SHARE Knowledge Hub

Five Whys for Incident Investigation

Five Whys helps investigators test what allowed an incident or control failure, but a chain is not proof. Verify each answer, branch when causes interact, choose stronger controls and check them in real work. Keep statutory reporting separate and use the correct Great Britain or Northern Ireland route.

Updated 2026-10-16UK focus

Understanding Five Whys for Incident Investigation

Five Whys helps investigators test what allowed an incident or control failure, but a chain is not proof. Verify each answer, branch when causes interact, choose stronger controls and check them in real work. Keep statutory reporting separate and use the correct Great Britain or Northern Ireland route.

Why this matters at work

Branch when more than one credible condition contributed, when causes interact, or when separate pathways explain occurrence and severity. A vehicle incident may involve layout, visibility, scheduling, route information and supervision. Forcing all those conditions into one line can suggest a false sequence. A timeline, fishbone or barrier review may better organise the wider system before selected branches are examined in depth. Stop a branch that is irrelevant, disproved or unsupported, recording why. Refer technical failures, complex processes and high-consequence uncertainties to suitable specialists. Do not continue asking why until speculation sounds persuasive. Shared conditions should be shown once with clear links rather than counted as multiple independent causes. Keep immediate protection in place while questions remain open. Final findings should explain what the evidence establishes and what remains unknown, with an owner for further work. A method is useful only if it helps choose effective controls; completion of a diagram is not evidence that the hazard has been resolved.

Treat culture as a prompt for observable evidence: conflicting instructions, ignored defect reports, resource decisions, repeated workarounds or incentives that affected a specified barrier. Interview people openly and preserve contrary evidence. Do not assume lack of a signature proves no briefing occurred, or that a signature proves understanding. Set a stopping point when further claims cannot be established and identify the next inquiry. Keep conduct questions separate from fair prevention-focused analysis. The resulting finding should name a change the organisation can implement and verify, not attach blame to a worker or use a vague root-cause label to avoid difficult design or management decisions.

How organisations manage this in practice

Map every final finding to a control purpose and an accountable owner with authority and resources. Prefer suitable changes to design, layout, containment, maintenance or planning before relying on reminders. Training and supervision can support those changes but do not replace physical protection. State interim measures, target dates, dependencies and evidence needed for technical acceptance or completion.

Worker health information requires actual lawful basis and special-category condition, necessary purpose, limited access, security and appropriate retention. Signature or interview participation is not automatic freely given employment consent. ICO guidance is under review after DUAA; clinical referral consent and work controls remain separate.

Map each supported finding to a specific control change, owner and acceptance criterion. Completion is separate from effectiveness. Verify safely under relevant conditions and check extent elsewhere; reopen, correct and safely retest failed protection rather than close from attendance or email delivery.

Let contradictory evidence change the branch instead of preserving a neat story

The conveyor-jam teaching chain is a set of hypotheses. Suppose the first explanation is “the worker had not been trained”, but records and demonstration establish relevant training. Reject that unsupported answer; investigate the recurring jam, access design, isolation method, production demands and previous defect response rather than using the fifth box to reach a predetermined blame conclusion.

A useful causal map can have fewer or more than five questions and more than one branch. Record the evidence for each link, the remaining uncertainty and what would disprove it. Choose upstream actions such as eliminating the recurring intervention or making safe isolation practical, then verify effectiveness at the workface. Retraining may support a control change, but it is not an adequate automatic ending to every chain.

The conveyor records contradict the first answer

Illustrative example. A hand-injury investigation initially blames missing training during jam clearance, but the evidence shows training was delivered and understood.

Decision. The team revises the branch to examine recurring jams, access, isolation and operational pressure.

Evidence to check. Machine condition, jam history, energy-isolation evidence, observation, accounts and training demonstration.

Hold or escalation point. Do not keep a disproved answer because it completes a five-step narrative.

Follow-up. Assign an upstream design or maintenance action and field-test that the hazardous intervention is no longer required.

Test each causal step instead of forcing five answers

Use one defined event throughout the chain. A why question needs a supported explanation, and the evidence may branch or stop short of five levels.

  1. Define the jam event

    Record the actual conveyor task, event and protected state; do not switch unnoticed to a different transport example.

  2. Explain rather than instruct

    A statement such as follow the procedure is an action, not evidence explaining why the event occurred.

  3. Check competing paths

    Compare maintenance, access and work-demand evidence; record contradictions and the observation still needed.

  4. Verify the action

    Match the corrective action to the supported control gap, with an owner and real-task test. Avoid a single blame-based root or an arbitrary five-question stopping rule.

