Skip to content

Worked examples · 10 chains

5 whys examples: worked chains with the evidence for each why

A 5 whys chain is only as good as the evidence behind each answer. A chain that sounds right but was never checked usually ends at a person, a procedure someone did not follow, or the fix the team already wanted. Below are ten finished chains. Every why shows the answer and the evidence that backs it, then the verified cause, the countermeasure and how the team proved the fix worked. Each one also names the wrong turn that was available.

Illustrative examples. These chains are illustrative: written from conditions common in plants, not from one company. Plants, roles and counts are made up for the example and are not benchmarks. Want a blank sheet for your own problem? The 5 whys template is free in PDF and Excel.

How to read each chain

Each chain starts with a problem stated as a measured gap, then asks why, one step at a time. Beside every answer is the evidence that checks it and how that evidence was gathered. The chain ends at a cause in how the work or the system is set up, not in a person.

Under the chain you will find the wrong turn that was available, the cause, the countermeasure, a line that reads the chain back to the problem, and the check that the fix worked.

A why chain read down with why and back up with soFour boxes, problem, why 1, why 2 and cause. Going down you ask why. Going back up you say so. If a link needs a maybe, it is weak.Problema measured gapwhy?soWhy 1answer + evidencewhy?soWhy 2answer + evidencewhy?soCausesomething you can fix
Down with why, back up with so.

Six steps for a chain you can defend

  1. State the problem as a gap. What happened, where, how often, against what standard. Numbers, not adjectives.
  2. Ask why, one step at a time. Answer from what you saw, not from what you assume.
  3. Write the evidence next to the answer. What was counted, measured or watched, and when. If there is none, mark it a guess and go and get it.
  4. Branch when two answers are true. Give each branch its own line and test both.
  5. Stop at a cause you can fix. A cause in the setup of the work or the system. If it names a person or says "did not follow", ask why again.
  6. Fix, then check. Put the fix in place, then measure the original problem for a set window.

Five is not the point

These chains run from two to five whys. The Lean Enterprise Institute lexicon puts it plainly: the specific number five is not the point; keep asking until the root cause is reached and eliminated. The two-why examples branch: the chain stopped because the answer split into two causes, and each needed its own test.

Six ways to get evidence for a why

  • Count

    Tally or sort data you already have: which cavity, which shift, which part number.

  • Measurement

    Put an instrument on it: temperature, current, vibration, dimension.

  • Observation

    Go and watch the work or the equipment, at the time it happens.

  • Record

    Read the procedure, the work instruction, the maintenance history or the change order.

  • Trial

    Change one thing on purpose and see whether the effect moves.

  • Interview

    Ask the people who do the job what happens, with open questions, never who.

Ten worked chains, by area

Each answer carries its evidence. The orange box in each chain is the wrong turn a team could have taken and what it would have cost, so you can spot the same one on your own sheet.

Area
Wrong turn

Showing 10 of 10 chains

Chain 1 of 10 · Manufacturing · 5 whys

Molding: short shots on press 7 after weeks of running well

Team:
Process technician, maintenance technician and shift supervisor
Problem:
Press 7 produced 62 short shots in 2,000 parts over the last five days, 3.1%. The month before it ran at 8 in 2,000, 0.4%.

The chain, with the evidence for each answer

  1. Why are the parts short?

    Cavity 4 is not filling completely.

    Evidence, count: Sorted the 62 rejects by the cavity number molded into each part: 54 of 62 came from cavity 4.

  2. Why is cavity 4 not filling?

    The melt reaching it is cooler than the recipe says.

    Evidence, measurement: Pyrometer reading on purged melt: 212 C. The recipe setpoint is 230 C and the same check three weeks ago read 229 C.

  3. Why is the melt cooler?

    Barrel heater zone 3 is not heating, although the controller shows it on.

    Evidence, measurement: Clamp-on ammeter on the zone 3 band read 0 A. Zones 2 and 4 read 6.2 A and 6.0 A.

