Today, I'm writing about some published reports about the alleged state of quality at the Tesla plant in Fremont, California.
If you don't know the history of the Tesla factory building, here it is in a nutshell:
First, the building was the GM Fremont plant from 1961 to 1982, where quality, productivity, and other measures were terrible. It was a poster child for the bad culture and poor results of a GM plant that was managed in the typical GM way.
John Shook, until just recently the CEO of the Lean Enterprise Institute, called the Fremont plant a “dysfunctional disaster” before the turnaround (something that he was a part of, as a Toyota employee). At the time, GM's then-CEO Roger Smith was chasing a “lights-out factory” model that relied heavily on automation. They wasted $90 BILLION dollars on this failed vision, company wide, as I wrote about here:
Hear Mark read this post (subscribe to the podcast):
Very quickly, the same plant with workers (selectively) hired from the old Fremont plant (and the same UAW) quickly became very successful with a completely-new management team and a new leadership style. This 1989 New York Times article described the NUMMI plant as having “extraordinarily high productivity and extraordinarily low level of defects and absenteeism,” even though it wasn't paradise, as the article suggests.
You can check out additional posts and podcasts of mine about the NUMMI story including:
During the financial crisis and their bankruptcy, GM pulled out of the NUMMI joint venture, which, among other factors, led to the closure of the plant.
After the closure, Tesla bought the plant and some equipment for a “bargain price” of $42 million. It became the Tesla Factory. Tesla's website calls it “one of the world's most advanced automotive plants.” I guess that depends on what “advanced” means. More on that later.
Some people who know the NUMMI story assume that Tesla has continued down that path of TPS, Lean, and quality. I'm not sure that's been the case.
Tesla and Toyota – A Partial Partnership
At the time they bought the building, Tesla had a partnership with Toyota, but that “fizzled” and ended, per a Bloomberg article from 2014: “Why the Tesla-Toyota Partnership Short-Circuited.” The partnership completely ended earlier this year.
Some of the recent news reports (more on this later) make me wonder if Tesla tried hard enough to learn automotive manufacturing practices from Toyota during their partnership. The VP of Manufacturing who started in 2010, Gilbert Passin, had significant experience with Toyota, including time spent running their plant in Cambridge, Ontario (and Tesla trumpeted this in their release about his hiring).
But I wonder how the Tesla culture developed over time with the leadership changes that took place. Passin was mentioned in a 2013 article I linked to here. Passin's LinkedIn page says he is still in that role, but his bio is gone from the company website and it seems that he was replaced sometime after late 2013. Through some Google searches, I can't find evidence of Passin moving onto another role.
Founder and CEO Elon Musk's goal with a group of 2011 executive hires was that by “blending the top minds from technology, automotive and manufacturing, Tesla is building the car company for the future.” The VP of Supply Chain hire at that time came from a semiconductor background which is, of course, very different than a high-volume automotive supply chain.
In 2016, Tesla hired an executive from Audi to run manufacturing, replacing a previous executive, Greg Reichow, who had semiconductor experience instead of coming from the auto world. Reichow, as VP of Production, and John Ensign, the VP of Manufacturing (who came from Honeywell), both left in 2016. They had separate VPs of production AND manufacturing? Interesting.
This article (with quotes from Reichow after he left the company) is an in-depth look at the somewhat “secret” second floor that produces battery packs and other components that Toyota never had to worry about: “Tesla's former VP of Production talks ‘secret second factory floor' and vertical integration”
It seems to me that the combination of Tesla technology and Toyota manufacturing know-how could have been very powerful… but did things get off track with non-automotive leaders running manufacturing? Or were there other factors? Or, are things really fine?
Phase 3: Tesla
In recent months, there have been reports from blogs, including one called “Daily Kanban,” about problems and delays with the Tesla Model 3 ramp up: Source: Tesla Responsible For Model 3 “Production Hell”
The slow, delayed Model 3 ramp up has been covered a lot in the mainstream media, including here: “Tesla Model 3 production delays start to take toll on company.”