Read the complete task-specific control examples

ERICPD is a planning cross-map: Discipline verifies and sustains every control. ESEAP compares effectiveness. A label does not authorise the task; retain the owner, evidence and stop or change decision.

Five Whys for Incident Investigation · ERICPD

Eliminate
Remove pedestrian movement from the vehicle zone through a redesigned delivery route.
Reduce
Use forward-only access or smaller delivery vehicles where geometry allows.
Isolate
Install robust vehicle-pedestrian separation and controlled crossings.
Control
Use booked deliveries, approved temporary route changes and contractor briefings.
PPE
Use visibility clothing for residual exposure, never as a substitute for separation.
Discipline
Review near-miss trend and field-test the revised traffic plan after changes.

Five Whys for Incident Investigation · ESEAP

Elimination
Remove pedestrian movement from the vehicle zone through a redesigned delivery route.
Substitution
Use forward-only access or smaller delivery vehicles where geometry allows.
Engineering controls
Install robust vehicle-pedestrian separation and controlled crossings.
Administrative controls
Use booked deliveries, approved temporary route changes and contractor briefings.
PPE
Use visibility clothing for residual exposure, never as a substitute for separation.

What are the key points about Five Whys?

Five Whys is a way to ask what allowed an incident or control failure to occur. A five-question chain is a prompt, not proof: each answer needs evidence, competing explanations and a control decision.

It starts with facts

The first problem statement should describe what happened, where, when and with what consequence or potential consequence.

Five is not a rule

Stop when the team reaches a correctable system weakness, not when it has counted five questions.

Branching is normal

Incidents often involve several causal paths: equipment, work method, supervision, environment, competence, change and organisational pressure.

Human error is not the end point

If a person made an error, the next question is why the task, equipment, information or organisation made that error possible.

Evidence must test each answer

Each why should be supported by records, observations, interviews, measurements or technical assessment.

Actions must strengthen controls

The output should change design, planning, supervision, equipment, maintenance, competence, verification or management arrangements.

HSE HSG245 sets out evidence gathering, analysis, risk-control selection and an action plan; it is a 2004 investigation workbook, not a statutory Five Whys rule.

What is Five Whys in incident investigation?

Five Whys is a causal questioning method used to explore why an incident, near miss or control failure happened.

The investigator starts with a neutral event statement and asks why the event occurred. The answer becomes the basis for the next why. This continues until the team reaches a cause that the organisation can control, correct and verify. The method is attractive because it is easy to understand, but that simplicity can be dangerous if it is used without evidence.

In safety investigations, Five Whys should expose conditions rather than personalities. 'The worker was careless' is not a useful final answer. A stronger answer might be that the procedure conflicted with normal work, the isolation point was not labelled, maintenance records were wrong, supervision did not detect drift or the task was authorised without a hold point.

What Five Whys is and is not
Aspect Good use Poor use
Purpose Move from visible event to correctable system causes. Find a quick person to blame.
Questioning Ask why a factual condition existed. Ask why someone lacked common sense.
Evidence Test each answer against records, observations and accounts. Accept assumptions because they sound plausible.
Depth Continue until the cause is controllable and useful. Stop after exactly five questions regardless of quality.
Output Verified findings and stronger controls. A neat chain with weak reminders.
Complexity Branch where several causes contributed. Force one linear chain for a multi-causal event.

Where does Five Whys sit in the investigation process?

Five Whys belongs after immediate response and evidence gathering, and before corrective actions are finalised.

HSE's investigation approach is a learning cycle: gather information, analyse information, identify risk control measures, then implement an action plan. Five Whys sits mainly in the analysis stage, but it depends heavily on the quality of information gathered earlier.

If a team starts asking why before it has preserved evidence, interviewed people or checked records, the chain can become a polished story rather than a reliable analysis. Used well, the method helps test the event sequence and identify which controls were missing, weak, defeated or not verified.

Investigation stages and Five Whys use
Stage What happens Five Whys contribution
Respond and make safe Treat casualties, prevent escalation and preserve urgent evidence. Do not analyse until the immediate risk is controlled.
Decide level and team Set investigation depth based on actual and potential severity. Use only where the method matches event complexity.
Gather information Collect scene, witness, document, digital and technical evidence. Create evidence-backed starting points.
Build timeline Describe what happened before, during and after the event. Check that each why fits the sequence.
Analyse causes Identify immediate, underlying and system causes. Drill down into selected causal paths.
Select actions Choose stronger controls and assign responsibilities. Make sure actions address verified causes.
Review effectiveness Check whether the action reduced risk. Test whether recurrence is less likely.