  4. Why is zone 3 drawing no current?

    The band's terminal screw has backed off and the connection is open.

    Evidence, observation: Photo of the terminal: loose screw, discoloured lug. Re-tightened with the press locked out, current read 6.1 A.

  5. Why did the screw back off without anyone knowing?

    Heater terminals are not on the preventive maintenance list, and the start-up check does not read heater current.

    Evidence, record: The quarterly PM task list for press 7 has 23 items, none of them heater terminals. The start-up sheet has no current column.

The wrong turn: jumping to a solution

Stopping at: Why 2: the melt is cooler. The fix would be: Raise the zone 3 setpoint by 10 C.

What happens: The controller already shows the band on, so a higher setpoint changes nothing. The display would look right while the plastic stayed cold.

Verified cause

A loose heater terminal that no PM task or start-up check looks at.

Read back: The terminal checks are missing, so the screw backed off unseen, so zone 3 drew no current, so the melt ran cooler, so cavity 4 did not fill, so the parts came out short.

Countermeasure

Replace the lug and band terminal. Add a terminal check with a torque value to the quarterly PM list for every press. Add a heater current reading for each zone to the start-up sheet.

How it was verified

Counted short shots after the repair over five production days and read the zone currents at each start-up.

Short shots, 5 production days after the repair: 62 of 2,000 (3.1%) before, 8 of 2,000 (0.4%) after.

Short shots, before and after the fixShort shots, 5 production days after the repair: 62 of 2,000 (3.1%) before, 8 of 2,000 (0.4%) after.Before3.1%After0.4%
Short shots, 5 production days after the repair: 62 of 2,000 (3.1%) before, 8 of 2,000 (0.4%) after.

Take to your own problem: 5 Whys template and breakdown analysis report.

Chain 2 of 10 · Manufacturing · 2 whys and a branch

Shipping: orders leaving a packing line late, with two causes

Team:
Production manager, planning manager and packaging team leader
Problem:
In four weeks, 14 of 80 customer orders from the packing line shipped after the promised date, 17.5%. The target is under 5%.

The chain, with the evidence for each answer

  1. Why did these 14 orders ship late?

    The packing line started them late.

    Evidence, record: Pulled start times from the line log: all 14 started more than 3 hours after the scheduled time.

  2. Why did the line start late?

    Two reasons show in the log. Changeovers overran, and the packaging film was not at the line.

    Evidence, count: Of the 14 late starts, 8 followed a changeover longer than the 40 minute standard and 6 waited for film.

The chain branches at why 2

Branch A: changeovers over 40 minutes. Branch B: film not at the line.

Evidence: 8 plus 6 equals the 14. No order appears in both groups.

The wrong turn: following only one branch

Stopping at: The larger group: 8 of the 14 are changeovers. The fix would be: Run a changeover kaizen and ignore the film shortages.

What happens: Even a perfect changeover fix leaves 6 late orders in 80, 7.5%, above the 5% target.

Verified cause

Branch A: the changeover kit is staged after the previous run ends, so the crew waits for tools and film cores. Branch B: the warehouse picks film from the planner's list, printed the evening before and not updated when the schedule changes.

Read back: Film is picked from an old list and kits are staged late, so film and tools are missing at the line, so starts slip, so orders ship late.

Countermeasure

Branch A: stage the next kit during the current run and add a kit check to the schedule. Branch B: the line leader requests film from the live schedule at the start of each shift. Both go on the tier 1 board.

How it was verified

Counted late orders over the next four weeks and logged the cause of every late start.

Orders shipped late, 4 weeks after both fixes: 14 of 80 (17.5%) before, 3 of 80 (3.8%) after.

Orders shipped late, before and after the fixOrders shipped late, 4 weeks after both fixes: 14 of 80 (17.5%) before, 3 of 80 (3.8%) after.Before17.5%After3.8%
Orders shipped late, 4 weeks after both fixes: 14 of 80 (17.5%) before, 3 of 80 (3.8%) after.