There have been reports of quality problems with the initial Model 3s that were sent to reviewers and investors, as discussed here:
It's probably not a fair comparison, but being behind schedule and having quality problems sounds more like the “old GM” than Toyota/NUMMI.
Bernstein analyst Toni Sacconaghi wrote in a note to clients [in November]:
“While we doubt that it would impact (or even be noticed by) most prospective buyers, we do worry that poor overall initial quality could undermine Tesla's brand and potentially overwhelm its service network.”
What were those problems?
“The analyst found a misalignment in the Model 3's glass roof, body panel gaps, rubber trim issues around windows and misaligned seams in the interior ceiling of the car, according to his note.
“Our inspection revealed widespread shortcomings in fit and finish … Tesla representatives acknowledged some of the fit issues, but stated that they believed that Model 3 was much further ahead than where Model X and S had been at this point in production,” he wrote.”
So Tesla's defense is, basically, that the early Model X and Model S cars were worse? Interesting.
The analyst said:
“…we can't help noting that Tesla likely chose to share with us its highest quality/best assembled units, so issues on other cars may be even more pronounced.”
Good point… Tesla didn't try saying “You should have seen the cars we did NOT send for the test drive.”
Model S and X Are Not a Model of Production Excellence?
It's one thing to have problems with pre-production or early production vehicles, as could be occurring with the Model 3.
But a recent report from Reuters talks about problems with the car that Tesla has been producing the longest – the Model S.
The headline is shocking and damning:
Some claim they were fired for being in favor of the UAW coming back into the plant to represent the Tesla workforce. The union has been trying to get back in for a while, so I might generally take reports with a grain of salt, given the feelings and agendas of those who might want to paint Tesla in a negative light. I'd expect to see negative articles about Toyota, Nissan, Honda, Volkswagen etc. during a UAW drive.
But, Reuters probably isn't going to go to press with uncorroborated stories or “fake news.”
From their article:
“After Tesla's Model S sedans and Model X SUVs roll off the company's Fremont, California assembly line, the electric vehicles usually make another stop – for repairs, nine current and former employees have told Reuters.
The luxury cars regularly require fixes before they can leave the factory, according to the workers. Quality checks have routinely revealed defects in more than 90 percent of Model S and Model X vehicles inspected after assembly, these individuals said, citing figures from Tesla's internal tracking system as recently as October.”
Reuters points out that “the world's most efficient automakers,” such as Toyota, have end-of-line defect rates of just 10%.
I've been in Toyota plants and have seen their brightly-lit “final inspection” areas. Yes, they have them.
Yes, a Lean thinker would say inspection is “waste” and they'd be right. Even though Toyota aims to build quality in from the start, per their TPS approach, they don't yet have a perfect process (or perfect quality) so, therefore, having final inspection (and doing it well) is better than NOT doing final inspection.
A defect getting to the customer is a far worse form of waste than doing final inspection is, in the short term.
Tesla claims their inspection process is:
“unusually rigorous, designed to flag and correct the tiniest imperfections.”
I'm sure that's the same at the Toyota and Lexus plants too.
Let's assume that Tesla final inspection is as good as Toyota's, and let's assume that means neither company misses many defects.
Either Tesla is indeed much pickier (which I doubt) or their built-in quality from suppliers and the assembly line is much worse.
The Reuters story has competing narratives:
“At Tesla “so much goes into rework after the car is done … that's where their money is being spent,” a former Tesla supervisor said.
The Silicon Valley automaker said the majority of its post-assembly defects are minor and resolved in a matter of minutes.”
So, if built-in quality is bad, Tesla says the defects are minor and fixed quickly. OK.
Except for the ones that aren't fixed quickly.
Again, there are competing stories, via Reuters:
“Trickier cases head to one of Tesla's outdoor parking lots to await repair. The backlog in one of those two lots, dubbed the “yard,” has exceeded 2,000 vehicles at times, workers told Reuters.
Tesla denied to Reuters that such “repair lots” exist.”
