TPS Foundations

Jidoka

Stop when something's wrong. Don't pass the defect forward.

Updated
·
4
min read
Definition

What is Jidoka?

Jidoka is the lean principle of building quality into the process by stopping production the moment a defect or abnormality is detected. Sometimes translated as autonomation or "automation with a human touch," jidoka gives any worker the authority to halt the line so the problem can be fixed at its source. It is the second pillar of the Toyota Production System.

Jidoka is the second pillar of the Toyota Production System and the most poorly understood one. It is usually translated as "autonomation" or "automation with a human touch," but neither phrase captures what makes jidoka work: a cultural rule that any worker, at any time, can stop production when something is wrong, and the organization treats that stop as valuable information rather than disruption. Without that cultural rule, jidoka is decoration.

"Stopping the line costs an hour. Letting the defect through costs a customer."

How jidoka works

Jidoka has two layers. The mechanical layer is automatic detection: machines that can sense their own abnormalities (a misfeeding part, a temperature outside spec, a missing fastener) and stop themselves before producing a defect. Toyota's first automated loom, built by Sakichi Toyoda in the 1920s, would stop the moment a single thread broke. That was the original jidoka.

The cultural layer is human authority. Beyond the machines, every operator has the authority to stop the line. In a Toyota plant, this is done with the andon cord, a pull-rope or button that any worker can engage when they see something wrong. The cord triggers a light and a buzzer, signalling a team leader to come help. The line keeps moving briefly while the leader assesses. If the issue cannot be resolved at a fixed point, the line stops at the next natural breakpoint until the cause is identified.

The discipline that makes jidoka work is finding the root cause, not the proximate cause. If an operator stops the line because a part is misfeeding, the wrong response is to clear the jam and restart. The right response is to ask: why did it misfeed? Was the upstream operation producing slightly out-of-spec parts? Is the feeder mechanism worn? Has the operator's standard work been changed in a way that lets the wrong orientation get past? The five whys discipline is jidoka's investigation tool. Without it, the same defect recurs the next shift.

Where jidoka fits on a small shop floor

Imagine a 15-person plastics injection molding shop running three SKUs for a kitchenware brand. The shop is profitable but customer returns are costing about 4 percent of revenue, almost all from cosmetic defects (sink marks, slight color variation) that survive into shipment. The owner has been blaming the molders.

A jidoka diagnosis would not blame anyone. It would install a simple rule: any molder who pulls a part off the press and sees a defect immediately calls the lead operator. The lead operator comes within two minutes, looks at the part, and asks: is this defect from the current shot, or has it been there for the last 20 shots? If 20 shots, the press stops and the team investigates. The hold time costs the shop maybe an hour of production. The find: a worn gate that has been producing borderline parts for two weeks. Replaced. Defects drop from 4 percent to under 1 percent over the next month.

This is jidoka at small scale. No expensive sensor equipment. No quality management system. Just the standing rule that defects stop the work, and the team has authority to investigate without management approval. The hardest part is the first few times: operators hesitate to stop because they are worried about hitting production targets. After the second or third time they stop and management says thank you, the hesitation goes away.

Common mistakes with jidoka

  • Installing the equipment without the authority. A stop button operators fear to use is decoration.
  • Fixing the symptom and restarting. Jidoka requires root-cause investigation. Clearing the jam without finding why the jam happened guarantees the next one.
  • Rewarding "fewest stops." The metric that matters is fewest defects passed downstream. Rewarding fewest stops trains operators to hide problems.
  • Treating stops as failures. Stops are the system working as designed. A shop with zero stops is either perfect or, more likely, has trained its workers not to pull the cord.
  • Skipping the post-stop standard-work update. Once a root cause is found, the standard work should be updated to prevent recurrence. Without the update, the lesson is lost when the operator who found it moves on.

Jidoka and related Lean tools

Jidoka is the second pillar of the Toyota Production System, paired with just-in-time. Its mechanical implementation is most often autonomation, and its visual signal mechanism is the andon cord or board. The investigative tool that makes jidoka stops productive is poka-yoke, which prevents the same defect from being possible to make again.

Common questions

The questions we hear most about this term.

How is Jidoka different from Just-in-Time?
JIT and jidoka are the two pillars of TPS and they answer different questions. JIT answers "when should we make it?" The answer is: only when the next station needs it. Jidoka answers "what do we do when something goes wrong?" The answer is: stop everything, fix the problem at the source, then resume. The pillars are designed to work together. JIT keeps inventory low, so a defect cannot be buried in piles of WIP. Jidoka stops the defect from being made twice.
Is Jidoka the same as Autonomation?
Yes and no. Autonomation is one English translation of jidoka, coined to capture the idea of "automation with a human touch": a machine that can detect a problem and stop itself, plus an operator who can intervene and fix the underlying cause. The Japanese word jidoka is broader. It includes the autonomation idea but also the cultural authority any worker has to stop the line, the daily problem-solving habit, and the discipline of fixing the cause rather than the symptom. Most lean texts use the terms interchangeably; technically jidoka is the bigger concept.
How does Jidoka work without bringing the whole shop to a stop?
Through smart line design. In a well-designed jidoka line, stopping happens at small intervals and gets fixed quickly. The andon cord is pulled, the upstream and downstream stations briefly idle, a team leader walks over within seconds, and either the cord gets released (problem solved) or a planned stop happens at the next natural breakpoint. The first time a shop installs jidoka, stops feel disruptive. After a few weeks, the team has fixed enough root causes that stops become rare and short. The early disruption is the system working as designed.
What are common mistakes when implementing Jidoka?
The biggest mistake is installing the equipment without granting the authority. Andon cords or stop buttons that operators are afraid to use are decorative. The second mistake is fixing the symptom and restarting, instead of finding the root cause. Jidoka requires a problem-solving discipline (often [five whys](https://arda.cards/glossary/five-whys)) so the same defect does not recur. The third is metric-gaming: rewarding teams for "fewest stops" instead of "fewest defects passed downstream." That trains operators to hide problems.
What does Jidoka look like on the shop floor of a small manufacturer?
Smaller than people expect. A 20-person shop running jidoka might have a simple rule: any operator who sees a defect raises their hand. The shift lead comes over within a minute. If the problem can be solved in two minutes, they fix it and the operator restarts. If it cannot, the operation pauses and the team huddles to figure out the cause before resuming. There is no infrastructure beyond the rule and the discipline. The hard part is not the equipment. It is the standing authority every worker has to stop.

Ditch the whiteboards and spreadsheets.

Same-day setup. No distributor lock-in. Zero stockouts. Top teams double revenue in 9 months.