The period of time discussed in the article is one in which I feel I have a great deal of insight to offer.
My sense is that the author is talking about some real events that occurred, but the timing is a bit smushed around, and maybe some of the terminology being used is causing confusion as well. Also, I think that everyone involved is mythologizing and potentially blowing things out of proportion, the author included.
Anyways, I don't have immediate access to my notes, so this is all off the top of my head, to take it with as many grains of salt as you wish.
March Slowdown Contradicted By Facts
First, the author is talking about a sales slowdown in March. This obviously didn't occur in the technical sense that we think of here on this site, since we are familiar with the Tesla business model. Sales are only "closed" at, or shortly before delivery, and so by definition March of 2013 was a banner month which capped off a great quarter.
I was directly tracking sales at the time, and prior to the conclusion of the quarter in late March I saw an opportunity to change the prevailing media narrative, so I published a
blog entry about this success, and then followed up with successful campaign to get the story out into the media (as I told the site admins prior to publishing the blog, I thought it only fair to publish on TMC instead of a regular news site because all of the data came from here).
Deferrals Were A Major Issue
However, there was another major Tesla story that was also apparent in the data during that quarter and the quarter following (which I did not feel like widely publicizing because it was not a positive story), and that was what appeared to be mass deferrals of existing reservations.
Without my notes and with the passage of time fuzzing my memory, I hesitate to discuss the underlying data that proved these deferrals occurred (or at least made it by far the most likely explanation for what was happening), but it did happen (I'll find the core links to my original posts if folks wish to debate this, but
here are
examples of the discussion, and the drop in wait times and the fact that the deposits were still on the books was fairly conclusive evidence that these were deferrals and not cancellations).
The cause of this was uncertain, and the ones discussed in the article are probably part of a spectrum of causes. My recollection was that the possibly illegal service contracts and how they implied that Tesla was not prepared to honor their warranty was a big issue that sort of encapsulated the many smaller issues.
Regardless, there was endless drama at the time and my sense in Q3-2012/Q1-2013 was that Tesla was acting like a company that didn't believe in itself and was setting itself to chisel as much money out of a small but devoted customer base as it could (a controversial statement maybe for Tesla fans, but as much as I love Tesla the whole service contract/warranty issue was evidence of pathology. I'll also point out that whatever I felt about Tesla's actions and what they indicated about their own level of confidence, *I* had confidence in their product and invested a substantial amount of money in them).
Anyways, mass deferrals was a fact. It happened, and it was a matter of concern to me at the time. The discussion in the article on this point rings very true to me, and the fact that it specifically mentions deferrals rather than cancellations goes pretty far in satisfying an itch that I've scratched for some time.
April Slowdown - Closed Factory
Next point. The so called March sales slowdown was bunk. The one in
April was very real. I listed 1,300 sales in April, but the data was very sketchy for the month because Tesla had done some batches severely out of order (note my
preliminary results were 1,700 (which was 90% assumption), but every time I looked back at it with more resolution sales looked weaker, and my strong recollection is that even 1,300 was giving Tesla the benefit of the doubt because I could not conceive of a reason for the factory to be shut down for more than a week).
The factory clearly did shut down at the beginning of the month, though I attributed it to the paint switchover and just assumed it did not last longer than a week. Based on this article I think it probably did last longer than a week, and my sense at the time that even 1,300 sales was being too generous was probably correct.
(Note: over long periods of time, tracking sales using VIN's was very accurate, but slicing and dicing the data into smaller time periods like a particular month, was somewhat less accurate.
Because of the peculiarities of the data in April I could not state how many deliveries there were with any precision at all, but I could do it with
high accuracy over the time period of April to the first half of June, after which I was able to use a simple assumption to nail the actual production results of that quarter. However bad the April data was, May had a bit more confidence, and by mid June the issues had resolved, and the length of time it took those issues to resolve was the key indicator for an extremely weak April, a modest May, and a strong June.