When should Five Whys be used?

Five Whys is best for bounded problems where the facts are reasonably clear and one or more causal paths can be tested.

The method is useful for repeated minor events, housekeeping problems, simple equipment defects, near misses, quality issues, procedural failures and focused lines of enquiry within a larger investigation. It is also useful as a teaching tool because it helps learners move beyond immediate causes.

Use it with caution where there is serious harm, major accident potential, technical uncertainty, conflicting accounts, contractor interfaces, fatigue, process safety barriers, design failure, emergency response weakness or possible legal sensitivity. In those cases, Five Whys may still help, but only as one tool among several.

Good and poor uses of Five Whys
Situation How to use it Watch point
Simple near miss Use a short evidence-led chain to identify why the control nearly failed. Still check potential severity.
Repeated minor incidents Use Five Whys on a representative event and compare with trend data. Repetition may indicate a system weakness.
Toolbox learning Use it to teach workers how surface causes lead to deeper controls. Avoid public blame discussions.
High-potential event Use only under competent investigation governance. Combine with timeline, fishbone and barrier review.
Technical failure Use for management questions but involve technical specialists. Do not replace engineering analysis.
Disciplinary concern Do not use Five Whys as a disciplinary shortcut. Separate fair investigation from blame-led assumptions.

How should Five Whys be completed step by step?

The method should be written as a chain of evidence-tested answers, with branches where several causes are credible.

Start with a neutral event statement. Good wording describes the event without judgement: 'Worker slipped on oil in workshop bay 2 and fractured wrist', not 'worker was careless'. Then ask why the event happened and write the answer as a condition that can be tested.

Continue only while the answers remain evidence-supported. If the team does not know the answer, mark it as a question and gather more information. If two credible answers exist, branch the chain rather than arguing one away. Stop when the answer is a controllable weakness that can be corrected and checked.

Five Whys step-by-step method
Step Instruction Quality check
1. Define the event Write what happened, where, when and the actual or potential consequence. Is it neutral and specific?
2. Confirm evidence Check photographs, statements, task records, permits, maintenance, training and relevant data. What facts are available and what is missing?
3. Ask the first why Explain why the event happened in factual terms. Is the answer supported by evidence?
4. Ask why that condition existed Move from the immediate event to the work condition behind it. Is the answer controllable by the organisation?
5. Branch where needed Create more than one line if equipment, people, environment or management factors interacted. Are we forcing a single chain?
6. Stop at useful causes Stop when the team identifies correctable system weaknesses. Would fixing this make recurrence less likely?
7. Test backwards Read the chain from final cause back to event. Does the logic still make sense?
8. Convert to actions Assign actions, owners, dates and effectiveness checks. Do actions strengthen controls or merely remind people?

What evidence standard should each why meet?

For every “why”, record the evidence, its source and what could disprove the explanation. Where several causes interact, branch the chain instead of forcing one path to a preferred answer.

A good answer is specific, testable and connected to the event. It should be possible to say what evidence supports it and what evidence might disprove it. If the answer relies on memory, opinion or assumption, that does not make it useless, but it should be marked as an open line of enquiry.

The evidence standard protects the investigation from hindsight bias. After an incident, the warning signs can look obvious. The job of the investigator is to understand what was known, expected, available and practical at the time.

Evidence checks for Five Whys
Evidence type Examples How to use it
Physical evidence Scene condition, damaged parts, guards, spills, labels, isolation points and PPE. Photograph and compare with expected condition.
Documentary evidence Risk assessment, RAMS, permit, inspection, maintenance, training and audit records. Check whether paperwork matched work-as-done.
Digital evidence CCTV, access logs, telemetry, alarm history, process data and messages. Use timestamps to test the sequence.
People evidence Witness accounts, worker experience, supervisor explanation and contractor input. Interview respectfully and test consistency.
Context evidence Workload, staffing, weather, lighting, shift pattern, fatigue, change and time pressure. Explain why decisions or conditions existed.
Technical evidence Engineering inspection, failure analysis, sampling, exposure results or specialist opinion. Use when the cause cannot be assessed by generalists.

When should a Five Whys chain branch?

A Five Whys chain should branch whenever more than one credible condition contributed to the event.

Many poor Five Whys worksheets fail because they force a single line. A forklift near miss may involve pedestrian route design, reversing visibility, delivery pressure, supervision, driver competence and a temporary storage change. One chain cannot hold that complexity without distorting it.