This reminds me of a story that Jim Womack, founder of the Lean Enterprise Institute, used to tell about cars coming off the end of the Cadillac line and then going to one of two places:
- Minor Repair
- Major Repair
As Jim recalled to me in 2007, some of the details of his story could be fuzzy or not 100% accurate, but he claims he “concluded that headcount in rework must be nearly equal to headcount on the line.”
Womack also said that the Cadillac plant manager claimed that their inspection was pickier than Chevrolet's.
That sounds familiar.
Former workers claim, but Reuters could not independently confirm problems, such as:
Defects included “doors not closing, material trim, missing parts, all kinds of stuff. Loose objects, water leaks, you name it,” another former supervisor said. “We've been building a Model S since 2012. How do we still have water leaks?”
Do those, if true, sound like allegations that would have come from a hypothetical “NUMMI 2” that combined Tesla and Toyota know how? Would a different, NUMMI-style partnership, have helped Tesla here?
The Right Cars Built Correctly?
Tesla customers are probably still characterized as “innovators” or “early adopters” in the technology adoption curve, so they are more likely to be forgiving of assembly quality problems that they might not tolerate in a Lexus, Cadillac, or BMW. Analysts say that's likely to change with the lower-priced Model 3, where Tesla aims to move into the mass market.
Some customers are going to be, net-net, happy with an exciting, attractive, fast, green car that has some minor assembly defects.
A successful automaker needs to bring the right product to market AND build it properly. Some Toyota critics say they don't create exciting cars (something CEO Akio Toyoda is trying to address).
One analyst said:
“We've never doubted Tesla's ability to make exciting products with top specifications, but there's a difference between unveiling something and then actually making it perfectly in large volume. Tesla has not perfected the latter yet,” Morningstar analyst David Whiston wrote earlier this month.
Tesla having an exciting car that they struggle to build without basic fit-and-finish issues makes me think of a hospital that's “world class” and does “cutting edge surgery,” but has basic problems around cleanliness and infection control practices…
Best and Most Advanced?
What do those labels mean, in the context of Tesla? Elon Musk seems to have a Roger Smith-esque bias toward automation and robotics.
If Smith was wrong about that in the 1980s, maybe he was just ahead of his time and Musk will get it right this time.
“Tesla Chief Executive Elon Musk has vowed the company would become “the best manufacturer on Earth,” helped by a new, highly automated assembly line and a simpler design for the Model 3.”
A “simpler design” that's easier to build with fewer labor hours is something Toyota and other automakers have generally worked toward over time.
In the automotive world, the most highly automated plants haven't necessarily been the “best” manufacturers. It's been said that Toyota plants tend to be slightly less automated than the Detroit automakers since Detroit tried getting as many people out of the process as possible (because of labor battles and management mindsets), while Toyota tends to automate jobs only when safety is an issue or when quality could be improved.
If Elon Musk wants to be the best manufacturer (and not just have the fastest, most exciting cars), can he combine the smart use of automation with the management system of Toyota that engages people's brains in continuous improvement? I explored this in an earlier post:
Tesla brags about its inspection process and its high standards (remember, GM/Cadillac used to do the same):
“Our goal is to produce perfect cars for every customer,” Tesla said in a statement. “Therefore, we review every vehicle for even the smallest refinement. Most customers would never notice the work that is done post production, but we care about even a fraction of a millimeter body gap difference or a slight paint gloss texture. We then feed these improvements back to production in a pursuit of perfection.”
And they talk a good game about continuous improvement, using the same “pursuit of perfection” language you would hear in a Lexus ad or a Lean presentation.
That said, if Tesla is fixing things that customers would never notice or care about, is that an example of the “waste of overprocessing?”
A Tesla defender might say I'm criticizing Tesla for poor quality AND criticizing them for fixing things. That's not inconsistent if they're fixing things that don't matter to the customer. Value (and quality) is defined by the customer.
When I worked at GM, engineering would brag about “aerospace tolerances” on Cadillac engine parts. Was that really producing a better engine or did that just sound good from an engineering and marketing perspective?
Quantity or Quality?