Take to your own problem: 5 Whys template and changeover checklist.

Chain 3 of 10 · Manufacturing · 4 whys

Assembly: a fastener left loose on variant B units

Team:
Manufacturing engineer, assembly team leader and quality engineer
Problem:
An end-of-line audit found 11 of 600 units, 1.8%, with one fastener under the specified torque. All 11 were variant B.

The chain, with the evidence for each answer

  1. Why were the fasteners under torque?

    The torque step was not done on those units.

    Evidence, record: The torque tool log for the 11 serial numbers shows five tightening events, not six.

  2. Why was the step not done?

    The operators did not know it was a step on variant B.

    Evidence, interview: Asked three operators on the station to show how they build variant B. All three did five tightenings and none mentioned a sixth.

  3. Why did the operators not know?

    The work instruction at the station lists five fasteners. Variant B has six.

    Evidence, record: Station work instruction revision A shows five. The variant B drawing lists six, the extra one added in the last engineering change.

  4. Why does the instruction not show the sixth?

    The engineering change process has no step to update work instructions at the stations affected.

    Evidence, record: The change order form has a drawing, BOM and PFMEA checklist. It has no line for work instructions or a station check.

The wrong turn: stopping at human error

Stopping at: Why 2: the operator did not do it. The fix would be: Retrain the three operators and post a reminder.

What happens: The next operators read the same wrong instruction. The cause was the instruction, not the person, and the next change order would do it again.

Verified cause

The engineering change process does not update the work instruction at the station.

Read back: The change process skips work instructions, so the instruction showed five, so the operators did five, so one fastener was missed on some units, so loose fasteners reached the audit.

Countermeasure

Correct the instruction and post it. Add work instruction update and a station check to the change order form. Set the torque tool to count tightening events, so a unit with fewer than six cannot pass the station.

How it was verified

Audited every unit at end of line for three weeks and checked the tool's count on each serial number.

Units with a fastener under torque, 3 weeks, 100% audit: 11 of 600 (1.8%) before, 0 of 600 (0%) after.

Units with a fastener under torque, before and after the fixUnits with a fastener under torque, 3 weeks, 100% audit: 11 of 600 (1.8%) before, 0 of 600 (0%) after.Before1.8%After0%
Units with a fastener under torque, 3 weeks, 100% audit: 11 of 600 (1.8%) before, 0 of 600 (0%) after.

Take to your own problem: 5 Whys template and standard work audit checklist.

Chain 4 of 10 · Quality · 3 whys

Packaging: cartons shipped with the wrong product label

Team:
Quality engineer, packaging supervisor and line operators
Problem:
Nine cartons in 1,800 packed in a month, 0.5%, carried the label of a different product. Two were found by a customer.

The chain, with the evidence for each answer

  1. Why did the cartons have the wrong label?

    The printer ran with the wrong label roll after a changeover.

    Evidence, record: Roll code on the printer at the time of the 9 events, read from the changeover photos: the wrong roll each time.

  2. Why was the wrong roll loaded?

    Two rolls look almost the same and sit on the same shelf.

    Evidence, observation: Side by side at the line: same colour and stock. The product code differs only in the last two digits of eight, printed in 8 point type.

  3. Why was the wrong roll not caught?

    The changeover checklist says "load label roll" and has no step that checks it against the order.

    Evidence, interview: Checklist read at the station: 14 steps, none verifies the roll code. Asked the operators what they check: that the roll fits.

The wrong turn: stopping at human error

Stopping at: Why 1: the operator loaded the wrong roll. The fix would be: Write a warning to the operators and add a second signature.

What happens: A signature on a form does not stop a roll that looks right from being loaded. The error returns when the line is busy.

Verified cause

Nothing in the process compares the label roll to the order, and two rolls are easy to mix up.