Branching does not make the method complicated; it makes it honest. Each branch can be short, but it should follow evidence. The final report can then explain how the branches interacted and which controls need strengthening.

Branching prompts for Five Whys
Prompt Why it matters Example branch
Did more than one barrier fail? Incidents often occur when controls line up badly. Guard missing, isolation not verified and supervision absent.
Did people, plant and process all contribute? Different cause types need different actions. Competence, machine design and procedure conflict.
Was there recent change? Change can create hidden mismatches. Temporary layout, new contractor or altered shift pattern.
Were there repeated weak signals? Previous near misses may show the system already knew. Defect reports, audit findings or informal workarounds.
Was work-as-done different from work-as-imagined? Procedures may not reflect real constraints. Task normally requires shortcut to meet production demand.
Was the final answer too neat? Overly simple chains may hide organisational causes. Final cause is 'carelessness' or 'did not follow procedure'.

How should immediate, underlying and root causes be separated?

Five Whys should help the team move from what directly happened to why the control system allowed it to happen.

Different organisations use slightly different language, but the investigation purpose is consistent. Immediate causes are closest to the event. Underlying causes explain the unsafe conditions or acts that allowed the event. Root or system causes sit deeper in management arrangements, design, competence, maintenance, planning, procurement, leadership or review.

The chain is useful only when it creates a clear line from event to action. A root cause that nobody can control is too vague. A root cause that points only to an individual's attitude is usually too shallow.

Cause levels in Five Whys
Cause level Plain English meaning Example
Immediate cause HSG245 describes the condition or action closest to harm; organisations may use different terminology. Oil underfoot and the resulting loss of footing are close to the harm; investigate how the leak and failed controls arose.
Underlying cause The unsafe condition or system gap that allowed the immediate cause. Oil leak had not been identified or contained before work started.
Root/system cause A deeper management weakness from which other failings grew. Inspection regime did not cover leak-prone equipment and defect escalation was weak.
Contributing factor A condition that increased likelihood or severity. Lighting made the oil difficult to see.
Latent condition A hidden weakness present before the event. Maintenance backlog had normalised small leaks.
Correctable cause A cause the organisation can realistically control. Revise inspection, containment and defect close-out arrangements.

HSG245 2004 and HSE GB human factors support investigation stages and evidence-led causal analysis. Five Whys chains are illustrative hypotheses, not a mandated method or proof; reporting law is separate.

How should Five Whys handle human error?

A decision or action by a worker may be one fact, but it rarely explains why the work system allowed the hazard. Test task design, equipment, information, staffing, supervision and organisational pressure.

People make decisions inside a work system. They see certain information, use available equipment, manage workload, respond to supervision, follow local norms and adapt to pressure. If an investigation stops at 'the worker forgot', it has usually stopped just when the useful questions begin.

Better questions include: why was the task easy to forget, why was the interface confusing, why was the hold point missing, why did the person feel unable to stop, why had supervision not detected drift and why had previous warnings not led to change?

Human factors prompts for Five Whys
Area Questions to ask Possible evidence
Task Was the task complex, rushed, abnormal, poorly sequenced or hard to perform safely? Task observation, method statement and worker accounts.
Equipment Was the equipment intuitive, reliable, guarded, labelled and suitable? Inspection, design information and defect history.
Environment Did noise, heat, lighting, space, weather or congestion influence the event? Scene photographs and environmental records.
Competence Was competence defined, trained, assessed and refreshed for the actual task? Training matrix, assessment record and supervisor check.
Supervision Were expectations clear and was work checked at the right points? Briefings, supervision logs and shift handover.
Organisation Did resources, workload, culture, procurement, maintenance or production pressure shape the conditions? Management review, audit findings and action trackers.

Use HSE human-factors guidance to test job, individual and organisational influences without stopping at blame.

What does a good Five Whys example look like?

The conveyor chain below is a set of possible explanations, not a verified finding. Check each step against machine condition, energy-isolation records, work observation, witness accounts and maintenance history before accepting a cause.

