What UX teams can learn from a 1980s car plant in Fremont
NUMMI, a joint partnership between GM and Toyota, offers lessons in teamwork, leadership, and quality control.
About a decade ago, I learned about a defunct car plant in Fremont, CA, called NUMMI through a This American Life episode. It’s one of my favorite This American Life episodes — it’s meticulously researched and well structured and narrated. It’s OK you can go listen to it now. I’ll be waiting right here.
The NUMMI story has stuck with me and I’ve even listened to the episode a few times since it’s original airing. As a Bay Area transplant, the idea that there was a time when actual products (let alone cars!) were manufactured in Silicon Valley fascinates me. Aside from Bay Area history, the NUMMI story offers a lot of lessons in organizational behavior, quality control, and innovation that paralleled some of my experiences in UX.
NUMMI was a failed GM plant known for having the worst workforce in the US auto industry. It was the type of place where workers were constantly fighting with management. Workers didn’t care about their jobs — they routinely drank on the job and produced cars with serious defects (like cars coming off the production line with a fender from a different model). GM eventually shut down NUMMI and laid off all the workers.
But then in 1984, GM and Toyota formed a partnership to build cars together at NUMMI. At the time, GM was the largest automaker and Toyota was a much smaller company. Toyota was known for building small, reliable cars. GM lost money on small cars and wanted to learn from Toyota’s manufacturing methods. Toyota wanted to learn from GM how to make cars in the US.
They rehired most of NUMMI’s old workforce and started flying groups of American workers to Japan to learn Toyota’s methods and be immersed in its culture. The American workers were amazed by what they saw. At Toyota, teams were formed of 4–5 workers, along with a team leader. To break up the monotony of doing one job over and over again, they’d switch jobs every few hours. When a worker got behind on a specific task in the manufacturing process, a team leader would offer to help. Japanese workers were also asked to contribute their ideas for making improvements on the production line so they don’t encounter problems again. And Toyota managers didn’t just solicit their feedback, they’d quickly iterate to build tools and solutions based on what workers described. Workers received bonuses if their ideas were implemented.
This approach was a far cry from what American workers were used to at GM. When American workers fell behind, they weren’t offered help. They were instead yelled at to hurry up! Doing so might have kept the line going but it meant that workers got sloppy and made mistakes in an attempt to not fall behind. These mistakes were harder to fix as cars continued down the production line and more and more components and parts were added on. The mistakes would add up over time, resulting in cars that were unreliable. GM workers were discouraged from stopping the line and could even get fired for doing so. Management’s assumption was that the workers were lazy and would want to stop the line all the time.
Meanwhile, Toyota’s plants were equipped with an andon cord, a system that empowered any worker to stop the production line. The first pull of the andon cord didn’t stop the line but it enabled workers to alert team leaders of serious issues. If the team leader and worker were unable to resolve the issue, they’d pull the cord again to stop production. Toyota’s thinking here is that it was easier to easier to fix problems at that point in the production line, before more parts were added on. And the worker at the source of the problem on the production line was likely more skilled at that particular process of building the vehicle so they were less likely to make mistakes when completing the repair. This whole system of teamwork was built on a trust between workers and managers.
The American workers’ immersion in Japan was pivotal to their adoption of Toyota’s methods. Not only did they learn this new system in theory, they saw it practiced in real production plants. When they returned to the US, the workers at NUMMI almost immediately started building quality cars. These workers had so much pride in what they were building. One worker recounted how whenever he’d encounter one of the cars he had helped build in a parking lot, he would leave a postcard with his contact information under the windshield wipers of the car. He wanted to hear from drivers about their feedback on the car. Another worker would go to car dealerships just to stare at the cars he had helped build.
NUMMI is an amazing story of turning around a plant and changing its culture. Sadly, what happened at NUMMI stayed at NUMMI and GM wasn’t able to replicate those methods at its other plants. And much of why GM failed at replicating NUMMI came down to organizational culture and behavior. A lot of GM’s management and workers were proud and defensive. They refused to believe the data that American cars were less reliable than Japanese cars.
