a rule. Stop when the supported cause is specific, controllable, evidence-based, and still connected to the original problem; move to fishbone analysis or broader root cause analysis when several cause paths remain open.
5 Whys analysis overview
- 1Define the problem and scope
- 2Ask why the problem occurred
- 3Use the answer as the next question
- 4Check each answer with evidence
- 5Continue, branch, or broaden
- 6Establish the supported occurrence cause
- 7Check the escape or detection weakness
- 8Link the next controlled action
What the 5 Whys method means
The 5 Whys is a focused method for testing one likely cause path. The team starts with a defined problem, asks why it happened, and uses each evidence-checked answer as the next question.
The method does not prove a root cause merely because the team reaches Why 5. The chain remains a possible cause path until evidence supports the conclusion and material alternatives have been ruled out or recorded as still open.
“5 Whys,” “Five Whys,” and “five-why analysis” are common names for the same method. Its purpose is to move past the visible symptom and identify the process or control condition that should be addressed through the appropriate corrective-action record.
History of the 5 Whys
ASQ attributes the method to Sakichi Toyoda, and the technique is commonly associated with Toyota problem solving. The practical lesson is more important than a long history: ask beyond the symptom, test the chain against facts, and stop when evidence supports a controllable cause.
Use of 5 Whys in manufacturing quality
In manufacturing quality, 5 Whys usually sits inside a wider record. It tests one focused possible-cause path; it does not control product disposition, corrective-action implementation, or effectiveness by itself.
| Record or method | Main job | How 5 Whys fits |
|---|---|---|
| Non-conformance report (NCR) | Controls the nonconformance and product decision. | Use 5 Whys after immediate product control when one focused possible-cause path needs to be tested. |
| Root cause analysis (RCA) | Tests several possible causes and evidence streams before a conclusion is accepted. | Use 5 Whys as one method inside RCA when one likely path can be followed without oversimplifying the problem. |
| 8D report | Structures a D1–D8 response covering containment, cause, corrective action, effectiveness, prevention, and closure. | Use it where appropriate during D4; D5–D7 and the linked action records own action selection, implementation, effectiveness, and prevention. |
| Corrective action | Changes the process or control associated with the supported cause. | The why chain records the evidence-supported cause path; the corrective-action record owns the action, owner, due date, implementation verification, effectiveness review, and closure. |
| Corrective action request (CAR) | Records a corrective-action request issued by a customer or internal quality authority and the formal response required from the organization. | A 5 Whys chain may support the cause explanation, but it does not replace the evidence, approvals, actions, and closure information required by the CAR response. |
| Supplier corrective action request (SCAR) | Requests a formal corrective-action response from a supplier. | A supplier may use a 5 Whys chain to support its cause explanation, but the SCAR response must still include the required evidence, actions, review, and closure information. |
When to use 5 Whys—and when to use a broader method
Use 5 Whys when the problem is specific, the process sequence is understood, and one likely cause path can be tested with evidence.
Use 5 Whys when
- the requirement and actual condition are clearly defined;
- affected and suspect scope is traceable;
- the process sequence is understood;
- one cause path is more likely than the others;
- records, parts, measurements, tests, interviews, or observations can check each answer;
- the issue is not so complex that one chain would hide interacting causes.
Use fishbone or broader RCA instead when
- several plausible cause paths remain open;
- the evidence conflicts;
- affected and conforming conditions differ in several ways;
- design, equipment, material, supplier, or process interactions may combine;
- the issue is high risk, cross-functional, repeated, or disputed;
- the chain begins to branch, skip logic, or depend on assumptions.
Do not force one chain through a multi-cause problem
Before starting a 5 Whys analysis
Do not start by asking why. First agree on the requirement, actual condition, affected and suspect scope, where the issue was found, where it was introduced, and what immediate control is already in place.
- What controlled requirement was not met?
- What was observed, measured, reported, or returned?
