-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
- Status
Very interesting work, Maekuz. I tried to dig into the source of the US figures. It is a consultancy called Ward's Auto. However, even after searching their site it is unclear how they have arrived at these numbers. I would guess that this would mean that these are registration numbers? Hopefully they are "clean numbers", i.e. individually verified. Sometimes consultancies do funny stuff to numbers - for instance they could have access to numbers for some states/countries and not others, and "correct" them by applying a factor to take into account the unknown territories.
You are coming up with estimates that are definitively higher than mine above. That is comforting, as I already feel that Tesla are doing well based on my own numbers.
You are coming up with estimates that are definitively higher than mine above. That is comforting, as I already feel that Tesla are doing well based on my own numbers.
Very interesting work, Maekuz. I tried to dig into the source of the US figures. It is a consultancy called Ward's Auto. However, even after searching their site it is unclear how they have arrived at these numbers. I would guess that this would mean that these are registration numbers? Hopefully they are "clean numbers", i.e. individually verified. Sometimes consultancies do funny stuff to numbers - for instance they could have access to numbers for some states/countries and not others, and "correct" them by applying a factor to take into account the unknown territories.
You are coming up with estimates that are definitively higher than mine above. That is comforting, as I already feel that Tesla are doing well based on my own numbers.
The data by WardsAuto is certainly a weak part. Rumor has it that WardsAuto get their numbers from Polk and Polk is known to rely on registration data. But thats about it. I would classify the US numbers as an indication. Its a guesstimate and it feels like WardsAuto (or Polk) are a bit more accurate than the Autodata numbers.
I guess the Q3 report will tell us a lot about the quality, since the European data should be quite accurate (except for including some loaners and showroom cars, which should be a relatively low number - I guesstimate we are talking ~20 cars in Norway, so maybe 50-60 cars for all of EU?).
Mario Kadastik
Active Member
So another data point for guidance numbers. According to the Munich event (11/20 Munich Tesla Event with Elon Musk) Elon's made now an official statement that battery commitment is there for up to 1200 cars / week production.
Assuming we enter 2014 with 650 cars / week (i.e. end of Q4) for pessimistic estimates and finish the year with 1200 cars / week with linear increase (each month the rate is increased by 50) we get 8400 cars in Q1, 10200 in Q2, 12000 in Q3 and 13800 in Q4 with total 44.4k.
Doing the same math, but assuming that the first half we increase the weekly rate every other month with third quarter increasing it every month and final quarter doubling the rate so somewhat a parabolic increase over the year we get to 8k in Q1, 8.8k in Q2, 10.2k in Q3 and 13.2k in Q4 with total 40.2k.
And now the reverse of the previous one with the parabolic rise happening early on (the most optimistic estimate) we get 9k in Q1, 11.4k in Q2, 12.8k in Q3 and 14k in Q4 for a total of 47.2k.
So just assuming various ramp-up speeds from 650 / week to 1200 / week we can get total FY2014 guidance between 40k-47k. Of course Panasonic having capacity for 1200 cars / week might not mean that we'll actually hit that number end of 2014 so take this as a mental exercise to give a ballpark measure.
Assuming we enter 2014 with 650 cars / week (i.e. end of Q4) for pessimistic estimates and finish the year with 1200 cars / week with linear increase (each month the rate is increased by 50) we get 8400 cars in Q1, 10200 in Q2, 12000 in Q3 and 13800 in Q4 with total 44.4k.
Doing the same math, but assuming that the first half we increase the weekly rate every other month with third quarter increasing it every month and final quarter doubling the rate so somewhat a parabolic increase over the year we get to 8k in Q1, 8.8k in Q2, 10.2k in Q3 and 13.2k in Q4 with total 40.2k.
And now the reverse of the previous one with the parabolic rise happening early on (the most optimistic estimate) we get 9k in Q1, 11.4k in Q2, 12.8k in Q3 and 14k in Q4 for a total of 47.2k.
So just assuming various ramp-up speeds from 650 / week to 1200 / week we can get total FY2014 guidance between 40k-47k. Of course Panasonic having capacity for 1200 cars / week might not mean that we'll actually hit that number end of 2014 so take this as a mental exercise to give a ballpark measure.
With regards to the decrease in Norwegian deliveries, I have another theory, that I would like to share.
