The andon cord is the most famous quality tool in manufacturing. Made famous by Toyota, it gives every operator the authority to stop the production line when they detect a problem. The principle is simple: the people closest to the process are the first to know when something is wrong, and stopping immediately prevents defects from accumulating.

I implemented andon authority at a European automotive plant. It was the most controversial decision I made in my career, and it produced the most profound cultural change I have ever witnessed.

The resistance

The plant manager was against it. "If operators can stop the line whenever they want, they will stop it constantly. We will never hit our production targets." The production supervisor was against it. "Operators are not qualified to decide what is a quality problem." The finance director was against it. "Every minute of line stoppage costs €2,800. This will destroy our margins."

Only the operators were not consulted. Nobody asked them whether they wanted the authority. Because the assumption — shared by every manager in the room — was that operators would abuse it. That given the choice, workers would stop the line to avoid work. This assumption reveals more about management's view of workers than it does about workers themselves.

When you do not trust your operators to stop the line, you are saying you do not trust them to care about quality. If you do not trust them to care about quality, why did you hire them?

The implementation

We installed andon buttons at every station — a physical button that, when pressed, triggered a yellow light and a musical tone. The team leader had ten minutes to respond. If the issue was not resolved in ten minutes, the line stopped and a red light replaced the yellow.

I trained every operator on what constituted a legitimate andon pull: visible defects, equipment malfunctions, safety hazards, missing or incorrect material, anything that felt wrong. The criteria were deliberately broad. I wanted operators to err on the side of stopping.

I also established a rule that I repeated in every meeting for six months: no operator would ever be disciplined for pulling the andon. Ever. Even if the stoppage turned out to be unnecessary. The only thing that would trigger a conversation was not pulling the andon when a defect was visible.

The first three months

Month one: the line was stopped 47 times. The production manager was apoplectic. I reviewed every stoppage. Of the 47, thirty-two were legitimate quality concerns that prevented defective parts from continuing down the line. Ten were equipment issues that, left unaddressed, would have produced defects within hours. Five were false alarms — operators being cautious, exactly as I had asked.

The cost of the stoppages: €68,000 in lost production time. The cost of the defects prevented: approximately €210,000 based on the historical cost of similar escapes. Net savings: €142,000 in the first month.

Month two: the line was stopped 31 times. Operators were learning to distinguish real issues from false alarms, and team leaders were responding faster. The average stoppage duration dropped from eight minutes to four minutes. The defect escape rate — parts that reached the customer — dropped by 40 percent compared to the same month the previous year.

Month three: the line was stopped 19 times. The andon pulls were stabilising at a sustainable rate. More importantly, the nature of the pulls changed. Instead of pulling for visible defects — a burr, a scratch, a misalignment — operators were pulling for process anomalies that preceded the defects. "The machine sounds different." "The material feels different." "The cycle time is two seconds longer than normal." These were early warning signals that no inspection system could detect, because they relied on the operator's tacit knowledge of the process.

The cultural shift

The quality improvement was significant — defect rates dropped by 55 percent in six months. But the cultural shift was more profound. Operators started talking about quality differently. They stopped saying "that is not my job" and started saying "I caught something on line three today." They proposed improvements. They challenged engineers about process changes that did not feel right. They engaged with the quality system as participants, not as subjects.

The production supervisor, who had been the most vocal opponent, came to me after four months and said: "I was wrong. The operators are finding things we would have missed. And they are faster than any inspection system because they know the process."

The plant manager, who had predicted production targets would be missed, found that after the initial disruption, productivity recovered to previous levels and then exceeded them — because the rework and re-inspection burden had dropped dramatically. The line was running more, not less, because it was running right.

What I learned

The andon implementation taught me three things that have shaped my approach to quality leadership ever since:

Trust is a quality tool. When you trust operators to detect and respond to problems, you get a detection system that is faster, more sensitive, and more adaptive than any technology. When you do not trust them, you get defects that pass through five inspection stations and reach the customer.

The people closest to the process know things you do not. An operator who has run the same machine for five years has tacit knowledge — sound, vibration, smell, rhythm — that no SPC chart captures. Giving them a channel to express that knowledge transforms it from anecdote into prevention.

Authority changes behaviour. When operators were given the authority to stop the line, they took ownership of quality. They stopped being passive participants in a system that inspected their work and became active guardians of the process. This shift in ownership did more for quality than any procedure, training programme, or audit I have ever implemented.

Give your operators the andon cord. Train them, support them, and trust them. The first month will be uncomfortable. The second month will be better. By the sixth month, you will wonder why you ever ran a line any other way.