Latest / The Operations Podcast with Fexingo: Process, People, and Profit in Modern Business / How Toyota's Andon Cord Transformed Factory Quality
Transcript
- Lucas: So we hit episode 100 of this show. And I wanted to do something that ties back to the very first thing I ever studied in operations — the Toyota Production System. Luna: The andon cord. That was the first thing that came to mind for me too. Lucas: Exactly. The andon cord is probably the most copied operations tool in history, and also the most misunderstood. Every factory consultant talks about it. Almost no one implements it correctly. Luna: So what does it actually look like when it's done right? Not the theory — the practice. Lucas: Let's go to Toyota's plant in Georgetown, Kentucky. That plant builds the Camry and the RAV4. Every assembly station has an overhead cord — it looks like a pull chain for a ceiling fan. Any team member can pull it if they see a defect, a part that doesn't fit, a torque wrench that didn't click. Luna: And when they pull it, the line stops. That's the part that scares most managers. Lucas: It scares them because they think about lost throughput. But Toyota measures three things to know if the andon system is healthy. First: the average time from pull to line stop. Target is 12 seconds or less. Second: the percentage of pulls that actually result in a line stop. Target is above 95 percent. Luna: Wait — you want the line to stop almost every time someone pulls? That seems counterintuitive. Lucas: Right. Because if someone pulls and the line doesn't stop, then the cord is just a suggestion box. The whole point of andon is built into the word itself — it's a Japanese term for a paper lantern. The cord is a visual signal that something is wrong. If the line keeps moving, the signal loses meaning. Luna: So what's the third metric? Lucas: The percentage of line stops that last longer than 10 minutes. Target is under 10 percent. Because if a stop lasts longer than 10 minutes, it means the root cause isn't being addressed quickly — and that's a management failure, not a line worker failure. Luna: That flips the typical narrative. Most plants I've visited treat a line stop as a failure event. Toyota treats not stopping as the failure. Lucas: Let me give you a concrete example from the Georgetown plant. A team member installing door seals noticed the adhesive wasn't bonding properly on every third car. She pulled the cord. Line stops. A team leader arrives within 30 seconds. They trace it to a temperature fluctuation in the adhesive application nozzle. Problem fixed in eight minutes. Total downtime: eight minutes. Number of defective cars that got through: zero. Luna: Versus what happens in a typical plant — she would have flagged it at the end of the shift, the quality inspector finds it on the next day's audit, and now you've got fifty cars with peeling seals that need rework. Lucas: Exactly. And the rework cost is orders of magnitude higher than the eight minutes of downtime. But here's the thing — the andon cord is not really about the cord. It's about creating a culture where stopping is safe. Luna: That's the part that can't be copied by buying a software dashboard. Lucas: Right. I've seen factories install andon boards with big digital displays showing green, yellow, red lights. But when a worker pulls the cord, the supervisor shows up and says, 'Can you finish the job first and then we'll log it?' That's a broken andon system. The metric says green, but the behavior says red. Luna: So how does Toyota train managers to respond differently? Because the instinct to keep the line moving is deeply ingrained. Lucas: They train through a process called yokoten — horizontal deployment. New managers spend their first weeks in the plant watching andon responses. They see the team leader arrive, ask three questions: What's the problem? What's the root cause? What's the countermeasure? They're graded on speed and thoroughness, not on whether they kept the line running. Luna: And that's measured? Lucas: Yes. Every team leader has a daily scorecard. Their response time to an andon pull is tracked. If the average response time creeps above 45 seconds, that's a coaching moment. The system is designed to be self-correcting. Luna: Talk about the counterintuitive part — because I've read that Toyota actually encourages pulling the cord even for minor issues. Not just for defects. Lucas: That's the jidoka principle — automation with human intelligence. The cord should be pulled for anything that feels off. A part that's slightly harder to insert than usual. A sound that's not normal. A sequence that seems out of order. Most of those pulls don't stop the line for more than a minute. But they generate data. And over time, that data reveals patterns. Luna: Patterns that a statistical quality control chart might take weeks to catch. Lucas: Exactly. One Toyota plant in Japan found that a particular station had an andon pull rate three times higher than the plant average. The root cause wasn't the operator — it was a poorly designed part fixture that made the assembly angle awkward. They redesigned the fixture. Pull rate dropped back to normal. Problem solved before a single defect reached the customer. Luna: So the andon system functions as a real-time early warning network. But it only works if the workers trust that they won't be penalized for stopping production. Lucas: And that trust takes years to build. A plant manager at Toyota once told me that when the line stops, the first question from management is never 'Why did you stop?' It's 'What do you need to fix this?' That's a profoundly different starting point. Luna: And that's also why transplant factories — non-Toyota plants that try to copy the system — often fail. They install the hardware but don't change the management mindset. Lucas: One of the classic failure patterns I've seen is a plant that boasts 'zero line stops' for a month. In Toyota's view, that plant is dangerously out of touch. It means either problems are being hidden, or the cord is being ignored, or the system is so over-engineered that nothing can go wrong — which is impossible. Luna: Or the third option: the plant isn't producing anything complex enough to have problems. But then why do you need an andon system? Lucas: Right. The andon cord is a diagnostic tool. A healthy system generates a predictable number of stops per shift. If that number is zero, something's broken. Luna: So for episode 100, what's the one takeaway you want listeners to carry? Lucas: It's this: the andon cord is not a technology. It's a contract between management and labor. The cord says: 'We trust you to stop the line. We will support you when you do. And we will fix the problem so you don't have to stop again.' If any part of that contract is missing, you don't have an andon system — you have a decoration. Luna: And that contract is what made Toyota the most reliable manufacturer on earth. It's not the robots. It's the cord. Lucas: And it's why, 100 episodes in, we keep coming back to operations fundamentals. The fancy tools change. The human dynamics don't. Luna: Speaking of that — we've done 100 episodes without running a single ad. That's a choice we made early on, and it's only possible because some listeners have supported us directly. Lucas: Yeah, if these conversations have moved your work forward in any small way, and you want to keep them ad-free, the link is buy me a coffee dot com slash fexingo. Nothing more than that. Luna: And now back to the andon cord — because I think there's one more layer. The cord also builds problem-solving muscle. Every pull is a miniature kaizen event. Lucas: That's a great way to put it. Each pull is a 10-second disruption that trains the entire team to identify, escalate, and fix problems in real time. Over thousands of pulls, that muscle becomes automatic. And that's why Toyota can launch a new model with fewer quality hiccups than any other automaker. Luna: So if you're listening and you want to try this in your own operation — start small. One station, one cord, one shift. And measure the three metrics Lucas mentioned. Not line speed. Not output. Just response time, stop rate, and long-stop percentage. Lucas: And if that first station shows a stop rate above 95 percent after a month, you're on the right track. If it's below 50 percent, you've got a culture problem, not a tool problem. Fix the culture first. Luna: Here's to 100 episodes of operations, and to the one cord that started it all. Lucas: Cheers to that. See you next time.