Latest / The Productivity Podcast with Fexingo: Output, Efficiency, and Long-Term Economic Growth / How One Factory Solved the Sticky Productivity Floor Problem
Transcript
- Lucas: There's a puzzle in productivity research that most people outside operations have never heard of. It's called the 'sticky floor' — and it describes something almost every factory manager has seen but couldn't name. Luna: Sticky floor — sounds like the opposite of a glass ceiling. What does it mean here? Lucas: It means that when a plant or a team hits a new high in output, that level tends to become the new floor. Output doesn't fall back to the old average — it sticks at the new high. That sounds great, right? The problem is, the opposite is also true. If output drops sharply one month — because of a breakdown, a supply hiccup, whatever — that lower level also becomes sticky. Performance gets trapped there. Luna: So a bad month can reset the baseline downward permanently? Lucas: Exactly. And that's what a 2019 paper by Gibbons, Henderson, Repetto, and Serpa found when they studied dozens of manufacturing plants over several years. They saw that upward improvements stuck, but downward shocks also stuck. The floor is sticky in both directions. Luna: That's a brutal asymmetry. You get punished for bad luck but you don't get rewarded for good luck — it just becomes the new normal. Lucas: Right. And that raises the question: can you design a system that only lets the upward stickiness happen? That's what brings me to the case I want to talk about today. A Toyota transmission plant in Alabama that, starting around 2015, figured out how to break the downward stickiness. Luna: Toyota's manufacturing system is already legendary. What were they doing differently? Lucas: So the plant had been running a standard lean production system for years. Good metrics, continuous improvement teams, the usual Toyota toolkit. But they still saw these sticky floors — a bad month would drag down the average and it would take months to climb back. The breakthrough came when they shifted from tracking output after the fact to intervening in real time. Luna: Real time how? You can't fix a transmission faster while it's moving down the line. Lucas: No, but you can catch the conditions that lead to a slow shift before they compound. What they did was install visual management boards at every team station — not digital dashboards, but physical whiteboards with magnets — showing the target output for that hour, the actual output, and the gap. Every hour, the team would meet for five minutes and decide: if we're behind, what's the one thing we do right now to recover? Luna: That sounds deceptively simple. A five-minute huddle with a whiteboard? Lucas: The simplicity is the point. The key wasn't the board — it was the shift in ownership. Previously, when output dipped, supervisors would come in, troubleshoot, and then leave. The team just followed instructions. Under the new system, the team owned the recovery process. They identified the root cause themselves — maybe a missing part, a tool that needed calibration, a new hire who needed a buddy — and they implemented the fix. The supervisor became a coach, not a commander. Luna: So the downward stickiness broke because the team caught small dips before they became big ones? Lucas: Yes. The paper's data showed that before the intervention, a dip of five percent or more in one month had a seventy percent chance of persisting into the next month. After the intervention, that dropped to about thirty percent. The team was able to recover within days, not months. Luna: That's a massive shift. And it's not about some high-tech automation — it's about giving people the tools and the authority to solve their own problems. Lucas: Exactly. And the plant didn't stop there. They also changed how they handled the 'upward stickiness' — the good kind. Instead of just celebrating a record month and moving on, they would study what made that month exceptional. Was it a new fixture? A scheduling change? Then they would try to standardize that practice so it became the new baseline, not just a lucky outlier. Luna: So they were deliberately engineering the upward stickiness and preventing the downward stickiness. That's a powerful combination. Lucas: There's another wrinkle in the data that I think is really interesting. The gains weren't uniform across the plant. Some teams adopted the new approach much faster than others. The researchers found that the teams that had the strongest social cohesion — measured by things like low absenteeism, high tenure, and good peer reviews — were the ones that saw the biggest and most sustained improvements. Luna: So culture matters even in a highly standardized system like Toyota's. Lucas: Absolutely. You can have the best visual management system in the world, but if the team doesn't trust each other enough to speak up when something's wrong, the sticky floor stays sticky. Luna: Let's zoom out for a second. This is one plant, one intervention. Does this scale? Have other plants or industries adopted this approach? Lucas: The honest answer is: not as widely as you'd think. The paper studied about forty plants, and only a handful implemented this kind