A perspective on why performance problems are usually visibility problems
You sit across from someone with a performance plan in front of you. You use words like ownership and accountability. Some part of you wonders whether you’re solving the real problem or just managing a symptom you’ve seen before. That doubt rarely comes from softness. It comes from pattern recognition. You’ve watched this play out, and the last hire couldn’t fix what this one can’t either. The quiet possibility worth sitting with: maybe the problem isn’t the person. Maybe it’s the system design they’re working inside.
The Plant That Proved the Point
In the early 1980s, General Motors ran an assembly plant in Fremont, California, that was considered one of the worst in the American auto industry. Drugs, alcohol, and absenteeism were rampant, and on any given day, one in five workers didn’t show up. Grievances piled up. Strikes were common. By 1982, GM had closed the plant and laid off thousands of workers.
Two years later, something unusual happened. GM partnered with Toyota to reopen the same plant as a joint venture called NUMMI. Toyota insisted on rehiring the same union workforce GM had just let go, and over 85 percent of NUMMI’s initial workforce were the same people laid off from Fremont in 1982.
The result wasn’t a slow improvement. Practically overnight, Toyota turned one of GM’s worst factories into one of its best. Defect rates plummeted, reaching quality levels comparable to Toyota’s own plants in Japan. The workforce stayed the same. The union stayed the same. Skills, habits, and history all stayed the same. The only thing that changed was the system design around them.
What Changed Wasn’t the People
This is the part worth sitting with longer than a quick case study usually allows. Nobody retrained these workers into different humans. Nobody replaced the workforce with more disciplined hires. Toyota changed what the workers could see and what they had authority to do about it.
Visibility
Under the old GM system, a defect introduced on the line might not surface until six stations later, if it surfaced at all. Workers described cars with engines installed backwards, missing steering wheels, mismatched parts stuck on and pushed out to the yard for someone else to fix later. The cardinal rule was simple and damaging: the line never stops. One former worker put it plainly: you saw a problem, you stopped the line, you got fired.
Under that kind of system design, a worker’s mistake was invisible to the person who made it. There was no feedback loop, only consequences that landed somewhere else, weeks later, on someone else’s desk. You cannot hold a person accountable for something they were structurally prevented from seeing.
Authority
Toyota’s most consequential change wasn’t a new tool or a new training manual. It was a thin nylon rope hanging along the assembly line called the andon cord. Any worker could pull it and stop the entire production line the moment something looked wrong. That single mechanism flipped the entire relationship between worker and outcome. Quality stopped being something management inspected for after the fact. It became something the person doing the work could catch, name, and fix in real time.
This is the piece that gets lost in most retellings of NUMMI. The story isn’t really about culture, or motivation, or a kinder management style, although those things mattered too. It’s about system design that gave people the ability to see consequences and the authority to act on what they saw. Statistician W. Edwards Deming spent decades making a related point to American management: most of the troubles and most of the possibilities for improvement in an organization belong to the system itself, not to the individuals working inside it. The number he used was 94 percent. His point wasn’t that people don’t matter. It was that looking to blame people is rarely a productive starting place for improvement.
The same logic shows up in how teams build the spaces they work inside. The post on workplace atmosphere as strategy makes a similar case from a different angle: the environment around a team isn’t a finishing touch, it’s a behavioral system that shapes what people do every day, whether anyone designed it on purpose or not.
Three Questions Before You Judge Someone
None of this means people never own real failures. Some performance problems are genuinely about the individual. But before reaching for that conclusion, three questions are worth asking, and they’re the same three the NUMMI story answers without trying to.
Can this person see the consequences of their work in real time, or does the feedback arrive somewhere else, too late to matter? Is there meaningful control over their output, or are they executing steps inside a process they didn’t design and can’t adjust? Do they understand how their specific role connects to the larger result, or is that connection only visible from a level of the organization they don’t have access to?
If the honest answer to any of these is no, the system design deserves scrutiny before the person does. GM fired an entire workforce and got the same outcome, because firing people doesn’t change a system that hides consequences from whoever replaces them. Toyota proved the inverse just as clearly. You can keep the exact same people and get a completely different result, if the structure around them changes first.
This reframing tends to unsettle leaders more than blame does, which is part of why blame stays so popular. Blaming a person is finite. You write them up, you let them go, you move on. Auditing a system means looking at decisions you made, structures you inherited, or workflows nobody questioned because they’d always been there. That’s a harder conversation to have with yourself than it is to have with an underperforming employee.
The Bottom Line
Clarity. Performance problems are usually visibility problems wearing a person’s name. The NUMMI plant didn’t succeed because Toyota found better workers. It succeeded because Toyota built a system design that let the same workers see what they were doing and act on what they saw.
Direction. Before the next performance conversation, run the three questions first. Can they see the result of their work. Do they have real control over it. Understanding how it connects to something larger matters too. If the system fails any of those tests, fix the structure before you fix the person.
Perspective. The most expensive mistake in most organizations isn’t a bad hire. It’s a bad system that keeps getting blamed on whoever happens to be standing inside it this quarter. What would change if you assumed the system first, every time, before you assumed the person?
Sources
- https://www.npr.org/transcripts/125229157
- https://www.thisamericanlife.org/403/transcript
- https://en.wikipedia.org/wiki/NUMMI
- https://deming.org/appreciation-for-a-system/