When I worked at GM and Dell, one common theme was that management quite often emphasized the goal of “making the numbers” – meaning we needed to hit production quotas as the primary goal. Sure, quality was perhaps considered “nice to have,” but I have so many stories about management making bad decisions that prioritized quantity over quality, including this one:
At Dell, I remember times when the ONLY end-of-quarter metrics that had incentives for factory workers were QUANTITY related. Nobody got rewarded for quality metrics. The implied message was clear: “Ship it. Get it out the door. We can fix quality problems later…” As they used to say at GM, “That's what warranties are for.”
One thing that's special about the Toyota Production System and Lean is that it's supposed to seriously put quality first. Any worker is allowed, if not expected, to “stop the line” if there is a defect, pulling an “andon cord” or pushing a button.
Many of us in healthcare have advocated for this mindset of putting safety first and being able to stop the line, including the “patient safety alert” system at Virginia Mason Medical Center.
The accusation at Tesla is that pressure to make the numbers leads to cutting corners or looking the other way on quality:
Employees who worked on Model S and Model X described pressure to keep the assembly line moving, even when problems emerged. Some told of batches of cars being sent through with parts missing – windshields in one case, bumpers in another – because there were none on hand. The understanding, they said, was that these and other flaws would be fixed later.
Quality inspectors would sometimes find more defects than those reported by workers in the internal tracking system when a car came off the line. “We'd see two issues, that's pretty good. But then we'd dig in and there would be like 15 or 20,” one person said.
One persistently tricky area was alignment, where body parts had to be “muscled,” in the words of the senior manager, to a certain degree of flushness. Not every team follows the same rule book, workers said, resulting in gaps of different size.”
I'm sorry, but that sounds like GM (and, again, the GM of decades ago), not Toyota.
Tesla, again, denies the stories.
Tesla claims that problems are found and caught. But this news story talks about something I'd call a quality problem: “Tesla Shipped Cars Without Seats And Digital Displays: Report“
I remember a story from the late 1990s or early 2000s where Ford had a seat supplier go on strike. This was during the boom of SUVs, so any lost production was VERY costly to Ford, losing a sale of their most profitable vehicles. Instead of stopping the line, I remember stories about Ford building parking lots full of SUVs without seats. It was an intentional choice, not a mistake. They figured it was cheaper and better to put the seats in later, even though it was outside of the planned, regular production process. I'm not sure I'd be comfortable with a car built outside of the validated production process. But, I'm an engineer.
Back to Tesla… the Financial Times reported (as shared by Jalopnik):
“…the company “shipped cars from the factory that lacked key parts,” including everything from digital displays to seats.
These parts were flown to Tesla-owned dealers, who then assembled them into the vehicle before completing the shipments to customers, according to several people familiar with the practice.
“This goes back years,” says a former regional executive, who declined to be named.
“It was common, common in every market — the seats, the displays were being flown in.”
Tesla doesn't deny this claim, but they give some “spin” that I don't believe:
“Unlike other car companies, which do not change their cars for at least a year at a time, Tesla is constantly improving its cars with over-the-air updates and often design and hardware improvements.”
“As but one example, that means occasionally we will even send, say, new certified parts to meet a car at the delivery center if those items have been upgraded after the car has shipped,” the spokesperson said. “This process may be unfamiliar to some, but has worked very well for us, as our customers know that if we can add value or make something better, we will do everything we can to do it right away.”
I totally understand how and why Tesla would send SOFTWARE updates to a vehicle. I don't buy the story that says, basically, “We didn't ship that car without seats… the customer is getting the benefit of an upgrade.” That's just not how things are done in the auto industry.
Of course, what I'd call a “bug” in the production process is claimed to be a “feature” by Tesla? Is Tesla building a new kind of car company or do they not really know how to build cars well?
So What's the Truth?
I don't know the real situation at Tesla. I'm writing this without being at the “gemba” (the actual place, the factory floor).
GM was in denial for a long time, and we know what happened there.
I want Tesla to succeed for many, many reasons. I hope the Reuters reports are lies made up by disgruntled former employees… but Reuters probably talked to many people (one of the stories above was corroborated by nine people). Either they're telling the truth (and Tesla has a lot of problems to solve), or they fooled the Reuters reporters… what do you think?