Never in my three decades in the auto industry have I seen a company throw so many suppliers under the bus quite so publicly in the way Tesla does. Going back a decade to the delayed launch of the Roadster when two different transmission suppliers were abandoned before Tesla finally redesigned its powertrain, the automaker has repeatedly blamed others for its own engineering deficiencies.
We saw this again Wednesday as Tesla claimed that suppliers were responsible for battery manufacturing snags at its Gigafactory that the company says are the main reason for the slower-than-promised ramp-up in production of its Model 3 sedan.
While I’m not privy to the contracts that Tesla has with its supply base, the reality is that they work with most of the same companies that every other automaker does to procure parts and manufacturing equipment. These are generally proven companies that have fine working relationships with other manufacturers.
High-volume manufacturing of any complex product is not a trivial challenge. As an engineering co-op student too many years ago, I spent my work terms in manufacturing environments before moving to product development after I graduated. Manufacturing equipment for mass production is expensive and surprisingly sophisticated.
Virtually all of it today is computer controlled, and it features sensors that can detect when tools need replacing or servicing is required. The goal is to produce products with a high degree of reliability and reproducibility the first time through. Anytime parts have to leave the main line to go through repair loops, it adds tremendously to the cost.
Developing this equipment is a very collaborative process. Today especially, a lot of parallel engineering takes place during product development. The manufacturing engineers work with the designers and development engineers from the beginning of a program to ensure that the product can actually be manufactured reliably and cost effectively. As the product comes together, the manufacturing team works out the steps in the build process, develops specifications and goes out to the equipment suppliers for bids. Once a vendor is selected, the automaker’s engineers work closely with the supplier to ensure that the machinery will be able to produce parts and vehicles to the required tolerances.
Before equipment is shipped to the factory, it is tested at the vendor’s shop and debugged by the engineers and technicians who designed and produced it. The automaker engineers usually spend time there verifying operation before signing off. Often there will be elements from multiple vendors that need to be integrated. For example, in a body shop, a transfer line might be built by one vendor while spot-welding robots come from another, with gauge equipment from still another.
When it works, it’s a remarkable mechanical ballet to watch. If you ever have the opportunity to visit a factory like the Ford Rouge assembly plant in Michigan, where the F-150 is built, it’s worth the trip.
However, long before opening night of any ballet, whether on a stage or in a factory, there are countless rehearsals. There are no shortcuts to getting this right. The automaker’s manufacturing engineering team ultimately serves as the choreographers and directors. If the dancers are staged in the wrong place, with lights shining in their eyes, those choreographers and directors are ultimately responsible. If the dancers are incapable of executing the required moves, they may well be replaced, but that is something that should be apparent well before opening night, and those who cast them must also take some responsibility, especially if they failed to recognize the problem soon enough.
In its third-quarter 2017 financial report, Tesla highlighted manufacturing issues at its Gigafactory in Reno, Nev., where its lithium cells and battery packs are produced.
The biggest challenge is that the first two zones of a four-zone process, key elements of which were done by manufacturing systems suppliers, had to be taken over and significantly redesigned by Tesla. We have redirected our best engineering talent to fine-tune the automated processes.
After the publication of this article, a Tesla spokesman pointed out that on a conference call with analysts after its earnings release, CEO Elon Musk did take responsibility for the problem – after a fashion.
I think at the end of the day, everything is our fault – and my fault most of all. If we pick the wrong subcontractor, we're the fault. So, I don't want to – just to be externalizing responsibility, really it's our fault for picking the wrong supplier and then not realizing it until way late in the game.
It’s certainly possible the supplier(s) were incapable of delivering what the sourcing contract called for, and perhaps de-sourcing them was the right call. However, Musk's claim that there was something “wrong” with the supplier is an obfuscation – the customer engineers from Tesla are ultimately responsible for managing a project like this. If they did indeed have to redesign the process to the degree indicated in the shareholder letter, this is a major project management failure. An issue of this magnitude indicates equipment that should never have been delivered to the Tesla plant in the first place.
Unlike the battery packs and cars that will ultimately be produced by the tens of thousands every month, the machines that build the machines are not mass produced. These are bespoke creations built to perform very specific tasks. Unlike a car, which should roll off the end of the assembly and function without rework, these machines almost invariably require modification before production begins in earnest. If Tesla has made numerous specification changes along the way, as has been reported by The Daily Kanban, the rework may be extensive and time consuming.
Again, the buck stops with Tesla. The original Roadster transmission design of 10 years ago was inadequately developed to begin with, and it should not have gotten as far as it did before Tesla shifted course while blaming its erstwhile suppliers. In 2016, Tesla used a Mobileye vision system for its original AutoPilot system in a way that Mobileye warned it was not adequate for. When it failed to do the job in the crash that led to the death of Joshua Brown, Tesla blamed Mobileye.
It's good this time to hear Musk take some responsibility for these recurring issues. However, the fact that such issues continue to happen year after year is still a sign of more fundamental management issues at Tesla. The consistent message coming from suppliers is that Tesla keeps shifting direction long after designs should be locked down. To a degree you can get away with this in the software business, but when you are building complex hardware, late design changes will result in delays.
As one of my first engineering managers told me many years ago, there comes a point in every project when you need to shoot the engineers and get on with the job. Musk talked about a pencils-down date for the Model 3 in mid-2016, but it appears that the pencils may still be scribbling.
Tesla has repeatedly failed in its own responsibilities to do adequate engineering and management, followed by Musk publicly casting blame elsewhere. If I were running a supplier business, I would probably pass on even bidding on a Tesla project.
Sam Abuelsamid, Contributor