- What is confirmed affected?
- What remains suspect?
- Where was the issue found?
- Where was it made or introduced?
- What containment and correction records already exist?
- What evidence is available now?
Keep the assumed cause out of the problem statement. Another reviewer should be able to understand the condition and scope without already knowing the investigation story.
| Entry | Example |
|---|---|
| Weak problem statement | Some housings leaked during test. |
| Stronger problem statement | Incoming test found leakage at the housing seal interface after assembly. The requirement is no leakage during the pressure-hold test, and product made since the last verified good setup remains suspect. |
How to use the 5 Whys method
The method uses each answer as the next question, but the chain is valid only while the logic and evidence remain connected to the original problem.
- 1
Define the problem and scope
State the controlled requirement, actual condition, affected and suspect scope, where the issue was found, and the originating NCR, complaint, CAR, SCAR, RCA, or 8D where applicable. Do not include an assumed cause.
- 2
Ask why the problem occurred
The first answer usually describes the visible process condition. Treat it as a possible cause path, not the final cause.
- 3
Use the answer as the next question
Ask why the previous answer was true. If the next question does not follow logically, the chain is drifting or skipping a step.
- 4
Test the answer with evidence
Review the part, record, drawing, work instruction, measurement, test, observation, interview, maintenance history, calibration record, or supplier information that can support or rule out the answer.
- 5
Check that the chain remains connected
Compare affected and conforming conditions. Each answer must explain the answer before it and remain consistent with the original problem and scope.
- 6
Continue, branch, or broaden
Continue while one evidence-supported path remains credible. If the path branches or evidence conflicts, record the open paths and move to fishbone analysis or broader RCA rather than forcing one story.
- 7
Establish the supported occurrence cause
Stop the occurrence chain when evidence supports a specific, controllable process or control condition and material alternatives have been ruled out or recorded as still open. Do not stop at a person, “training issue,” or Why 5.
- 8
Check detection and link the next action
For a downstream or customer escape, run a separate chain for the escape or detection weakness. Record the supported conclusions and link them to the record that owns corrective action, implementation verification, effectiveness review, and closure.
When to stop asking why
Stop when evidence supports a specific, controllable cause that remains connected to the problem—not when the team reaches the fifth question.
Stop when
- the evidence supports the cause path;
- the answer explains the affected condition and distinguishes it from conforming conditions;
- the cause names a process or control condition;
- the cause is within the control of the responsible process owner or supplier;
- material alternative paths are ruled out or recorded as still open;
- the conclusion provides a credible basis for corrective action.
Do not stop when
- the team has merely reached Why 5;
- the answer is “operator error,” “training issue,” or “inspection missed it”;
- evidence is missing or conflicting;
- the chain skips a logical step;
- several credible paths remain open;
- the corrective action was selected before the cause was supported.
Stop rule
How to test each why answer with evidence
A logical chain is not evidence. Record what was reviewed and classify each material answer as supported, ruled out, or still open.
| Result | What it means | Required next step |
|---|---|---|
| Supported | The reviewed evidence is consistent with the answer, explains the affected condition, and is not contradicted by material facts. | Continue to the next why or use the answer as the supported conclusion when the stop criteria are met. |
| Ruled out | The reviewed evidence contradicts the answer or shows that it does not distinguish affected from conforming conditions. | Close that cause path and record the evidence used. |
| Still open | The evidence is missing, incomplete, or conflicting. | Assign the next evidence check and do not use the answer as the conclusion. |
Evidence to review
- failed or returned product and photographs;
- inspection, measurement, and test results;
- drawings, specifications, work instructions, and control plans;
- lot, batch, work-order, and shipment records;
- setup, maintenance, and calibration records;
- supplier or customer information;
- interviews and direct process observation.
“No evidence found” is not the same as “ruled out”
Occurrence cause and escape or detection weakness
When an issue reached a downstream process or customer, one why chain is usually incomplete. Ask separately why the nonconformance occurred and why the expected control did not detect it.