I’m wondering if this “drop” in deliveries for Norway in October, has something to do with the new plant in the Netherlands that opened on the 22.ndAugust. The big rush in Norwegian deliveries was in late August and September, and then it became more silent. As you all know the cars to the Norwegian market are delivered directly to Norway, because we are not a “real” EU member. I expect it was as simple, as there was a need to fill up the plant in the Netherlands with enough cars to keep the workers busy, and simultaneously getting the deliveries out to please EU signature and early reservation holders, which leads to fewer deliveries in Norway.
Cut from Tesla’shomepages:
“With a size of 18.900 square meters, this new state-of-the-art facility iswell prepared to receive the brand new Model S which is shipped over from the US.Having crossed the ocean and reached Europe, Model S arrives at the Tilburg plant for finalassembly before being delivered across the continent.”
Just my thoughts about the decrease in Norwegian deliveries. Any thoughts?
Last edited:
Julian - I'm quite confident that Tesla did not enter Q3 at over 500 cars/week. In Q2, they most likely produced 5450 cars. There were 13 weeks of production in Q2. So 5450 divided by 13 = 419 cars/week avg in Q2. They started Q2 likely at slightly below 400/week and probably ended Q2 at 480/week at the most IMO. So I think using 550/week as a starting number for Q3 would not be accurate.
I'm also growing more convinced that Tesla did not reach 600 cars/week by end of Q3 (ie., DB report, Elon statements, and other).
There were also only 12 weeks of production (because first week of July was vacation for the factory).
So if Tesla started Q3 at 480/week and ended at 570/week, then the avg would be 525/week x 12 weeks = 6300 cars produced.
For cars delivered, I'm thinking of subtracting 500-600 cars for transit, loaners and store cars.
So, Q3 cars sold might be 5700-5800.
- - - Updated - - -
I think Tesla stopped delivery of European cars starting last week of August (or so) until the end of September so they could reserve all the cars produced during that time for U.S. customers to minimize time of transit and insure no (or little) cars were on the boat to Europe at end of Q3. Considering transit time, I would think actual deliveries to Europe in October would be very, very small. (Can someone in Europe confirm if October deliveries are minimal?)
I also think they restarted European shipments at the first week of October, so that we'll see another big increase in European deliveries starting November.
DaveT, As always, I, like others, always appreciate your input. Maekuz, did a nice analysis of registrations in the US and Europe for Q3 and comes up with a number of deliveries that exceeds your prediction by roughly 1,000. I claim no special insight and based on VIN assignments I believe the delivery total will be over 6,000 and perhaps even 6,500. Can you give us your reasons for not embracing Maekuz's numbers?
Thanks Al
Thanks Al
Johann Koeber
Happy Owner
Elon did say in Munich that the production is restrained by the parts situation. As an example he mentioned the batteries.
He also said his time at Tesla is split between ramping up production and designing model X.
My impression ist that he really is pushing hard to ramp up production, because all the cars Tesla can sell help improve the bottom line. This will be used to fund future developement and grow the company.
He also said his time at Tesla is split between ramping up production and designing model X.
My impression ist that he really is pushing hard to ramp up production, because all the cars Tesla can sell help improve the bottom line. This will be used to fund future developement and grow the company.
They take a week off every quarter, including the first week of April, IIRC. So 5450/12 = 454 car average. Also, I remember at least 2 people that took factory tours in June that reported ~540/week. There was also the Kimbal Musk tweet from 6/5/13 saying 80 cars per day, but I'm not sure if that is a 5 or 6 day week.
Do you have a link to the 540/week from 2 people on factory tours in June?
The tweet from Kimbal Musk could mean that Tesla was making 40 cars/shift (total 80 cars/day). 5 day work week (which I thought they were doing) would be 400 cars/week. 6 day would be 480 cars/week.
- - - Updated - - -
I appreciate Maekuz' work. I do have doubts on the actual data behind the numbers because it not available in "first data" form. Meaning, if we could see the actual registration data (for all states) then that would be a legitimate "first data" point. As it stands, the registration data that we have from the links in the Excel sheet is more of a "second data" point. Meaning, the sources aren't showing us the actual data but just giving us a summary number (ie., # registrations for a month). But we don't know exactly how they came up with that number as we're not privy to the actual hard data.
This is why I think while registration data like Maekuz' can be helpful in some ways, it's not the best way to gauge an accurate prediction of the actual # cars delivered in Q2. But if we can verify the actual hard data, then the registration data could be very accurate. So, I guess it's a matter of how the source collected and analyzed the data which we don't know and can't confirm the actual numbers.
In terms of "first data", VINs are a "first data" point because they're the actual hard data from users. (One of the problems with VINs is that delivery can happen 1-3 months after, or even longer if in Europe so it's difficult to use VINs to get an accurate # of actual deliveries in Q3.)
