Skip to content

TESLA manufacturing issues?

April 2, 2018

In this post I want to comment on recent concerns about Tesla’s ability to ramp up manufacturing of its Model 3.

(1) Is the the quarter-million-per-year production capacity goal appropriate?

As with any pure EV, the ideal use case is: The second of two vehicles for a comfortable suburban household, for use in daily commute and charged overnight at home, while the other conventional or hybrid vehicle can be used for trips that are longer or break the daily pattern.

The price/prestige level of the Model 3 seems to straddle the bottom end of the aspirational-luxury category, and the top end of the sensible-car category. Despite statements to the contrary, it’s definitely a kind of status symbol for a middle class buyer, like the Iphone was when it was new.

The brand appeal is key – buyers motivated strictly by utility have more economical options, both in pure EV’s and even more so in hybrids. Car buyers with a single vehicle (i.e., single people or households with 1 vehicle) are going to be few and far between.

Point being, even though it’s clearly not as broadly useful as a liquid fueled or hybrid vehicle, the market for a premium EV is more than big enough to absorb a run of several million of these things over the life of the vehicle design (nominally 5 years between “major” redesigns in the auto industry). So I think their production capacity goal of a quarter to half a million per year are entirely appropriate.

(2) Was the decision to go with the robot-porn approach too much?

I’d argue that for a US company, extensive automation is the way to go. US management is really not good at handling relationships with workers. Too often it becomes abusive, in which case quality goes downhill. Nasty as it is, making gigantic investment in all-seeing all-knowing automated production systems is a good fit for how we do business in this country.

The other thing that has to be there is quantity. As a rule of thumb, quantities of 100k assemblies per year don’t quite justify custom tailored end-to-end-full-automation level of investment. At that level you would automate individual process steps and have operators fill in the gaps where it’s more economical. At a million per year, you talk about true full automatic. With a product as complicated and oddly shaped as an automobile, the industry doesn’t even do it at that level. But Tesla and Elon Musk being who they are, it’s go big or go home. With a premium product, you can get away with it.

Also keep in mind that most of the labor actually isn’t in the final assembly anyway, it’s in the myriad of subassemblies. We don’t see or hear much about how the battery packs or drivetrain components or electrics or accessory assemblies or wire harnesses are put together. For all we know, the wiring and seats and climate control parts are made in the same shops in Mexico that makes them for all the other North American built cars.

What we are shown are shiny new robots picking up car bodies and moving them around and installing doors or windshields. That’s nice, and maybe a bit of an impractical show-off move. But I have a feeling that the “production hell” isn’t actually due to getting the shiny new robots to pick up and put down parts in the right place.

Not that it’s trivial- I participated in an automation project that had a couple of robot arms at the center of the whole mess. There’s a lot of fudging and tweaking to get the robotics right — the nominal specs the manufacturer gives you make some subtle assumptions that have to be wroked through by the automation engineers doing the initial set-up of the automation. The production engineers or techs operating the system then have to know enough to understand which of the resulting fudge factors to touch when the system needs to be tweaked 6 months down the line.

I think Tesla is well equipped to overcome the technical/engineering challenges, and I suspect they are NOT the main part of their current struggles. It’s just something a journalist will focus on because a misaligned door or body panel is easy to understand, compared to a logistical SNAFU that involves 3-4 tiers of suppliers, managing phase-in of engineering change orders, mixtures of old and new parts, long lead times, unwieldy quality/information systems, personnel changes, project (mis)management and development resource (un)availability, hardheaded procurement departments, and so on.

In other words, it *is* the “machine that builds a machine”. That machine isn’t a robot, it’s a living breathing business, a tech development, manufacturing, documentation, logistics, and quality organization. And besides that, a social organism. They’re building a business. That’s what is not going as fast as they hoped. Of course adding an entire additional layer of automation to this already major challenge doesn’t help.

(3) Will Tesla overcome its manufacturing capacity shortcomings in a timely manner? Would it be a good thing if they did so at all costs?