Supported occurrence cause
What process or control condition allowed the nonconformance to occur?
Follow the path through the method, setup, material, equipment, document, maintenance, measurement, or process control that created the condition.
Escape or detection weakness
What control should have detected the condition, and why did it not?
Review the inspection characteristic, test coverage, sampling plan, checklist, gauge use, review step, and release decision.
The escape point is where the issue should have been detected. The detection weakness explains why that control did not work as required.
Do not merge the two chains
WORKED MANUFACTURING EXAMPLE
Leaking HSG-88 housing customer return: a worked 5 Whys example
This example shows one focused occurrence chain, a separate detection chain, the evidence beside each answer, and the stop decision before corrective action.
Example records are fictional demonstration data.
| Summary item | Detail |
|---|---|
| Linked records | NCR-2026-120 / 8D-2026-042; linked corrective action ACT-3012; linked corrective action request CAR-2026-031. |
| Customer, part, and owner | Apex Industrial; HSG-88 housing; Tom Alvarez. |
| Lot and work order | HSG88-LOT-0726A; WO-2026-7741. |
| Problem statement | Apex Industrial returned HSG-88 housings from HSG88-LOT-0726A that leaked at the seal interface during incoming test. Requirement: housing must seal at the mating face without leakage during the pressure hold test. |
| Affected and suspect scope | 12 returned / 48 suspect. |
| Containment | Hold suspect HSG-88 stock and related WIP in the red-tag hold area; perform 100% seal-face visual check and leak test before release. |
| Correction and disposition | Affected returned housings were handled under the approved NCR-2026-120 disposition and kept separate from the cause analysis and corrective action. |
Occurrence chain — why did the leak occur?
| Step | Why question | Answer | Evidence reviewed | Decision |
|---|---|---|---|---|
| Why 1 | Why did HSG-88 housings leak at the seal interface? | A seal was damaged or mis-seated during assembly. | EVD-120-01 customer return photos; EVD-120-02 leak test record; EVD-120-03 seal inspection photos. | Supported path—continue asking why. |
| Why 2 | Why could a damaged or mis-seated seal be installed? | Seal orientation was not confirmed after the cleaning step. | EVD-120-03 seal inspection photos; EVD-120-04 WI-ASSY-221 Rev C gap review. | Supported path—continue asking why. |
| Why 3 | Why was seal orientation not confirmed after cleaning? | WI-ASSY-221 Rev C did not require a seal-orientation confirmation after the cleaning step, so a damaged or mis-seated seal could be installed before the pressure test. | EVD-120-04 WI-ASSY-221 Rev C gap review. | Supported occurrence cause—stop the occurrence chain. |
Supported occurrence cause
The team stops at Why 3 because the conclusion is evidence-supported, specific, controllable, and still connected to the leak. Forcing Why 4 and Why 5 would add length without improving the decision.
Detection chain — why did the issue reach the customer?
| Step | Why question | Answer | Evidence reviewed | Decision |
|---|---|---|---|---|
| Detection Why 1 | Why did the seal-face condition reach the customer without being detected? | The process relied on final pressure testing and did not include a seal-face visual check before packing. | EVD-120-02 leak test record; EVD-120-03 seal inspection photos; EVD-120-05 IP-HSG-88 Rev B review. | Supported path—continue asking why. |
| Detection Why 2 | Why was a seal-face visual check not required before packing? | IP-HSG-88 Rev B required final pressure testing but did not require a seal-face visual check before packing, so the team relied on the final test alone. | EVD-120-05 IP-HSG-88 Rev B review. | Supported detection weakness—stop the detection chain. |
Supported detection weakness
Linked corrective action
- update WI-ASSY-221 to Rev D and require seal-orientation confirmation after cleaning;
- update TC-SEAL-04 to Rev B and add seal-face visual checking before packing;
- train assembly and test operators on the revised controls;
- review similar seal-assembly controls for the same gap.
