Latest / The Operations Podcast with Fexingo: Process, People, and Profit in Modern Business / How Toyota's Kanban System Quietly Runs the World
Transcript
- Lucas: So here's a weird fact. Toyota's kanban system — the little card that tells a worker to make one more part — was inspired by a trip to a Piggly Wiggly supermarket in the United States in the 1950s. Luna: Wait — Piggly Wiggly? The grocery chain? How does a supermarket inspire one of the most copied operations tools in history? Lucas: Exactly. Taiichi Ohno, the engineer behind the Toyota Production System, noticed that American supermarkets didn't stockpile mountains of inventory. They replenished shelves based on what customers actually took. Each item had a reorder point, and when a customer grabbed a can of soup, that triggered a signal to restock one can. Pull, not push. Luna: Pull, not push — that's the core of kanban. Instead of saying 'let's produce 10,000 units and then try to sell them,' you only produce what the next step in the process demands. Lucas: Right. In Toyota's system, that signal is a physical card — the kanban. It's literally a piece of paper or plastic that travels between workstations. When a downstream process uses a part, it sends the kanban back upstream, authorizing production of exactly one more. No kanban, no production. Luna: I've always found that rule fascinating. It means you can't just produce for the sake of keeping machines running. You have to wait for a signal. Lucas: And that waiting is intentional. One of the biggest mistakes companies make when they copy kanban is they think it's about speed. It's not. It's about deliberate delay. You let inventory shrink until it almost breaks, then you fix the bottleneck. That's how you expose waste. Luna: Let's get specific. What's a real number from Toyota or a supplier that shows the impact? Lucas: Sure. A Japanese auto parts supplier called Denso — which was actually part of Toyota originally — implemented kanban across its radiator plant in the 1970s. Within two years, they reduced work-in-progress inventory by 87 percent. That's not a typo. Eighty-seven percent. They went from having thousands of half-finished radiators stacked on the floor to just a few hundred in the entire system. Luna: And what happened to productivity? Usually when you slash inventory, you'd expect chaos — machines starving for work, workers idle. Lucas: Productivity actually went up — about 30 percent over the same period. Because they weren't spending time moving, stacking, and counting excess inventory. They were spending time making good parts. The kanban forced them to solve the real problems instead of hiding them behind piles of stock. Luna: So that's the theory. But kanban has spread way beyond car factories. I'm thinking of hospitals, software teams, even cloud computing. Lucas: It's a good point. Let's take a hospital operating room. A classic problem: surgeons want all possible instruments available for every case, 'just in case.' That leads to massive sterile supply inventories, expired items, and waste. Luna: And that's where a kanban-style system comes in. Some hospitals now use two-bin systems for surgical supplies. When the first bin is empty, you use the second bin, and the empty bin triggers a reorder. That's exactly the same logic as a kanban card. Lucas: Exactly. One study from Virginia Mason Medical Center in Seattle — which famously adopted the Toyota Production System — showed that using kanban for surgical supplies reduced inventory by 53 percent and cut stockouts by 74 percent. Nurses spent less time hunting for supplies and more time with patients. Luna: I want to go back to something you said earlier — that most companies implement kanban wrong. What's the most common mistake? Lucas: The most common mistake is treating kanban as a scheduling system rather than a waste-reduction system. Companies calculate how many cards they think they need based on historical demand, and then they just maintain that level. They never reduce the number of cards deliberately to force improvement. Luna: So they miss the whole point. The number of kanban cards is a dial. You turn it down, see what breaks, fix it, turn it down again. Lucas: Exactly. Taiichi Ohno famously said, 'Without a standard, there can be no improvement.' The kanban is that standard. You set a rule — 'we will only have ten cards in this loop' — and then you challenge the team to make it work with nine, then eight. Each time, you find a new way to reduce waste. Luna: That's so counterintuitive to typical management thinking, which is often 'let's add more resources to meet demand.' Kanban says, 'let's deliberately starve the system to find the fat.' Lucas: It's a discipline. And it's why kanban has survived for over 70 years while other management fads have come and gone. It's simple, visual, and brutally honest. If a kanban card is sitting on a board without moving, you can see that you have a problem. Luna: Okay, but is kanban still relevant in the age of AI and real-time data? Why use a physical card when you could have a digital signal? Lucas: Great question. In many modern settings, kanban has gone digital. Amazon's fulfillment centers use a digital version — when an item is scanned out of a bin, that signal propagates to the suppliers. In software development, tools like Jira or Trello are essentially digital kanban boards. Luna: But the principle is the same. The card — whether physical or digital — is the token that limits work-in-progress. And that's the key insight. Lucas: Exactly. In fact, one of the most interesting applications I've seen recently is in cloud infrastructure. Companies like Netflix use a kanban-like approach to manage server capacity. Instead of provisioning servers based on forecast, they have a pool of 'cards' that represent available capacity. When a team needs a server, they take a card. When they're done, they return it. No card means no new server until capacity is freed up. Luna: That's a brilliant way to prevent runaway cloud costs. So many companies just spin up virtual machines and never turn them off. A kanban system forces you to justify every resource. Lucas: And it scales. One of the great things about kanban is that it works whether you're managing 10 parts or 10,000. The rules are the same. You limit the number of cards, and you pull when you need something. Luna: But there's a dark side, right? Kanban can be brutal on workers if it's implemented as a speed-up tool rather than a learning tool. Lucas: Absolutely. When Toyota introduced kanban, they paired it with a culture of respect for people and continuous improvement. If you just cut inventory and tell workers to 'figure it out,' you get stress and burnout, not improvement. The whole point is to surface problems so that teams can solve them together. Luna: I've seen that happen in a Danish insulin factory run by Novo Nordisk. They use kanban for the entire supply chain of insulin cartridges. The cards are color-coded by urgency, and workers have the authority to stop the line if a shortage looms. It's not about blame — it's about visibility. Lucas: That's a great example because the consequences of a stockout in insulin production are literally life-threatening. Kanban gives them a way to manage that risk without overproducing and wasting expensive biological material. Luna: It also creates a shared language. Everyone from the shop floor to the logistics team can look at the kanban board and know the state of the system in ten seconds. Lucas: That visual management piece is underrated. When I've visited factories that use kanban well, the first thing you notice is that there's no hiding. The board is front and center. Every card is accounted for. You can literally see the flow of work. Luna: You know, this kind of deep dive into operations is exactly why I love working on this show. We get to explore these elegant solutions that most people never think about. Lucas: And that's why we keep it ad-free. We want these conversations to be useful to you without any noise. If they've helped you think about your own work in a new way, and you'd like to support that, there's a simple way — buy me a coffee dot com slash fexingo. It's not a subscription, no perks, just a way to keep the show independent. Luna: Yeah, it's a small thing that makes a big difference. We don't run ads, so listener support is what keeps the lights on. Seriously, every contribution helps. Lucas: Anyway, back to kanban. I think one of the most powerful things about this system is that it's endlessly adaptable. You can use it in a bakery, a hospital, a software team, or a rocket factory. Luna: Let's end with a specific takeaway. If someone wants to try kanban in their own operation, what's the first step? Lucas: Map the value stream. Identify one loop — one product or one process — and start with just one card. Let one unit flow through the entire system. Then add a second card, and a third, but never more than you can handle. Watch where the cards pile up. That's where your problems are. Then fix those problems before adding more cards. Luna: Start small, make it visual, and treat every blockage as a learning opportunity. That's kanban in a nutshell. Lucas: And remember, the goal isn't to have zero cards. It's to have just enough to keep the system flowing while exposing the next problem. It's a system designed for continuous improvement, not perfection.