Claiming 1,300 sales in April, 1700 in May and 2150 in June came from assuming week long shutdown at the beginning of April, followed by a production rate of ~400/week through May, and then upped to ~500/wk in June. But Tesla had clearly demonstrated a production rate of 500/wk in March, so I was further assuming that they had accomplished that with extraordinary means (overtime, etc) that they would prefer not to do on a regular basis.
But it was fundamentally a Tesla friendly assumption. The actual data points to most of the deliveries actually occurring after some time in mid May. If I just weight the data the way I see it it could easily fit a production rate from the start of April to mid May of ~275/wk (ignoring possible shutdowns) followed by an increase to ~500+/wk for the rest of the quarter. I do know that in early May I was shocked by the fact that Tesla hadn't overtaken my outliers from March and the very low production rates it potentially implied.
I stopped tracking sales at the end of June because it was clear that Tesla had started intentionally assigning VIN's in a way designed to defeat tracking efforts.)
Underlying Reservation Rate
What this doesn't tell us is how many new reservations were coming in. During Q1 Tesla aggressively worked to shut down our information channels that gave us insight into reservations. I maintained some insight into underlying reservations in Europe for some time afterwards thanks to an oversight on their part, but by the end of Q1 I am confident that we were effectively blind in terms of tracking new reservations in the U.S.
Fortunately, it didn't matter. Wait times dropped to a level consistent with there being no real wait list (which proved the mass deferrals, possibly 10k+ if I am doing the quickie analysis from memory correctly), so at that point deliveries were a direct proxy for new reservations, and so we could directly sample demand based on deliveries for the quarter. The whole notion of "production constrained" is misunderstood I think.
Tesla was making only a modest effort to make sales, and that effort was scaled to their production capacity, but after Q1-2013 it was never the case that there were huge numbers of people waiting in line for the car (at least in the U.S. where deliveries were actually occurring). Elon also gets an out on the whole "production constrained" issue because at the time there were also global wait lists that Tesla was not in a position to satisfy thanks to production constraints, though in that case it was more the case that they were not set up yet to produce those cars.
So What Really Happened?
There is a lot of data that could be used to support the general narrative in this story. I doubt that Tesla was ever in imminent danger of bankruptcy in March and in April their stock soared after they released their production results, and as a result of the greatly improved media environment, which gave them access to vast amounts of capital that were just unrealistic in March.
But it seems reasonable to me to think that Tesla was under intense stress, and the request for a $5b capital injection to build the gigafactory makes a lot of sense to me from a strategic standpoint. Until April 2013, a ~$5b capital raise was probably pure fantasy. Elon certainly didn't have the cash. His billions were much more of a theoretical exercise in early 2013, and not remotely liquid, so I could see Elon scheming to make a deal like this before the state change in April.
All of the color about executives not telling Elon "how bad things were" sounds like a bunch of nonsense. I'm sure there were some employees who felt the situation was dire. But Elon probably had a better sense of reality than they did, and their freakout about deferrals and his response to it sort of confirms it for me. These were deferrals, not cancellations. Fundamentally these people still wanted to buy the car, but were just on the fence for whatever reason. Pushing to close those sales is just CEO 101.
The bigger issue that is potentially buried in the story is the potential impact on NEW reservations. Based on April (and May, which wasn't particularly strong either) sales it seems clear that they had cratered as well. Working to close sales from existing reservations is one thing, but that is a finite resource. In the best case, closing all of those sales could not have carried Tesla for more than a couple of quarters. So we know that at some point the underlying reservation rate came back up.
My guess is that it came back up as a result of the improved media environment in April, when Tesla suddenly became a media success story, while Tesla simultaneously worked to eliminate the underlying problems with service and warranty concerns. I'm actually skeptical that closing sales on deferrals was ever a huge factor. The math on deposits never did seem to indicate a major movement to resolve deferrals.
But a quick uptick in new reservations following the positive media environment of early April is very consistent with a major pickup in production in mid May.
Edit: I should point out that the last time I tried to analyze reservations, the math still didn't seem to resolve unless you assumed there were still thousands of deferred $5k Model S deposits sitting on the books.