Read back: No check compares roll to order, so a look-alike roll is not caught, so the wrong roll runs, so cartons carry the wrong label.

Countermeasure

Scan the roll code against the order at changeover, with the printer set to refuse to start on a mismatch. Move look-alike rolls to separate shelves with large codes.

How it was verified

Loaded a wrong roll on purpose ten times to confirm the printer refused each time, then counted mislabelled cartons for three weeks.

Cartons with a wrong label, 3 weeks after the scan check: 9 of 1,800 (0.5%) before, 0 of 5,400 (0%) after.

Cartons with a wrong label, before and after the fixCartons with a wrong label, 3 weeks after the scan check: 9 of 1,800 (0.5%) before, 0 of 5,400 (0%) after.Before0.5%After0%
Cartons with a wrong label, 3 weeks after the scan check: 9 of 1,800 (0.5%) before, 0 of 5,400 (0%) after.

Take to your own problem: 5 Whys template and poka-yoke worksheet.

Chain 5 of 10 · Quality · 4 whys

Machining: burrs on parts that reached the customer

Team:
Quality engineer, machining supervisor and inspector
Problem:
The customer returned 24 of 3,000 shipped parts, 0.8%, for burrs on a bore edge.

The chain, with the evidence for each answer

  1. Why did burrs ship?

    Inspection did not catch them.

    Evidence, record: Reviewed the inspection records for the returned lots: all passed.

  2. Why did inspection not catch them? (A first try said "it is not thorough enough", which only restates the problem.)

    The visual check cannot reliably see a burr this small at line speed.

    Evidence, trial: Mixed 20 parts with a known burr into a lot and ran the normal check. The inspector found 14 of 20, 70%.

  3. Why were burrs made at all?

    The deburr tool runs past its wear point.

    Evidence, measurement: Measured burr height against parts since the last tool change: over the 0.15 mm limit at about 4,500 parts. The tool change is set at 6,000.

  4. Why is the change set at 6,000?

    The number was copied from a different part and never measured for this one.

    Evidence, record: Setup sheet history: 6,000 appears on the sheet for another part number and was carried over when this part started.

The wrong turn: going in circles

Stopping at: Why 2, as first worded: inspection is not thorough, so inspectors do not look closely. The fix would be: Tell inspectors to look harder and add a second inspector.

What happens: The answers loop back to "burrs got through". Nothing new was learned. The trial showed that even a diligent inspector finds 70%.

Verified cause

Two causes, found by splitting the question. Made: the tool change interval was never set for this part. Escaped: a visual check that sees 14 of 20 known burrs.

Read back: The interval was copied, so the tool wore past the limit, so burrs were made, and the check sees only 70%, so burrs reached the customer.

Countermeasure

Set the tool change at 4,000 parts, from the wear data, and review it after the first ten tool changes. Add a go/no-go edge gauge for the bore at the station.

How it was verified

Repeated the seeded-burr trial with the gauge in use, and counted customer returns on the next 3,000 parts.

Seeded burrs found by the check, Repeat trial, 20 seeded parts: 14 of 20 (70%) before, 20 of 20 (100%) after.

Seeded burrs found by the check, before and after the fixSeeded burrs found by the check, Repeat trial, 20 seeded parts: 14 of 20 (70%) before, 20 of 20 (100%) after.Before70%After100%
Seeded burrs found by the check, Repeat trial, 20 seeded parts: 14 of 20 (70%) before, 20 of 20 (100%) after.

Take to your own problem: 5 Whys template and defect catalogue template with boundary samples.

Chain 6 of 10 · Quality · 4 whys

Quality lab: parts pass in the plant and fail at the customer

Team:
Quality lead, lab technician and machining supervisor
Problem:
The customer's incoming inspection rejected 20 of 2,500 shafts, 0.8%, for diameter, while the plant's final check passed all of them.

The chain, with the evidence for each answer

  1. Why do the plant and the customer disagree?

    The two measurements of the same diameter are not the same.