A suspected wrong part needs identity evidence across the chain

A suspected wrong part needs identity evidence across the chain. A normal machine-workshop review shows a component inside a closed box, review papers and a closed machine enclosure. Picture key: 1 Closed component box · 2 Review papers · 3 Closed machine enclosure. The boxed component is not shown fitted. Three separate analytical comparisons distinguish the specified requirement and revision, ordering/receipt identity and revision, and the actual fitted component identity. Two dashed conceptual links compare those questions; a separate coupling-and-shaft insert is for identity only, not assembly, tolerance or live testing. The box, machine and records do not establish a match, mismatch, alloy, fit, energy state, fault or permission. Four unnumbered checks protect the task, trace identity and revisions, test competing explanations and verify an effective correction through competent arrangements. Keep uncertainty open; a neat Why chain does not justify restart.
Normal component-and-record review. The boxed component is not fitted; the separate analytical comparison supplies no match, alloy, fit, state or restart finding.
Full written explanation

A normal machine-workshop review shows a component inside a closed box, review papers and a closed machine enclosure. Picture key: 1 Closed component box · 2 Review papers · 3 Closed machine enclosure. The boxed component is not shown fitted. Three separate analytical comparisons distinguish the specified requirement and revision, ordering/receipt identity and revision, and the actual fitted component identity. Two dashed conceptual links compare those questions; a separate coupling-and-shaft insert is for identity only, not assembly, tolerance or live testing. The box, machine and records do not establish a match, mismatch, alloy, fit, energy state, fault or permission. Four unnumbered checks protect the task, trace identity and revisions, test competing explanations and verify an effective correction through competent arrangements. Keep uncertainty open; a neat Why chain does not justify restart.

Machine-workshop comparison: specification, order, receipt and fitted identity are questions to check, not assumed failures.

  1. Protect the equipment and task. Use competent arrangements and preserve evidence; do not recreate the failure.
  2. Trace identity and current revisions. Compare actual specification, ordering/receipt records and the fitted part.
  3. Test competing explanations. Keep supported, contrary and missing evidence distinct before assigning a cause.
  4. Verify the effective correction. Use competent change/acceptance checks and confirm the outcome in real work safely.

When the evidence does not fit: An uncertain identity or neat Why chain does not justify restart. Keep the affected task protected and unresolved causes open for competent review.

Who acts: Investigation lead, engineering/maintenance owner, procurement or information owner and affected operations team.

Changed workplace

Distribution centre: Test interacting physical-route and delivery/change explanations with their sources.

Machine-workshop part identity: Trace specification, records and fitted identity without assuming which organisation failed.

Evidence check: Illustrative hypotheses only. No actual near miss, component mismatch, blame finding, fit/tolerance, assembly instruction or permission to restart is assigned.

Read the evidence-led control hierarchy and retain the current conveyor worked example

Primary references: HSE: HSG245 investigation stages, 2004 · HSE: human factors in investigation · HSE: machinery precautions

Open the complete HD image (new tab)Download image

Read the evidence-led control hierarchy and retain the current conveyor worked example

Primary references: HSE: HSG245 investigation stages, 2004 · HSE: human factors in investigation · HSE: machinery precautions

Consider a worker who cuts their hand while clearing a jam on a powered conveyor. A weak investigation might conclude that the worker failed to follow the procedure. A stronger Five Whys chain tests why the jam had to be cleared manually, why isolation was not used, why supervision did not detect the practice and why previous jam reports had not led to redesign.

The result should be practical. If the chain ends with 'worker did not follow procedure', the action will likely be a reminder. If the chain reaches design, isolation, planning and maintenance weaknesses, the action can reduce risk at source.

Worked example: hand injury during jam clearing
Why level Hypothesis to test Action implication Evidence to check
Event Worker cut hand while clearing jammed material from powered conveyor. Make area safe, treat injury and preserve evidence. Injury and incident record, machine state, photographs.
Why 1 Hand entered nip point while stored energy remained available. Review guarding, isolation and energy control. Guard/access condition and isolation test.
Why 2 Jam clearing was routinely done through access panel during production stops. Observe work-as-done and revise method. Work observation, production and training records.
Why 3 Procedure required isolation, but production targets and frequent jams encouraged quick clearing. Address workload, planning and supervisor expectations. Procedure, shift targets, interviews and supervision records.
Why 4 Frequent jams had been reported but treated as production inconvenience, not a safety design issue. Review defect escalation and engineering change criteria. Jam/defect logs and escalation trail.
Why 5 The maintenance and improvement process did not classify repeated jams as a safeguarding concern. Introduce risk-based escalation for repeated equipment faults. Maintenance priority and safeguarding review criteria.
Possible combined finding if evidence confirms The injury arose from exposed energy during a normalised workaround, driven by recurring equipment jams and weak escalation. Engineering review, interlocked access, jam reduction, permit/LOTO verification and action effectiveness checks. Converging physical and documentary evidence; competing explanations.

Observe or test tasks only under a safe competent plan. Stop affected machinery work until guarding, isolation and task controls are verified; investigate past records or use controlled simulations instead of recreating an injury mechanism.

How do Five Whys and fishbone analysis work together?

Fishbone analysis widens the investigation; Five Whys drills down into selected causes.

Fishbone analysis is useful when the team needs to map many possible causes across people, plant, method, environment, materials and management system factors. Five Whys is useful when the team wants to examine one branch more deeply.

Used together, they reduce two opposite risks. Five Whys alone can become too narrow. Fishbone alone can become a long list of possibilities without depth. The stronger workflow is to map the landscape, select important branches, drill down, test evidence, then write actions.

Using fishbone and Five Whys together
Stage Tool Output
Define event Both Neutral problem statement.
Map possible causes Fishbone Broad categories and possible causal factors.
Prioritise branches Team judgement Evidence-backed lines needing deeper enquiry.
Drill down Five Whys Deeper cause chains under selected branches.
Test evidence Both Confirmed, possible, disproved and open findings.
Choose actions Action review Corrective actions linked to verified causes.

How should Five Whys connect to barriers and controls?

Every cause should be connected back to the control that should have prevented, detected or mitigated the event.

Barrier thinking makes Five Whys more useful. Instead of asking only why the event happened, ask which barrier was missing, weak, defeated, bypassed, unavailable, misunderstood or unverified. This keeps the analysis close to risk control.

For higher-risk work, the final output should show whether prevention, detection, control, mitigation, emergency response and assurance arrangements were reliable. If the analysis cannot name the failed control, it may still be too vague.

Barrier prompts for Five Whys
Barrier question What it reveals Example
What should have prevented the event? Missing or weak prevention controls. Guarding, segregation, isolation, ventilation or planning.
What should have detected the problem earlier? Inspection, monitoring or alarm gaps. Defect reporting, gas detection, audits or supervision.
What should have controlled the situation? Procedures, permits, competence or engineering controls. Permit hold point or lock-off check.
What should have reduced the consequence? Emergency response, PPE or mitigation weakness. First aid, spill kit, fire response or eyewash.
How was the barrier assured? Whether controls were checked, tested and maintained. Audit, inspection, maintenance or competence verification.
Could the same barrier be weak elsewhere? Extent of condition and wider learning. Other lines, sites, shifts or contractors.

How should Five Whys findings become strong corrective actions?

Close an investigation action only after a named owner installs the measure, checks it in real work and tests whether the failure recurs. Training alone is weak when the verified problem is equipment or layout.

The action plan is where investigation quality becomes visible. Weak actions tell people to be careful, repeat training without evidence of a competence gap, or update a procedure nobody uses. Strong actions change the work system: design, engineering controls, access, maintenance, planning, supervision, procurement, competence verification and leadership review.

Actions should also be tested. Completion is not the same as effectiveness. A revised procedure is complete when it is issued; it is effective only when the task is observed, workers understand it and the original weakness has been reduced.

Corrective action strength
Action type Strength Example
Eliminate or redesign Strong Remove the need to access the danger zone during clearing.
Engineering control Strong Interlock access panel, improve guarding or redesign equipment to reduce jams.
System control Strong to medium Introduce risk-based defect escalation and independent isolation verification.
Planning and resources Medium to strong Allow downtime to clear defects rather than normalising workarounds.
Competence and supervision Medium Target training and supervisor checks to a verified gap.
Procedure update Medium to weak unless verified Revise method and check use in real work.
Reminder or poster Weak if used alone Use only as supporting communication, not the main action for a serious finding.

How can the hierarchy of controls turn Five Whys findings into effective risk reduction?

Use ERICPD and the HSE five-level hierarchy on the hazard that evidence reveals. The Why chain helps locate a failed barrier; it is not itself a protective measure.

A near miss can have interacting explanations, not one blame chain

A near miss can have interacting explanations, not one blame chain. A normal distribution-warehouse scene provides rail, draft-layout-board and clipboard context beside a separate branching Why enquiry. Picture key: 1 Warehouse rail · 2 Draft layout board · 3 Review clipboard. A reported reversing near miss is the neutral analytical starting point, not a pictured incident. Two independent dashed candidate branches ask about route/separation and actual work conditions, and about delivery/change information, revisions, bookings and communication. Both join a demonstrated-control-gap question before owned action and effectiveness checking. Retain contrary evidence and interacting explanations; exactly five answers are not required. Three complete paired explanations establish the event basis, test branches and choose stronger action. The scene proves no event, cause, safe route, barrier adequacy, role, equipment state or permission.
Normal distribution-centre review beside two candidate Why branches. Neither the draft board nor rail establishes a near miss, cause or approved traffic arrangement.
Full written explanation

A normal distribution-warehouse scene provides rail, draft-layout-board and clipboard context beside a separate branching Why enquiry. Picture key: 1 Warehouse rail · 2 Draft layout board · 3 Review clipboard. A reported reversing near miss is the neutral analytical starting point, not a pictured incident. Two independent dashed candidate branches ask about route/separation and actual work conditions, and about delivery/change information, revisions, bookings and communication. Both join a demonstrated-control-gap question before owned action and effectiveness checking. Retain contrary evidence and interacting explanations; exactly five answers are not required. Three complete paired explanations establish the event basis, test branches and choose stronger action. The scene proves no event, cause, safe route, barrier adequacy, role, equipment state or permission.