I also consider interviews from Elon Musk (or other management) as "first data" points as well. This is why his interviews in August stating they were slightly over 500 cars/week should be taken seriously.
The DB report which states that Tesla should reach 600 cars/week in Q4 is an interesting data point. It's not "first data" but it's interesting because as sleepy noted these analysts often do have access to management (and at least Tesla's investor relations) and often can confirm important "first data" points.
Top reporters (ie., CNN, Bloomberg, etc) often also have access to Tesla's PR lead and can confirm data points as well.
- Bloomberg noted on July 11th that production had passed 400 cars/week (Elon Musk: Tesla Model S Output Speeds Past 400 a Week - YouTube)
- CNN noted on August 5th, that Tesla started Q2 at 400 cars/week and ended it slightly below 500 cars/week (Tesla shares surge after surprise profit - Aug. 7, 2013 )
- CNBC noted on August 21 that Tesla had passed 500 cars/week production (Elon Musk: Tired but optimistic about Tesla's future)
There is also data points that people can't share because they might be under NDA (ie., factory tours).
Putting this altogether, I think Tesla began Q3 at under 500 cars/week and ended it at under 600 cars/week.
There has been talk of 700 cars/week (from conversations from store employees). I was open to this but now I think that these employees might have heard ORDER numbers not PRODUCTION numbers from Tesla. Meaning, Tesla probably shared with their employees that their at 700 orders week (ie., 500 north america, 200 europe) and would be at 800 orders per week in Q4 (and 1000 orders/week by end of 2014). I think some employee(s) might have confused this with production (thinking it was the same as orders) and shared that.
Just found this on the Q2 2013 Shareholder Letter (http://files.shareholder.com/downlo...9/Q2'13 Tesla Motors Shareholder Letter.pdf):
"During Q2, we improved our production rate by 25% from 400 to almost 500 vehicles per week."
Is there any reason to believe they didn't increase again by 100 to 600? So a 550 average * 12 weeks is 6600 produced. There's always some debate about produced versus delivered, but I'm not sure I see much reason to believe the quantity of unrecognized Q3 cars is all that much larger than those same types of cars held over from Q2.
So, continuing our discussion. If we can say that we exited Q2 just under 500 (if not 500 per week) and exited Q3 at a production rate of roughly 600 could we justify a conclusion more in line with both Maekuz's Registration analysis and CfOH's VIN analysis and come up with 6600 production (12weeks * 550/week) with 500 in transit and come to a reasonable consensus of methods to predict 6000+ deliveries for Q3?
I agree with this and that is why I have not re-posted my article, since the store employee seemed a little too optimistic. With the exuberance running at an all-time high, I did not want people to get even more excited and put their whole life's savings into deep OTM Nov. call options (sounds like something I would do :wink: ).
I just wanted to add that there is evidence that Tesla did in fact build 550, 550, and 600 cars in the last three weeks of Q2. This data comes from witnesses on the factory tours. It looked to me as if Tesla was doing 400/week and then all of a sudden a supplier constraint got solved and they were able to ramp up to 500 virtually over night. That 600 number was probably an anomaly though.
Precisely. They achieved that in Q2, but when Elon was speaking he was talking about The current state (Q3).
That's what I originally thought (ie., last week)... 500 beginning of Q3, 600 end of Q3... avg 550 x 12 weeks = 6600 production. Then minus 500-600 for transit, loaners, store cars.
But I'm actually now doubting if Tesla actually reached 600/week in Q3, and thinking it could be a goal that they're targeting during Q4 (ie., DB report says this).
Also, regarding CfOH's VIN analysis, it's tough to use his data to come up with an accurate Q3 production or delivery number. His data IMO is helpful to see the pace of orders more so than deliveries or production. The reason being is that often a VIN is assigned not to long after an order is confirmed, but delivery can take any where from 1-3 months(?) depending on country and also when you ordered. And delivery times might be getting longer as well (but this doesn't show up in CfOH's analysis since he tracks just the initial VIN date and not production or delivery dates).
Whatever the number is, it should really reflect getting ready for their annual target for 2014. If Tesla wants to deliver 50K cars in 2014, 6 to 7 K deliveries in Q3 should provide them the appropriate growth rate needed to make that target for 2014 (they should be able to achieve a 20 to 25% q over q growth rate in production). Any more will prepare them for Model X and any Model S surge from China, etc for 2014. Important is gross margins for q3 and guidance.
- Status
Similar threads