    Evidence, measurement: Measured 10 returned shafts on the plant micrometer and on the customer's gauge. The plant reads about 0.012 mm smaller on every one.

  2. Why does the plant micrometer read small?

    It is out of calibration.

    Evidence, measurement: Checked against a 25.000 mm gauge block: it reads 24.988 mm.

  3. Why was it out of calibration?

    It was not on the calibration schedule, so it was never checked.

    Evidence, record: The calibration register lists micrometers by asset tag. This one carries no tag.

  4. Why does the micrometer have no tag?

    It is a spare that replaced a damaged one in the lab, and nobody tagged it when it entered use.

    Evidence, interview: The purchasing record shows the spare was bought 14 months ago. The lab technician confirms it went into use that week.

The wrong turn: jumping to a solution

Stopping at: Why 1: the measurements differ. The fix would be: Tighten the plant's internal limits by 0.015 mm.

What happens: Good parts would be scrapped, and the uncalibrated gauge would stay in use and affect every other part it measures.

Verified cause

Gauges enter use without being tagged and added to the calibration register.

Read back: A spare gauge was never tagged, so it was never calibrated, so it read small, so the plant passed shafts the customer measured as out of tolerance.

Countermeasure

Calibrate and tag the micrometer. Audit every gauge in the lab against the register. Add a rule that no gauge leaves stores without an asset tag, and review the past 14 months of results taken with it.

How it was verified

Audited all 42 gauges in the lab for a tag and an in-date calibration, then counted customer rejects on the next 2,500 shafts.

Shafts rejected by the customer, Next 2,500 shafts after the audit: 20 of 2,500 (0.8%) before, 0 of 2,500 (0%) after.

Shafts rejected by the customer, before and after the fixShafts rejected by the customer, Next 2,500 shafts after the audit: 20 of 2,500 (0.8%) before, 0 of 2,500 (0%) after.Before0.8%After0%
Shafts rejected by the customer, Next 2,500 shafts after the audit: 20 of 2,500 (0.8%) before, 0 of 2,500 (0%) after.

Take to your own problem: 5 Whys template and gauge calibration log.

Chain 7 of 10 · Safety · 3 whys

Warehouse: five forklift near-misses at one aisle crossing

Team:
EHS lead, warehouse supervisor and a forklift operator
Problem:
Five near-miss reports in six weeks at the same crossing, each a person on foot stepping into a forklift path. No one was hurt.

The chain, with the evidence for each answer

  1. Why did people step into the forklift path?

    They came through the door from packing straight into the aisle, where they cannot see an approaching truck.

    Evidence, observation: Watched the door for 40 minutes at the morning peak. A wall blocks the view of the aisle from the doorway.

  2. Why do they use that door?

    It is the shortest way to the break room.

    Evidence, count: Counted 31 people leaving packing in the 40 minutes: 22 used that door and 9 went the long way round.

  3. Why does the shortest way end in the truck aisle?

    The floor marking leads from the door across the aisle, with no barrier and no stop point.

    Evidence, record: Floor plan and photos: the painted walkway crosses the aisle at the door. Trucks pass about every 3 minutes.

The wrong turn: stopping at human error

Stopping at: The near-miss reports said "pedestrian not paying attention". The fix would be: Remind everyone to look both ways at the door.

What happens: A rule that depends on every person, every time, stays in place at a spot where they cannot see. The next report will say the same thing.

Verified cause

The route people take ends at a blind crossing with nothing to stop them or the truck.

Read back: The shortest route ends at a blind crossing, so 22 of 31 walk it, so people step into the truck path, so near-misses happen.

Countermeasure

Fit a barrier so the door exit turns along a protected walkway to a marked crossing with a clear view. Add a convex mirror at the corner as a second layer, not the main fix.

How it was verified

Repeated the 40 minute observation after the change and tracked near-miss reports for six weeks.