GM also didn’t understand the systemic and cultural factors that made NUMMI successful. They focused on the cosmetics of making other plants look like NUMMI, without investing in shifting the entire ecosystem and culture around those plants. For instance, they tried unsuccessfully to implement the same culture and production system at a plant in Van Nuys, CA. For one thing, the Van Nuys plant wasn’t affiliated with Toyota. That meant that the Van Nuys workers didn’t get immersion trips to Japan and their training wasn’t as extensive as the Fremont workers. That also meant that when workers had suggestions for reworking parts there wasn’t a system in place to do anything with the feedback. At NUMMI, Toyota would take that feedback and rework the parts that it supplied NUMMI. In short, an isolated plant can’t do such a massive overhaul of their culture and production system on their own.
Another factor that made NUMMI successful is that the workers in Fremont knew what it was like for their plant to be shut down and to lose their jobs. The Van Nuys workers didn’t have that experience and were skeptical about all the changes being implemented. To make things even worse, the bonus system in Van Nuys still rewarded making cars so managers weren’t incentivized to stop the line and fix problems on the production line.
Five Lessons for UX teams
I see so many parallels between the NUMMI story and my work in UX. Here are five lessons I took away:
- Empower anyone to “stop the line.” On many teams, UXers are often not only the voice of the user but also advocates for higher quality. Through research and iterative design processes, we are often able to identify potential issues and problems with products early in product development. Given that UX is still a somewhat nascent field in product development and given that so many companies incentivize launching products, UX isn’t always empowered to “stop the line.” Yet the NUMMI story shows us how critical it is to enable those closest to the fine details of the work to stop production when they identify quality issues. Software and hardware development teams should think about how to build systems in place where anyone can metaphorically “stop the line.”
- Design and UX processes aren’t cosmetic. GM managers thought they could replicate NUMMI’s success by reworking their plants to look like NUMMI, when NUMMI’s success was due to a whole ecosystem and culture that supported innovation and quality control. The same can be said for design and UX. It doesn’t matter how many UXers you’ve hired, how many studies your team has conducted, how many sprints you’ve run, or how many iMacs litter your design studio. If you don’t have a culture and system to support quality and innovation, UX isn’t part of your development process and you haven’t invested in it. Everyone has to be onboard with integrating UX into the development process.
- It should be OK to make mistakes. A key to Toyota’s focus on quality was not penalizing workers for making mistakes. They encouraged their teams to recognize and correct problems, rather than hide them. This idea comes up often in a lot of Amy Edmondson’s research on psychological safety. Organizations with high psychological safety create an environment where it is OK to make mistakes.
- Adopt the teamwork principle. I found a lot of parallels between the relationship between workers and management in the auto industry and the relationship between UX, product management, and engineering in tech. If you’ve ever worked on a team where that relationship is contentious, you know how hard it is to build good products in that environment. A lot of folks in tech could benefit to adopting the Japanese principle of teamwork. Of course, that’s easier said than done, especially if you’re starting from a place where trust between different team members has been eroded.
- Culture change is hard! No duh, right? But it really is important to recognize how hard it is to change organizational culture and the time and effort it takes to do so. Even if you’re not trying to change how an auto manufacturer builds cars, there’s a good chance that at some point in your career you may propose a new process, meeting, or way of doing things. That small idea you may have may not seem like a big deal to you but you’re essentially asking your colleagues to slightly change their team culture. And there’s a pretty good chance you’ll be met with some resistance.
Further Listening and Reading:
- NUMMI (2015) — This American Life
- How to Change a Culture: Lessons from NUMMI (this article is behind a paywall but if you sign up for a free MIT Sloan Management Review account you can read 3 articles for free)
- Creating Psychological Safety in the Workplace — HBR Ideacast
- The Fearless Organization
This post includes affiliate links to Amazon.