Depends on your definition of timely, eh? In practice, that means before the financial pressure becomes too great, and before their brand value erodes – a thing of tremendous value Tesla has right now. In terms of financial pressure, they have a good 12 to 18 months to get to 5000/week??? Just based on looking at their debt schedule (see previous post). As long as they make steady progress to that, if they are selling at that pace a year from now, they’ll be able to borrow more and keep going just fine. If they make un-steady progress, the market will punish their stock, and in turn the media will punish their brand image.

Potential damage to brand value is by far the deadlier killer. They need to maintain the reputation of their product. I think slowing down production is the lesser of two evils in that regard. Of course having the autopilot crash a car into a large non-moving object is a painful blow there. But that aside, for a half-finished automated manufacturing situation, increamental steps are the way to go.

Since they planned their product flow for full auto from the beginning, they already have the process broken down into well sequenced steps. To build up the rate of final assembly to their target rate, they do this: Some of the steps, you pull the robot and put in workers. Maybe the whole line slows down a bit. You eat that for a while. One by one, you get the stations to work as originally intended.

However, I don’t think it’s a matter of final assembly anyway. As I said above, it isn’t getting the robots to work together, it’s getting the *company* and the supply chain and the design / manufacturing / quality / docs / management / change / project-management parts of the organization to work together. Oh and very likely, getting the product design itself right.

In all of the above, I think Tesla has 3 advantages over their competitors, and 1 disadvantage. The first advantage is again the prestige of the company name. They will actually be able to hire an “A” team. A lot of manufacturing organizations get “B” level tech staff just because the work isn’t sexy. The second advantage is they’re starting from scratch. They poach experienced staff in positions where they want that, they can hire young-innocent-and-enthusiastic staff in positions where that is beneficial. Clean slate. The third advantage is they can throw money around like they’re a silicon valley startup. That helps in the beginning (at the risk of developing some really bad habits).

The big disadvantage is also one of the advantages. Starting from scratch. Big car companies had decades to “build the machine that builds the machine” (not a robot, remember). Often growing it organically.

(4) Uneducated gut feel

I think Tesla will continue to ramp up production and reach their goals. Late. But they will get there.

If you see them “aggressively cutting costs” *before* they had a chance to borrow more money and expand the service side of their business, you know they didn’t make it.

As an aside, about “building the machine that builds the machine”. If you are deep in chronic crisis-mode when building your organization, your organization will not come out healthy. That is a risk for them, but it would be something that manifests itself in future generations of product.

(5) Postscript : Do I even like Tesla?

Kindof yes, kindof no. They made EV’s cool for the luxury car buyer, and therefore for the entire market. That’s a huge accomplishment. On the other hand, the use case for the EV’s they’re building is suburban commute but better for the environment. There is already a solution to that: multi-occupant vehicles. Buses. Vans. Trains. Park-and-Ride’s are especially perfect for suburban life. These could exist in both luxury and unwashed-masses mode. No new tech required. So I’m constantly disappointed by the opportunity cost to our society, by hyping a solution to a problem that’s already been solved. But this post isn’t about that, it’s just about whether I think they’ll make it business-wise and manufacturing-wise.

(6) Additional Resources

Review of the Model 3:

https://www.caranddriver.com/reviews/2018-tesla-model-3-test-review

Current rate of Model 3 manufacturing:

https://www.bloomberg.com/news/articles/2018-03-23/tesla-shows-signs-of-a-model-3-surge

https://www.bloomberg.com/news/articles/2018-03-29/tesla-urges-workers-to-prove-haters-wrong-ramp-up-production

Video, Tesla Model 3 production promo:

Video: Model S production (mostly body, not much final assembly shown, unfortunately):

Video: Nissan Leaf production promo:


Disclosure- I am neither long nor short Tesla or any of its competitors, partners, suppliers etc. – in any form including stocks, bonds, options, business relationships etc. This is an unpaid piece.

From → Uncategorized

3 Comments
  1. I’m impressed with the depth of research in your recent posts.

    • I had a brief and misguided career detour trying to do short-term stock trading (not worth the risk and stress). But it ended up being tremendously educational for all kinds of detail having to do with the energy business, and also a crash course in corporate-finance shenanigans.

      Lately the place I work has been getting involved with clients doing industrial automation – the Tesla story kindof brings it all together.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: