Latest / The Operations Podcast with Fexingo: Process, People, and Profit in Modern Business / How an Auto Parts Supplier Cut Machine Changeover Time by 85 Percent
Transcript
- Lucas: When you walk a factory floor, the moment you notice is the moment everything stops. Changeover — the time between the last good part of one batch and the first good part of the next. For decades, most plants treated it as a fixed cost. But there's a case from a stamping plant outside Detroit that cut changeover from four hours to under forty minutes. That's an 85 percent reduction. Luna: That sounds like classic single-minute exchange of die — SMED — from Shigeo Shingo. But those principles have been around since the 1980s. Why aren't more plants doing this? Lucas: Good point. The principles are well known, but implementation is where it falls apart. This particular plant — we'll call it Midwest Stamping — was supplying door panels to two of the Big Three. They had sixteen presses, and each changeover was costing them about 40 minutes of downtime per shift. By the time you add it up, they were losing nearly two hours of production per day across the line. That's roughly 15 percent of capacity sitting idle. Luna: So what did they actually change? I imagine it wasn't just one thing. Lucas: No, it was a systematic application of SMED's three steps. First, separate internal from external setup. Internal means steps that require the machine to be stopped. External means steps that can be done while the machine is running. At Midwest Stamping, they found that nearly half of the setup time was spent on external tasks that they could do in advance — gathering dies, preheating, positioning equipment. By moving those outside the downtime, they cut changeover time from 240 minutes to about 140 minutes right off the bat. Luna: So 100 minutes saved just by reorganizing what people did when. That's huge — and essentially free. Lucas: Exactly. The second step is to convert internal setup to external wherever possible. They invested in quick-release clamps instead of bolts — that cost about $200 per press. And they standardized die heights so they didn't have to adjust shims every time. That alone shaved another 30 minutes. The third step — streamline all aspects of setup — involves things like color-coded tooling racks and shadow boards so workers aren't searching for the right wrench. They also created a standardized checklist and trained a dedicated changeover team. After six months, they got changeover down to 36 minutes. Luna: Eighty-five percent reduction. What did that mean for the bottom line? Lucas: They calculated it as roughly $2.3 million in recovered production capacity per year — without any overtime or new equipment. That's pure profit from operations. And because they could switch batches faster, they reduced their batch sizes, which cut work-in-process inventory by 40 percent. That freed up floor space and reduced carrying costs. Luna: So the ROI on $200 clamps was incredible. That's the kind of story that makes me want to share this podcast with colleagues. Lucas: If these conversations have helped you think about your own operations, buy me a coffee dot com slash fexingo — that's where listener support keeps us ad-free and independent. A couple of dollars a month genuinely makes a difference. Luna: Yeah, it really does. It's how we keep doing deep dives like this without any sponsors. Lucas: So what's the lesson for other manufacturers? The biggest barrier isn't technical; it's cultural. The plant manager told me the hardest part wasn't buying clamps — it was convincing operators that a 240-minute setup was normal. Luna: How did they overcome that? Lucas: They started with one press as a pilot. The changeover team documented every step, filmed the process, and did time studies. Then they shared the results with the whole plant. When operators saw that the external steps could be done without stopping the press, they bought in. Within a month, the team had reduced setup on that one press from 240 to 90 minutes. That proof of concept spread like wildfire. Luna: So the key was making it visible. Show, don't just tell. Lucas: Exactly. And they also tied it to the operators' own metrics. They started tracking changeover time on every shift and posting it on a whiteboard. Teams that improved got recognition. It became a game — who could beat the previous best time. Within six months, the plant-wide average was under 40 minutes. Luna: What about the engineering side? Weren't there some steps that had to be internal no matter what? Lucas: Sure. The actual die change itself — lift the old die, place the new one — that's inherently internal. But SMED says you can still streamline those steps. They added a die cart with rollers so the operator didn't need a forklift. They pre-staged the next die so it was waiting at the press. They also synchronized the swap so two operators worked in parallel — one removing, one installing. That cut the internal