New to this topic?
We recommend reading these guides first to get the most out of this one:
Lean Manufacturing
Zero
Defects Goal
3
Device Types
1961
Coined by Shingo
100%
Inspection at Source

What Is Poka-Yoke?

Poka-yoke (“POH-kah YOH-keh”) is Japanese for “mistake-proofing.” Developed by Shigeo Shingo at Toyota, it refers to any device or process change that makes it physically impossible — or immediately obvious — to make an error.

The insight behind poka-yoke is simple: people will make mistakes. Systems should not let mistakes become defects. Instead of relying on inspection to catch errors after they happen, design the process so errors cannot happen.

Everyday Poka-Yoke

You already use poka-yoke every day. A USB-C cable only fits one way. Your car will not shift out of park unless you press the brake. A microwave stops when you open the door. The best poka-yoke is invisible — you never even realize it is preventing an error.

Three Types of Poka-Yoke Devices

TypeHow It WorksStrengthShop Floor Example
PreventionMakes the error physically impossibleStrongest — eliminates the defect entirelyAsymmetric fixture that only accepts the part in the correct orientation
DetectionDetects the error immediately and stops the processStrong — catches defects at the sourceSensor that stops the machine if a component is missing before assembly
WarningAlerts the operator to an abnormal conditionWeakest — relies on human responseLight or buzzer when torque is out of spec
Prevention
❯
Detection
❯
Warning
Always prefer prevention over detection, and detection over warning

How to Implement Poka-Yoke

Identify the ErrorUse root cause analysis to find where errors occur. Pareto your defect data. Focus on the top 3 defect types — they likely cover 80% of your quality losses.
Find the Root CauseAsk 5 Whys. What condition allows this error? Is it a confusing sequence? Similar-looking parts? Missing feedback? The root cause tells you where to put the device.
Design the SolutionPrefer prevention over detection. Use physical shapes, colors, sensors, interlocks, or sequence controls. The simpler and cheaper, the better. Many great poka-yoke devices cost under $50.
Test and ValidateTry to defeat your own device. If an operator can still make the error, iterate. The standard is zero defects reaching the next process.
StandardizeUpdate standard work to include the device. Train operators. Add to maintenance checks. Share the solution across similar processes.

Practical Examples

ErrorPoka-Yoke SolutionTypeCost
Part installed backwardsAsymmetric locator pinsPrevention$20-50
Missing fastenerParts counter in bin; machine will not cycle if count is wrongDetection$100-500
Wrong material loadedColor-coded containers matching color-coded stationsPrevention$10-30
Operator skips a stepSequence-locked controls — step 3 is not possible until step 2 is confirmedPrevention$200-1000
Over-torque on fastenerTorque-limited tool that clicks and stopsPrevention$50-300
Label applied upside downLabel dispenser with orientation guidePrevention$30-100
✅ Quality Culture
  • Every defect triggers a poka-yoke investigation
  • Operators suggest error-proofing ideas
  • Simple, cheap, physical devices preferred
  • Quality built in, not inspected in
❌ Inspection Culture
  • Rely on final inspection to catch defects
  • Blame operators for mistakes
  • Over-engineered solutions that operators bypass
  • Accept a “normal” defect rate

🎯 Key Takeaway

The cheapest defect is the one that never happens. Poka-yoke shifts your quality strategy from “inspect and sort” to “prevent and confirm.” Start with your top defect, apply 5 Whys, and design a device that makes the error impossible. Most effective poka-yoke costs less than $100 and pays for itself in one shift.

Interactive Demo

Try an assembly task with and without error-proofing. See how poka-yoke devices prevent defects.

⚡
Try It Yourself
Poka-Yoke Error-Proofing
▼
Try assembling 4 steps in order. Without poka-yoke, wrong sequences produce defects. Toggle poka-yoke types to see how each prevents errors. Try clicking steps out of order!
No Poka-Yoke: No error-proofing. Any step can be done in any order.
Assembly Steps (click to perform)
0
Total Attempts
0
Errors Made
0
Blocked by Poka-Yoke
0%
Defect Rate
Ready for the full knowledge check? Test your understanding with guided scenarios and data export.
PROTake the Pro Knowledge Check →
Keep Learning
📖
Guide
Jidoka
📖
Guide
FMEA
📖
Guide
Cost of Quality
🎓 Part of the IE Methods & Tools track — view all 15 guides →
Free forever · Every feature included

Stop reading, start modeling

Model your process flow, run simulations, optimize staffing with TOC math, and test your knowledge with 107 interactive checks — all in one platform.

Open Workbench → Pro Knowledge Check

Take this to a room

The running order

For a floor team with a recurring defect. They should leave with one candidate device for their top defect and a preference for prevention over detection.

7 beats · 8 min
  1. 1

    Start with the ones they already use

    A USB-C cable only fits one way. A car will not leave park until you press the brake. A microwave stops when the door opens.

    • Nobody had to be trained not to make those mistakes.
    • The best error-proofing is invisible - you never notice it working.
    • This is not a quality programme. It is a design choice.

    Ask the room What is something here that is impossible to get wrong, and why?

  2. 2

    Three kinds, in strength order

    Prevention, detection, warning. Always reach for the leftmost one you can afford.

    • Prevention: the error is physically impossible. Asymmetric fixture, one orientation only.
    • Detection: the process stops the moment the error happens. Sensor on a missing component.
    • Warning: a light or a buzzer. Weakest, because it relies on a human noticing.
  3. 3

    They are cheap

    The price list surprises people, and it is the argument that unblocks the budget conversation.

    • Part backwards: asymmetric locator pins, 20 to 50 dollars.
    • Wrong material: colour-coded containers matching stations, 10 to 30 dollars.
    • Over-torque: torque-limited tool that clicks and stops, 50 to 300.
    • Skipped step: sequence-locked controls, 200 to 1,000.
  4. 4

    Find the condition, not the person

    The device goes where the root cause is, so you need the root cause first.

    • Pareto the defect data. The top three types are usually 80 per cent of the loss.
    • Then 5 Whys: what condition allows this error? Confusing sequence, similar-looking parts, no feedback?
    • If the answer is operator error, you have not finished.
  5. 5

    Try to defeat your own device

    An error-proofing device that an operator can work around is a decoration.

    • Attempt the error deliberately. If it is still possible, iterate.
    • Over-engineered devices get bypassed, which is worse than no device.
    • The standard is zero defects reaching the next process.

    Ask the room What would someone here do if this device slowed them down?

  6. 6

    Then make it stick

    A device that is not in the standard and not on a maintenance check will quietly stop working.

    • Update standard work to include it.
    • Add it to maintenance checks - sensors drift and fixtures wear.
    • Share the solution to every similar process, not just this one.
  7. 7

    What we do next

    Take our top defect, run 5 Whys, and design something under a hundred dollars this month.

    • Pull the defect Pareto for the last quarter.
    • Take number one and find the condition that permits it.
    • Prefer a physical shape over a sensor, and a sensor over a light.
    • Most effective devices pay for themselves in a shift.

    Ask the room What is our number one defect, and what condition lets it happen?

Related Guides

📊 Six Sigma
Data-driven quality improvement
🔍 Root Cause Analysis
Find the real cause of defects
✅ First Pass Yield Calculator
Measure quality at the source