Distribution-centre example: each Why is a hypothesis until its evidence and competing explanations are checked.

  1. Start from a verified event basis. Separate observation, source and uncertainty before asking why a barrier failed.
  2. Branch and test the explanations. Compare physical protection, task/change information and management evidence.
  3. Choose and test stronger action. Act on the supported gap and verify the actual outcome, reopening repeated failure.

Change the pathway

  • Five is not a proof requirement. Stop a line when evidence does not support it; follow interacting branches where needed.
  • A plausible answer is not a source. Record contrary material and what remains to be checked instead of forcing agreement.
  • A Why chain is not a safeguard. Select a feasible source-level or physical control before relying only on reminders.

Evidence check: Illustrative hypotheses only. No actual near miss, component mismatch, blame finding, fit/tolerance, assembly instruction or permission to restart is assigned.

Read the evidence-led control hierarchy and retain the current conveyor worked example

Primary references: HSE: HSG245 investigation stages, 2004 · HSE: human factors in investigation · HSE: machinery precautions

Open the complete HD image (new tab)Download image

Read the evidence-led control hierarchy and retain the current conveyor worked example

Primary references: HSE: HSG245 investigation stages, 2004 · HSE: human factors in investigation · HSE: machinery precautions

HSG245 2004 and HSE GB human factors support investigation stages and evidence-led causal analysis. Five Whys chains are illustrative hypotheses, not a mandated method or proof; reporting law is separate. GB traffic/DSEAR sources support the respective illustrative control examples.

A delivery change causes a reversing vehicle to pass near a worker. Test the route plan, barriers, booking and supervision before accepting “worker error” as a cause.

A workshop has repeated solvent spills. Test container design, couplings, task timing and maintenance against physical evidence before choosing a corrective measure.

What are the limitations of Five Whys?

A single linear chain can miss parallel failures and time sequence. For complex or high-potential events, use a timeline, fishbone or barrier analysis and specialist evidence alongside Five Whys.

The method can become linear, biased and too shallow. It can also produce different answers depending on who is in the room. A manager may see behaviour; a worker may see poor equipment; an engineer may see design; a safety adviser may see missing assurance. The value comes from testing these perspectives, not choosing the quickest one.

For major injury, major accident potential, process safety, complex technical failure, occupational disease, high-value asset damage or legal sensitivity, Five Whys should be supported by competent investigation governance and other tools.

Five Whys limitations and safeguards
Limitation How it can mislead Safeguard
Linear thinking Forces one chain even where causes interact. Branch the analysis.
Hindsight bias Makes the event look obvious after it happened. Ask what was known and available at the time.
Blame endpoint Stops at carelessness or failure to follow rules. Ask why the behaviour was possible or likely.
Weak evidence Turns guesses into causes. Mark evidence status and gather missing information.
Over-simple actions Produces retraining and reminders. Review action strength and effectiveness checks.
Wrong tool for complexity Misses technical, organisational or barrier interactions. Combine with fishbone, timeline, barrier or technical analysis.

What should a Five Whys action register include?

The register should connect each verified cause to a control improvement, owner, timescale and effectiveness check.

Five Whys action register fields
Field Purpose Example
Event reference Links the analysis to the incident or near miss. INV-2026-021
Problem statement Defines the event neutrally. Worker slipped on oil in workshop bay.
Cause chain or branch Shows the evidence-led why sequence. Leak not detected because inspection route excluded rear of machine.
Evidence status Separates confirmed, possible and open findings. Confirmed by inspection record and photo.
Control weakness Names the failed or missing control. Inspection regime and defect escalation.
Action States the practical improvement. Revise inspection route and add leak-response standard.
Owner and due date Creates accountability. Workshop manager, 30 September 2026.
Effectiveness check Confirms whether the action works. Three inspection samples and worker feedback after one month.
Wider applicability Checks whether the weakness exists elsewhere. Review similar machines across all workshops.