Implementation verification
Confirm that WI-ASSY-221 Rev D and TC-SEAL-04 Rev B are issued, obsolete versions are removed, the revised checks are in use, the EVD-120-06 retraining record is complete, and the first three batches after release are reviewed to confirm the revised controls are being used.
Effectiveness target
No repeat leaks in 90 shipped units during the defined review window.
5 Whys handoff
The why chains support the D4 cause decision. ACT-3012 and CAR-2026-031 remain the controlling records for action ownership, implementation verification, effectiveness review, and closure.
Strong and weak why answers
A strong answer names a reviewable process condition and identifies the evidence behind it. A weak answer names a person, broad category, or unsupported conclusion.
| Weak answer | Why it is weak | Stronger answer | Evidence |
|---|---|---|---|
| The operator installed the seal incorrectly. | It names a person-level event without identifying the control that allowed it. | WI-ASSY-221 Rev C did not require seal-orientation confirmation after cleaning. | EVD-120-04 WI-ASSY-221 Rev C gap review. |
| Final test missed it. | It names the escape symptom without identifying the missing detection control. | IP-HSG-88 Rev B required final pressure testing but did not require a seal-face visual check before packing. | EVD-120-05 IP-HSG-88 Rev B review. |
| Training issue. | It is broad, difficult to test, and preselects training as the action. | The released work instruction omitted the seal-orientation confirmation after cleaning. | EVD-120-04 WI-ASSY-221 Rev C gap review. |
| Why 3 is the root cause. | The number of questions is not evidence. | The supported occurrence cause is the missing seal-orientation requirement in WI-ASSY-221 Rev C. | EVD-120-03 seal inspection photos and EVD-120-04 WI-ASSY-221 Rev C gap review. |
Common 5 Whys mistakes
| Mistake | Better practice |
|---|---|
| Starting with the assumed cause | Define the requirement, actual condition, and affected and suspect scope before asking why. |
| Forcing exactly five questions | Ask fewer or more questions and stop only when the evidence and decision criteria are met. |
| Stopping at operator error or training issue | Identify the process or control condition that allowed the event and the detection weakness that allowed it to pass. |
| Building a logical story without evidence | Record what supports, rules out, or leaves each material answer open. |
| Using “no evidence found” as ruled out | Rule out a cause only when reviewed evidence contradicts it or shows that it does not explain the affected condition. |
| Forcing one chain through several possible causes | Move to fishbone analysis or broader RCA when the path branches or evidence conflicts. |
| Mixing occurrence and detection | Use separate chains for why the nonconformance occurred and why the expected control did not detect it. |
| Choosing corrective action before the cause is supported | Complete the evidence checks before selecting the process or control change. |
A tidy chain without evidence is a story, not an investigation.
References and further reading
Use these references as background, then apply your own company, customer, product, process, supplier, and industry requirements.
ASQ — Five Whys and Five Hows
Definition, intended use, method steps, and the point that fewer or more than five questions may be required.
Toyota — Toyota Production System
Official background on Toyota’s production-system philosophy, abnormality control, continuous improvement, and prevention of recurrence.
BMJ Quality & Safety — The problem with “5 whys”
A critical discussion of limitations such as forcing one causal pathway or one root cause onto a complex problem.
This guide follows established manufacturing quality frameworks (ISO 9001, AIAG, ASQ definitions) and is reviewed internally before publishing.
This is provided for general manufacturing quality use. Adapt it to your company, customer, product, process, supplier, and industry requirements before use. It is not intended for medical device, pharma, FDA, Part 11, or validated regulated workflows.
Use the 5 Whys template
Use the matching template page when the team needs a controlled worksheet for the problem statement, occurrence chain, escape or detection chain, evidence beside each answer, supported cause, linked action, and follow-up.