People crossing the aisle at the door, 40 minute observation, morning peak: 22 of 31 (71%) before, 3 of 31 (9.7%) after.

People crossing the aisle at the door, before and after the fixPeople crossing the aisle at the door, 40 minute observation, morning peak: 22 of 31 (71%) before, 3 of 31 (9.7%) after.Before71%After9.7%
People crossing the aisle at the door, 40 minute observation, morning peak: 22 of 31 (71%) before, 3 of 31 (9.7%) after.

Take to your own problem: 5 Whys template and near miss report form.

Chain 8 of 10 · Safety · 2 whys and a branch

Wash area: four slips in eight weeks with two causes

Team:
EHS lead, sanitation supervisor and maintenance technician
Problem:
Four slips on the wet floor in the wash area in eight weeks. One person lost a day of work.

The chain, with the evidence for each answer

  1. Why did people slip?

    There was water where they walked.

    Evidence, record: Incident forms: all four slips at or near the two hoses and the floor drain.

  2. Why is there water on the walkway?

    Water is arriving from the hose couplings and is not leaving through the drain.

    Evidence, observation: Watched a wash shift: both couplings drip. A drain test with a measured 10 litres took 95 seconds to clear on the near side.

The chain branches at why 2

Branch A: hose couplings leaking. Branch B: a partly blocked drain grate.

Evidence: Replacing one coupling stopped the dripping at that hose only. The grate had trim waste across about half of its slots.

The wrong turn: following only one branch

Stopping at: Branch A only: the couplings leak. The fix would be: Replace the couplings and close the problem.

What happens: Water would still pool at the drain. Each branch wets the walkway on its own, so slips would fall but not stop.

Verified cause

Two causes. Coupling seals are replaced only when someone reports a leak. The grate has no cleaning step, so trim waste blocks half the slots.

Read back: Seals are not replaced on a schedule and the grate is not cleaned, so water leaks and does not drain, so the walkway is wet, so people slip.

Countermeasure

Put coupling seals on the weekly PM list. Add grate cleaning to the end-of-shift sanitation checklist with a photo standard. Fit anti-slip matting at the two hose positions as a second layer.

How it was verified

Repeated the 10 litre drain test after the cleaning step started and every week for a month, and counted slips over the following eight weeks.

Time for 10 litres to clear at the near-side drain, After the grate cleaning step started: 95 seconds before, 28 seconds after, a change of -71%.

Time for 10 litres to clear at the near-side drain, before and after the fixTime for 10 litres to clear at the near-side drain, After the grate cleaning step started: 95 seconds before, 28 seconds after, a change of -71%.Before95After28
Time for 10 litres to clear at the near-side drain, After the grate cleaning step started: 95 seconds before, 28 seconds after, a change of -71%.

Take to your own problem: 5 Whys template and safety observation card template.

Chain 9 of 10 · Maintenance · 3 whys

Maintenance: a conveyor motor bearing failed three times in five months

Team:
Maintenance supervisor, two maintenance technicians and a reliability engineer
Problem:
The drive-end bearing on the same conveyor motor failed three times in 150 days. Each failure stopped the line for about three hours.

The chain, with the evidence for each answer

  1. Why does the bearing fail?

    It overheats and the grease breaks down.

    Evidence, measurement: Bearing housing temperature on the running motor: 78 C, against 61 C on the identical motor beside it.

  2. Why does it run hot?

    The bearing housing is overfilled with grease.

    Evidence, observation: Opened the housing during a stop: the cavity was packed full and grease had pushed into the motor side.

  3. Why is it overfilled?

    The PM task says to grease until fresh grease shows at the relief, with no quantity, and the two grease guns deliver different amounts.

    Evidence, record: PM card wording read as written. Weighed 10 strokes from each gun: 15 g and 50 g.

The wrong turn: jumping to a solution

Stopping at: Why 1: the bearing overheats, so it must be the wrong bearing. The fix would be: Buy a higher-grade bearing for the motor.

