Reinventing the car, episode 3
How Tesla cracked the code of automobile innovation
Tesla is several years ahead of other carmakers. A view from the inside can help us grasp the key differentiators that sustain its lead. That’s episode 3 of our series on the reinvention of the automobile (See Ep. 1 and 2)
By Philippe Chain* Frederic Filloux
On a chilly Wednesday morning in February 2012, we are four months ahead of the launch of the Tesla Model S. The car is still under development. Not much has been shown about the car that will become a game-changer in the automobile industry. A huge amount of work remains to be done.
There are twelve people in a conference room, on the 2nd floor of Tesla’s main plant, in Fremont, California: Jérôme Guillen, Model S program director, the head of manufacturing Gilbert Passin, the head of supply and purchasing Peter Carlsson, myself as the director in charge of quality and several key engineers.
The task force has been hastily assembled for a reason. A few hours ago, we received a preliminary report from a testing facility in Ohio where we are conducting some critical safety tests. The news is not good and it could halt the entire development of the car. Elon Musk had been informed of the issue. His reaction was a terse “Solve it, guys”.
In the United States, the car industry works under a self-certification system. The list of the 25 mandatory tests is set by the National Highway Traffic Safety Administration (NHTSA) and performed by carmakers themselves or by third-party test labs.
The Ohio facility has just performed test n°21 which involves a low-speed frontal collision against a concrete wall. The car actually passed the test. But there is a hitch. The deformation of frontal crash-boxes exceeded the design specifications.
A crash box is a piece of metal that plays the role of a one-time shock absorber in the event of a front crash. It is a critical safety element. When a car hits an obstacle head first, the released kinetic energy translates into a brutal deceleration expressed in G-force. At 30 mph (48km/h), a 160 lb (72 kg) individual wearing a seatbelt will typically experience a 30 g deceleration which is the equivalent of 2.1 tons acting on the body. It is generally admitted that a deceleration above 60 g’s is fatal. Hence the importance of a deformable structure that will spread the shock like the crash boxes.
There is one on each side of the car. It is made of steel, and carefully designed for an optimal collapse:
Here is how it plays out on a Tesla: the boxes in blue are folding like an accordion, absorbing part of the shock :
Practically, if the crash-box collapses excessively at low speed, it means that the car would fail the next collision performed at 35 mph (56km/h). And ours is scheduled for the following Monday at 2 pm at Karco Engineering, a remote testing lab in the San Bernardino desert, northeast of Los Angeles. We have four days to solve this and we can’t afford to delay the tests, nor can we waste too many cars. We have eleven prototype cars scheduled to be crashed. A few of them are there to be completely destroyed while others are to be re-used for another kind of crash (a large car maker might waste dozens of vehicles).
Back to the situation room. The engineer in charge of the crash-box design is clearly upset, split between guilt and incomprehension. We don’t waste time with finger-pointing. We need a solution, quickly. The conclusion is that we won’t change the design of the box, but we must use a stronger grade of steel. The purchasing manager summons one of his buyers, briefs him and the guy starts placing calls in the corridor. He’s back after ten minutes: “OK. There is a coil of steel available in North Carolina. Should I buy it?” Sure! No question!
In the next 24 hours, the one-thousand pounds coil is on its way to a special processing plant somewhere in the Midwest to be cut, formed, and welded to make the crash-box. As Murphy’s law would have it there is a huge snowstorm in the central United States, grounding the flight carrying our precious cargo.
The pieces finally arrive on a Saturday night at Tesla’s main plant for a final hardening in an oven. Bad luck, the equipment fails. The repairman is called in the middle of the night, and at dusk, the cooking, which takes several hours, starts. The Karco test is thirty-hours away. On Sunday night, the still-warm elements are loaded in the trunk of my car. Exhausted by a sleepless night, I don’t feel like driving all night, my deputy takes the wheel, and after six hours on the dangerous I-5 highway, we arrive at Adelanto, in the middle of nowhere, 60 miles northeast of Los Angeles, just in time to affix the two crash-boxes to the car. The test is passed with flying colors and overall, the Tesla Model S was eventually granted a 5-star rating on all counts by the NHTSA, the maximum score a vehicle can get.
To me, this anecdote is a perfect illustration of why Tesla is several years ahead of its direct competitors.
I have been in a similar situation elsewhere, in which a showstopper incident suddenly occurs, impairing the entire design process of a new car.
What took a single meeting and five days at Tesla would have required a six-month-long process at Renault or Audi. It would have started with a thorough investigation going back to the source of the error, looking for the person to blame internally, investigating a possible blunder of the supplier, etc. Then, a politically delicate sequence would have unfolded to set up a series of meetings and workshops. Multiple hierarchical strata would have been involved, going up probably to the very top of the organization.
That is the key to the Tesla Way: expedited, value-driven procedures, implemented with a maximum delegation, and a fast decision-making process in a flat structure.
