#354 1966 · Toyota Motor Corporation (assembly line) · Manufacturing
Toyota built the andon cord as every line worker's daily obligation, not a permission — with a team leader sworn to arrive within seconds and thank them for pulling it
the problem
Defects rode silently down the line unflagged by the worker who spotted them
background
Under Ford-style mass production, the assembly line was sacred: keeping it moving was the metric everyone was measured on, and a worker who stopped it — even to flag a defect — risked discipline or firing. At GM's old Fremont plant, for instance, stopping the line for any reason, including quality problems, was grounds for termination. Defects were caught downstream, if at all, after thousands of identical units had already passed the flawed step.
Toyota inherited the idea of self-stopping machinery from founder Sakichi Toyoda's 1896 automatic loom, which halted itself the instant a thread broke rather than weaving flawed cloth — the root of 'jidoka,' automation with a human touch. Rebuilding postwar with none of Detroit's capital, Taiichi Ohno extended the same logic to people: Toyota installed stop-button andons on the Crown line in 1955, and by 1966, when andon was installed at the Kamigo engine plant, Toyota's own corporate history describes the system as complete.
what everyone would do
The Ford-style answer was to keep the line moving at all costs and catch defects downstream through inspection — stopping the line for any reason, including a quality problem, was grounds for discipline or firing at plants like GM's old Fremont facility. That fails because it makes the worst-positioned person to fix a defect (an inspector far downstream, with no memory of the specific moment it occurred) responsible for catching it, while punishing the one person who actually saw it happen — the line worker — for saying anything.
what they saw
Ohno saw that the worker who spots a defect in real time has the cheapest, fastest fix available if given permission to raise it immediately — the problem was never a lack of a mechanism to flag defects, factories already had ways to report issues. It was that raising one carried real professional risk, so workers rationally stayed silent. A cord alone changes nothing; what has to change is whether pulling it is treated as an obligation fulfilled or a confession of failure.
the move
Pulling the cord does not stop the line outright — it lights a signal and plays music, and a team leader, kept at roughly one per eight workers, is trained to arrive within seconds. If the team leader fixes the issue before the car reaches the end of its cycle-time station, the line never actually stops. Critically, the team leader's job is to thank the worker for pulling it, never to question or blame them — pulling the cord was built into training and management response as unconditional positive reinforcement, so it read as duty fulfilled, not failure confessed.
why it works
Pairing the cord with a team leader trained to arrive within seconds and thank the worker — never question or blame them — converts pulling it from a personally risky act into an expected, celebrated part of the job, which is why Toyota workers pull it roughly 2,000 times a week and treat that volume as evidence the system is healthy rather than failing. Because most pulls get resolved within the station's own cycle time, the line rarely actually stops, so the cost of using the signal stays low while the benefit — catching a defect at the exact moment and station it occurred, when the fix is cheapest and the cause is freshest — compounds every time it's used. NUMMI is the clean natural experiment: when GM's dysfunctional Fremont plant got the SAME cord under Toyota's incentive and response structure, absenteeism fell from 45% to 3% and quality matched Japanese output within two years — proof the cultural mechanism, not the physical device, did the work.
the payoff
The system succeeds by being pulled constantly, not rarely: at Toyota's Georgetown, Kentucky plant workers activate the andon roughly 2,000 times a week, which Toyota treats as healthy — problems surfacing, not hiding. When GM handed over its dysfunctional, strike-prone Fremont plant to Toyota to run as NUMMI in 1984 and had the same authority and response structure installed alongside the cord, absenteeism fell from about 45% to 3% and the plant matched Japanese-built quality within about two years — the clearest evidence the mechanism, not the cord, did the work.
where it breaks
Installing the cord without the response structure — no team leader obligated to arrive fast, no explicit prohibition on blaming the worker who pulled it — produces almost nothing, exactly what happened at US plants that copied the artifact: one comparison found a Ford plant's cord pulled only a handful of times against a comparable Toyota plant's roughly 2,000 weekly pulls, because workers correctly read that pulling it still carried social or professional risk. The mechanism also depends on team leaders actually having the capacity to respond within seconds — understaffing the leader-to-worker ratio turns the cord back into an empty gesture, since a slow or absent response teaches workers the same lesson punishment would: don't bother. And it only helps with problems visible to the person on the line; it does nothing for defects or risks that no line worker is positioned to notice in the first place.
what came after
Outside NUMMI, US automakers that copied the physical cord without the underlying obligation and response culture got little from it: one comparison cited by lean-manufacturing analysts found a Ford plant's cord pulled only a handful of times against Toyota's roughly 2,000 weekly pulls at a comparable plant. The contrast is now a standard teaching case in lean-manufacturing literature for why copying an artifact without its incentive structure produces nothing.
references
- [1]What Is an Andon Cord? How Toyota's Andon System Really WorksLean Blog (Mark Graban), 2012leanblog.org
- [2]Andon — TPS EncyclopediaArt of Lean, 2023artoflean.com