What happens: The new bearing is greased the same way and overheats too, at a higher price. The cause is how it is greased.

Verified cause

The grease task has no quantity, and the grease guns are not matched.

Read back: The card has no quantity and the guns differ, so the housing is overfilled, so the bearing runs hot, so the grease breaks down, so it fails.

Countermeasure

Set the grease quantity from the bearing maker's guidance and write it on the PM card in grams. Mark each gun with its grams per stroke and put the stroke count in the task. Take a housing temperature at each PM.

How it was verified

Measured housing temperature at each weekly check for eight weeks after the housing was refilled to the new quantity, and tracked bearing stops.

Drive-end housing temperature, Weekly checks after the grease change: 78 degrees C before, 62 degrees C after, a change of -21%.

Drive-end housing temperature, before and after the fixDrive-end housing temperature, Weekly checks after the grease change: 78 degrees C before, 62 degrees C after, a change of -21%.Before78After62
Drive-end housing temperature, Weekly checks after the grease change: 78 degrees C before, 62 degrees C after, a change of -21%.

Take to your own problem: 5 Whys template and maintenance job plan template.

Chain 10 of 10 · Maintenance · 5 whys

Maintenance: a hydraulic press stopping on motor overload

Team:
Maintenance technician, maintenance planner and press operator
Problem:
A hydraulic press tripped its motor overload six times in two weeks. Each trip cost about 40 minutes.

The chain, with the evidence for each answer

  1. Why did the press stop?

    The motor overload tripped.

    Evidence, record: Control panel fault log: six overload trips, all under load.

  2. Why did the overload trip?

    The pump motor was drawing more current than rated.

    Evidence, measurement: Clamp-on ammeter at the pump motor: 31 A under load. The nameplate rating is 24 A.

  3. Why is the motor drawing more?

    The pump is working against a blocked suction strainer.

    Evidence, observation: Strainer pulled and inspected: about 70% of its area covered with fine debris.

  4. Why is the strainer blocked?

    Debris is entering the tank through an access cover that has no gasket.

    Evidence, observation: The cover shows no gasket. A tank oil sample contained fine metal and paper fibres.

  5. Why is the gasket missing?

    The tank clean procedure has no step to refit and inspect the cover, and no sign-off.

    Evidence, record: Procedure read as written: 11 steps, ending at "refill the tank". The tank was cleaned five weeks before the first trip.

The wrong turn: stopping at the symptom

Stopping at: Why 2: the motor draws too much current. The fix would be: Fit a higher-rated overload relay, or replace the pump.

What happens: A higher relay hides the problem until the motor or pump is damaged. A new pump pulls on the same blocked strainer and fails sooner.

Verified cause

The tank clean procedure does not cover refitting the cover gasket.

Read back: The procedure does not cover the cover gasket, so debris entered the tank, so the strainer blocked, so the pump drew more current, so the overload tripped, so the press stopped.

Countermeasure

Clean the tank and strainer, fit a new gasket and change the oil. Add refit and inspect the cover, with a second person's initials, to the tank clean procedure. Add a pump current reading to the monthly PM.

How it was verified

Read pump motor current at each monthly PM and counted overload trips over the next four weeks of production.

Pump motor current under load, After cleaning and the gasket fix: 31 amps before, 23 amps after, a change of -26%.

Pump motor current under load, before and after the fixPump motor current under load, After cleaning and the gasket fix: 31 amps before, 23 amps after, a change of -26%.Before31After23
Pump motor current under load, After cleaning and the gasket fix: 31 amps before, 23 amps after, a change of -26%.

Take to your own problem: 5 Whys template and breakdown analysis report.

Nothing on this page replaces your site's own safety rules, lockout procedure or quality system. For a safety event with serious consequences, use a full investigation, not one chain.

Where 5 whys chains go wrong

Five wrong turns account for most weak chains. Each is shown in at least one example above.