This thinking involves the acceptance of a certain level of risk, which is unthinkable elsewhere. I’m not talking of customer safety features (on which there is no compromise whatsoever at Tesla) but more about development processes.
Another good example is the long-term durability evaluation of a car. When I discussed it with Elon, I told him our engineers’ calculations led to at least a million equivalent miles of driving required before launching the car — a six-month phase required to discover potential weaknesses and fix them. My request was actually very limited in regards to the industry practices: German manufacturers don’t release a car that has not clocked 10 million kilometers and two winters. Elon, in his customary laconic way, answered: “OK, do it. But we are not delaying the launch date for it…
— But we might encounter issues that will require some modifications of the production models…
— Yeah, I know, but we will make the changes afterward if we have to…
— Even if it involves recalling some cars?
— Yes. And for the rest, we will adjust by pushing some OTA upgrades (Tesla’s main software is maintained and upgraded remotely on a regular basis, just like a PC).
Tesla does not abide by the customary rules of the car industry.
While sometimes risky, it has largely contributed to its competitive edge. The company will never sacrifice the pace of its innovation for the sake of the sacrosanct process — a cardinal value in the car industry. For instance, legacy carmakers won’t ship a car with excessive “gaps and flush” (the space between body panels). Ours gave me a hard time as we couldn’t adjust what we call the “body geometry”. For months, we shipped cars afflicted with wider than usual gaps and the first models needed some manual adjustments with mallet and foam… What would have been deemed as unacceptable by any carmakers was seen as part of an ongoing process by Elon Musk who believed, rightly so, that the user experience of driving a truly innovative automobile would outweigh minor defects that will be eventually corrected. As the success of the brand (and market research) shows, Elon was right in his intuition.
Where Tesla’s peculiar DNA is the most visible is in the recruiting. At the time I was there, the VP of Human Resources was Arnnon Geshuri, former HR director at Google where the staffing grew from 2500 to more than 20,000 people under his watch. Again, Elon handed him a straightforward roadmap: a) hire nothing but champions in their field, and b) make sure they have the right cultural fit. The result is a cohort of problem-solvers selected for their knowledge and expertise, but also for their creativity, sense of autonomy, and boldness. By contrast, a traditional automaker — a German one for instance- will seek for the same level of core competencies, but with a different cultural fit: up there, the engineer will be asked to study good solutions, while someone else, higher in the ranks will decide upon it.
This principle leads to three key differences:
One, as we saw earlier with the crash-box example, Tesla is obsessed with speed — and it applies it to every operational decision.
Two, it leads to a much flatter hierarchy: At the time we launched the Model S, there were only two layers below Elon. Later when I joined Audi to build the e-tron, I was dealing with four hierarchical levels just for the engineering department, supplemented by two other echelons above. In short, we are talking of at least a 3 to 1 ratio. As a result, Tesla moves incomparably faster than Audi for instance. Where the Model 3 took 3 years in development, it would have been the customary 60 months time frame at Audi.
Three, the intensity of the workload at Tesla leads to higher turnover among executives and engineers. With a 27 percent replacement rate, it is even higher than in prominent startups like, for instance, Lyft (23 percent). And for Elon Musk’s direct reports the turnover hit a record of 44 percent last year, according to Alliance Bernstein tech analyst Toni Sacconaghi. We used to say that a year at Tesla equals seven years elsewhere, just like “dog years”.
In its quest for speed, Tesla has developed an approach closer to the tech world than heavy industry. It will not hesitate to make selective acquisitions if it means saving time (in software for instance), where a carmaker will not deviate from the “not invented here” principle.
I will conclude with a question I face quite often. Can Tesla maintain its lead or will carmakers like Volkswagen Group eventually catch up? Legacy manufacturers are without any doubt injecting agile methods in their processes. But, their key value remains the process, following operational experience from which strict rules and methods derive. The departure from this culture might take a long time. While young talented engineers joining these firms have been trained with new work approaches, chances are the bulky technostructure will curb their enthusiasm. In the meantime, I am pretty convinced that Tesla will do whatever it takes to retain its lead in key areas such as software, assisted driving capabilities (not Level 5 autonomy, we will discuss that later), performance, user experience, etc. The gap is here to stay. That explains why Tesla, with a market cap of $271 billion, vs $172 billion for Toyota, has now become the most valued carmaker in the world.
— Philippe Chain with Frederic Filloux
* Philippe Chain is an international expert on electric vehicles. Trained at a top French engineering school, as well as INSEAD and Stanford GSB, he has accumulated nearly 30 years of experience in the automotive industry. At Renault, Chain was a Chief Vehicle Engineer before switching to an electric strategist position. At Tesla, he was VP for quality during the development of the Model S, and later, he worked for Audi and for Faraday Future. Today, he assists several carmakers and governments willing to venture into the electric car industry.
The Monday Note is almost 13 years old and we are going to give it a refresh. As a first step, we would like to get your opinion about our work. Please take a moment to complete this short survey. This matters a lot to us.
Thanks! — Frederic