die-change time from 18 minutes to seven. Luna: Parallel work is such a simple idea. Why is it so rare in practice? Lucas: Because most plants have a culture of 'one person per press.' But when you cross-train people and create a dedicated changeover team, you can split the work. The team at Midwest Stamping literally timed each step and rearranged the sequence so one operator did left-side tasks, the other did right-side. They also color-coded the wrenches so no time was wasted selecting tools. Luna: And the results went beyond just the machine hours. You mentioned inventory. What exactly happened there? Lucas: Because they could change over so quickly, they could run smaller batches. Instead of producing a week's worth of door panels, they could make a day's worth. That cut work-in-process inventory by 40 percent, which freed up floor space and reduced the risk of obsolescence. They also improved quality — less inventory meant less time for parts to get damaged or lost, and fewer defects carried over from one batch to the next. Luna: So it's a classic lean win: shorter lead times, lower inventory, better quality. And all from a project that cost next to nothing. Lucas: I'd say the total investment was about $15,000 in tooling, training, and some new tool carts. They got payback in under a week. That's the kind of ROI that should get every plant manager's attention. Luna: But I imagine not every plant can do this. There must be constraints — like union rules, or product complexity. Lucas: That's a fair point. This plant was non-union, which made cross-training easier. But the principles are universal. I've seen similar results in unionized facilities where the team had to negotiate new job classifications. It's slower, but it works. As for product complexity, SMED actually works best when you have many different parts — because that's when changeover time really adds up. A custom job shop with 200 different part numbers can benefit enormously. Luna: So what about industries outside stamping? Can this apply to, say, a bakery or a pharmaceutical line? Lucas: Absolutely. SMED is industry-agnostic. I've seen it applied to packaging lines, injection molding, even software deployments — though there the terminology changes. The core idea is always the same: identify what must be done while the process is stopped, and do everything else while it's running. In a bakery, that might mean pre-measuring ingredients for the next batch while the current bread is baking. In pharma, it's pre-sterilizing molds and pre-labelling containers while the previous batch runs. Luna: So the common thread is: preparation eliminates downtime. I think that's a lesson that goes beyond operations. Lucas: Exactly. And the beauty is that SMED doesn't require a big IT project or a capital expenditure request. It's just a fresh look at what your people are doing during changeover — and a willingness to challenge the belief that it has to take that long. Luna: I want to come back to that cultural piece. You said the hardest part was convincing people that 240 minutes wasn't normal. How do you institutionalize that mindset? Lucas: You build a system that makes continuous improvement the default. At Midwest Stamping, they created a 'changeover board' in each cell — a visual management tool that shows the target time, actual time, and any issues encountered. Every changeover is a mini-experiment. The team meets for 45 seconds after each swap to discuss what went well and what could be improved. That becomes a habit. Over time, the improvements compound. Luna: I've seen similar things in hospitals with surgical setup. So it's really about creating a feedback loop that's short and frequent. Lucas: Precisely. And the data doesn't lie. Once you have the baseline, you can see the improvement curve. That stamping plant went from 240 minutes in January to 36 minutes by June. Their target is now 20 minutes by the end of the year. They think they can get there with some additional investments in quick-change die holders. Luna: Twenty minutes from four hours. That's a 92 percent improvement from the starting point. And they're not done. Lucas: That's the thing about operations — the best plants never declare victory. They just raise the bar. For any listener managing a factory or a service operation, I'd say: don't assume your changeover time is fixed. Challenge it. You might be surprised what a $200 clamp and a team with a stopwatch can do. Luna: And maybe a dedicated changeover team and a willingness to film themselves. Lucas: Right. The camera doesn't lie either. So that's the story: 85 percent reduction, $2.3 million in recovered capacity, and a cultural shift that keeps delivering. Luna: It's a great example of how operations, when done right, can be a competitive advantage. Lucas: Absolutely. And it doesn't require a massive budget — just the discipline to look at your process with fresh eyes. For The Operations Podcast, I'm Lucas. Luna: And I'm Luna. Thanks for listening.