Who needs Five Whys competence, and what proves it?

Test evidence handling, neutral questioning, jurisdiction screening and action effectiveness with a realistic incident, not attendance alone.

Who needs to demonstrate what
RoleLearning to testCompetence demonstrationAssessor and recordReassessment trigger
Workers and witnessesPreserve facts and distinguish observation from interpretation.Describe the timeline and evidence source.Supervisor records briefing.Incident or process change.
SupervisorsSecure scene and screen GB or NI reportability.Work through sample report route and evidence capture.Safety lead reviews decision.Jurisdiction or threshold change.
InvestigatorsTest each Why and branch where evidence supports multiple causes.Build a chain with counter-evidence and causal rationale.Competent investigation lead reviews.Complex event or new evidence.
Technical specialistsValidate engineering or human-factor findings.Use safe analysis, inspection or controlled tests to validate barrier design; do not recreate exposure to live machinery or other hazardous failure.Independent authority signs technical finding.Design or failure mode change.
Action ownersInstall stronger controls and verify their effect.Show field test, owner and recurrence review.Sponsor signs closeout.Control fails or event recurs.

How does Five Whys support study and professional development?

Five Whys helps learners explain the movement from immediate causes to underlying and root/system causes.

For practical learning, place Five Whys inside the wider investigation process. Start with evidence gathering, then analyse causes, select controls and review whether actions work. Do not write a generic chain that ends with 'worker error'. Show how work design, supervision, equipment, competence and management arrangements shaped the event.

Scenario-based answers should be specific. In a machinery injury, discuss guarding, isolation, maintenance, jams, supervision and production pressure. In a slip incident, discuss contamination, flooring, cleaning, footwear, lighting, reporting and inspection. In a process safety event, discuss barriers, alarms, procedures, competence, change control and assurance.

Study and professional application
What to know How to apply it Common trap
HSG245 process Use Five Whys during analysis, after evidence gathering. Starting with causes before facts.
Cause levels Separate immediate, underlying and root/system causes. Calling every answer a root cause.
Evidence status Label confirmed, possible, disproved and open answers. Treating assumptions as facts.
Human factors Explore job, individual and organisational conditions. Stopping at human error.
Action strength Choose controls that reduce recurrence risk. Relying on reminders and retraining only.

How can BS EN ISO 45001 support five whys for incident investigation management?

A voluntary occupational health and safety system can connect evidence-led investigation, verified causes, stronger controls and effectiveness checks without treating a Why chain as proof.

HSE public guidance supports practical management themes in Great Britain. ISO clause mapping, conformity and certification claims remain held for current licensed-text and specialist review; this page does not establish compliance or certification.

What are the warning signs of poor Five Whys analysis?

Poor Five Whys analysis usually sounds neat, quick and certain, but does not improve the work system.

Five Whys warning signs
Warning sign Why it weakens learning Better practice
The chain starts with blame The investigation begins from a conclusion. Use a neutral event statement.
Every answer is opinion The chain cannot be defended. Show evidence or mark as open.
There is only one line for a complex event Interactions are hidden. Branch the chain.
Human error is the final answer System causes are missed. Ask why the error was possible.
The answer repeats the event No deeper cause has been found. Ask what condition allowed it.
Actions are reminders Controls are not strengthened. Use stronger design, engineering, planning or assurance actions.
No effectiveness check exists Actions can close without reducing risk. Define field verification or trend review.

What do the acronyms and key terms mean?

Acronyms should be explained so investigation language remains accessible to workers, supervisors and learners.

Acronyms and meanings
Acronym Meaning Plain English use
HSG245 HSE guidance on investigating accidents and incidents Main HSE investigation workbook.
HSE Health and Safety Executive Great Britain workplace health and safety regulator.
RCA Root Cause Analysis Structured analysis of deeper causes and system weaknesses.
RIDDOR Reporting of Injuries, Diseases and Dangerous Occurrences Regulations Statutory reporting regime for specified work-related events.
GBGreat BritainGB RIDDOR 2013 route.
NINorthern IrelandSeparate NI RIDDOR 1997 route.
ERICPDEliminate, Reduce, Isolate, Control, PPE, DisciplineControl-selection teaching aid applied to verified hazards.

Reader navigation, terms and proposed documents only; no linked-page evidence, instructional asset or reuse approval is implied.

Documents to download

Download editable resources from the SHAREhub environment to support your work. Working forms and spreadsheet input areas are blank. Reports, plans, guides, briefings, procedures and policies contain completed worked example content to adapt.

Return to SHARE home