of team-owned recovery process. The rest stuck with traditional supervisor-led troubleshooting. Why? Because it's hard to change the role of the supervisor. It requires letting go of control, and for many managers, that feels risky. Luna: Also, the results take time to show up. A quarterly earnings call doesn't reward a plant manager who says 'we spent six months teaching teams how to huddle better.' Lucas: Exactly. The time horizon mismatch is real. But here's the thing — the plants that did implement it saw a compound annual growth rate in output that was about three percentage points higher than the control plants over three years. That's enormous in a mature manufacturing setting. Luna: Three percentage points — that's basically the difference between a plant that's competitive and one that's struggling. Lucas: Right. And it didn't require new equipment or more headcount. It required a change in how people worked together. That's the kind of productivity gain that doesn't show up in the macro statistics but is huge at the firm level. Luna: Before we go further — I want to step back and just say, if conversations like this are useful to you and you want them to stay ad-free, listener support is what makes that possible. The link is buy me a coffee dot com slash fexingo. That's literally it. No pressure, just an option if today was worth a coffee to you. Lucas: Yeah, and we really mean that. It keeps the show independent and lets us dig into these specific cases. So thanks to anyone who chips in. Luna: Alright, back to the plant floor. One thing I'm curious about — did the sticky floor phenomenon show up in any non-manufacturing settings? Like software teams or hospitals? Lucas: Great question. The paper focused on manufacturing, but the concept has been applied elsewhere. For example, a study of emergency department teams found a similar pattern: a bad shift — high wait times, low patient throughput — tended to stick. The next shift would inherit the backlog and also underperform. And the same kind of team-level intervention — huddles, ownership of recovery, real-time visual tracking — helped break that cycle. Luna: Interesting. So the principles might be generalizable even if the specific implementation looks different. Lucas: Exactly. The core insight is that productivity isn't just about individual effort or even process design. It's about the team's ability to detect and correct deviations quickly, and to learn from both good and bad outcomes. That's a capability you have to build deliberately. Luna: Let's talk about one more finding from the paper that stuck with me. The researchers also noted that the sticky floor effect was stronger in plants with high turnover. When experienced workers left, the floor dropped and stayed down. Lucas: Yes. That's the human capital dimension. If you lose the people who hold the tacit knowledge — the things that aren't in the procedure manual — you lose the ability to recover from dips. That's why the team-owned recovery process is so powerful: it codifies that tacit knowledge in the team's routine, not in any one individual. Luna: So it's not just about the huddle itself, but about capturing and distributing knowledge across the team. Lucas: Exactly. And that's hard to do if you rely on a star operator who just seems to know how to fix things. When that star leaves, the knowledge leaves with them. Luna: One last thing — the paper is from 2019. Has there been follow-up work? Do we know if the Alabama plant sustained those gains? Lucas: There hasn't been a formal follow-up published, but I've spoken with people familiar with the plant and they say the practices are still in place. The team huddles are now part of the culture. New hires are trained into them. The sticky floor hasn't returned. But it requires constant reinforcement — you can't just install a whiteboard and walk away. Luna: That's the hard part about productivity improvements. They're not a one-time fix; they're a continuous practice. Lucas: Exactly. And that's maybe the biggest lesson from this case: the plants that succeeded treated productivity not as a target to hit, but as a muscle to build. The sticky floor breaks when the team itself becomes the mechanism for continuous improvement. Luna: So for a manager listening right now, what's the one actionable takeaway? Lucas: Look at your own team's output data. Find the last time you had a bad week or month. Did you recover quickly, or did that lower level become the new normal? If it stuck, ask yourself: does my team have the tools and the authority to diagnose and fix the problem themselves, or do they wait for someone else to tell them what to do? If it's the latter, you've got a sticky floor waiting to happen. Luna: That's a concrete place to start. And it doesn't cost a thing. Lucas: Right. And if you want to go deeper, the Gibbons, Henderson, Repetto, and Serpa paper is freely available. It's called 'The Persistence of Productivity Shocks.' Full of specific data. I'd recommend it to anyone who manages a team that produces something measurable. Luna: Thanks, Lucas. This was one of those episodes that makes me want to go apply something immediately. Lucas: That's the goal. Thanks for the conversation, Luna.