  1. Stopping at human error

    What it looks like: The chain ends with "the operator skipped the step", "the pedestrian was not paying attention" or "needs retraining".

    Why it happens: A person is easy to name and needs no further work. It also feels like an answer, because a person did do the thing.

    What to do: Treat the error as a link in the chain, not the end of it, and ask what made it easy or likely: the instruction, the layout, the tool, the time, the check that was not there. The RCA2 report lists one or more individuals identified as causing the event among the warning signs of an ineffective review (page 20).

    See: A fastener left loose on variant B units, Cartons shipped with the wrong product label, Five forklift near-misses at one aisle crossing

  2. Jumping to a solution

    What it looks like: By the second why someone says "so we need a better bearing" or "raise the setpoint", and the remaining whys are filled in to fit.

    Why it happens: Someone arrived with a fix in mind, and the chain becomes a way to approve it.

    What to do: For every why, write down the evidence and who saw it. If the only support is that it sounds right, it is a guess to test, not an answer. Run the trial before you buy anything.

    See: Short shots on press 7 after weeks of running well, Parts pass in the plant and fail at the customer, A conveyor motor bearing failed three times in five months

  3. Following only one branch

    What it looks like: The chain is one tidy line, yet the problem keeps coming back at a lower rate after the fix.

    Why it happens: Real problems often have more than one cause. A single line of whys finds one, and the team stops there.

    What to do: At each why, ask whether another answer is also true. Write both branches down and test each. Fix every branch the evidence confirms, and use the counts to see how much each contributes.

    See: Orders leaving a packing line late, with two causes, Four slips in eight weeks with two causes

  4. Going in circles

    What it looks like: The last why restates the first: "burrs got through because inspection missed them", "inspection missed them because they were not caught".

    Why it happens: Each answer rewords the problem, so nothing new is learned and no evidence can be collected.

    What to do: A why must name something different that you can go and look at. When the answers loop, change the question: ask separately why the defect was made and why it was not caught.

    See: Burrs on parts that reached the customer

  5. Stopping at the symptom

    What it looks like: The chain ends at a part: "the overload tripped", "the pump failed". The part is replaced or reset and the failure comes back.

    Why it happens: The part is the first thing that can be fixed on the day, and it brings the line back up.

    What to do: Keep asking until the answer is something in how the work, the equipment or the system is set up. Taiichi Ohno's stopped machine is the classic case: without repeatedly asking why, managers would simply replace the fuse or pump and the failure would recur (Lean Enterprise Institute lexicon).

    See: A hydraulic press stopping on motor overload

How to verify a root cause, before you fix it and after

Three checks, in this order. A cause that fails one of them is a candidate, not a root cause.

  1. Read it back

    Start from the cause and read each answer with "so" until you reach the problem. If any link needs a "maybe" or an "and also", it is weak. Each example above ends with that read-back line.

  2. Test the cause on its own

    Change or restore the cause and see whether the problem moves, before the full fix if you can. Re-tighten the terminal and read the current. Load the wrong roll on purpose. Measure the motor with the strainer clean. A cause you cannot test is a theory.

  3. Measure the result

    Take the measure from the problem statement, run it for a set window and compare it with the baseline. Count the same thing the same way. If the number does not reach the target, a second branch is still open: go back to the chain.

Where the method comes from

The method and the cautions follow these two sources, both opened on the day this page was written. The chains, plants and counts are ours.

Free templates for your own chains

The blank 5 whys sheet, and the forms the cause and the countermeasure go into next.

Go deeper

Key terms

Lean glossary

Doing this in LeanSuite

FAQ

Questions about 5 whys

More examples

See all example galleries, the fishbone diagram examples, the worked A3 examples, the root cause analysis guide and the lean glossary.

Chains that end in a notebook?

See on a call how LeanSuite takes a problem from the first why to a tracked corrective action, with the evidence attached and the result checked.

Schedule Demo