Demand

[bonaire]

Someone posted here that the lot at Charlotte had appx 300 cars. But there are off-site storage fleet sites too, like Lathrop.

Dennis has good number estimates but the amount of service loaners is quite large and if you compute in the old $2000 per car for the supercharger network, I assume there is a similar, if not more, $3000 per car to handle service fleet costs that can be reduced as the cars are sold and the miles driven on them written off along with the discounting for age and other changes needed to complete the sale of the loaner vehicle. Dilution of the fleet valuation through discounted loaners does impact the eventual trade-in and value of used cars longer-term. They're able to sell more so the uniqueness drops as well so I don't get why someone wants to pay full price when buying "their own" custom order other than "they made it for me!" novelty. I have to think if the 12000 number is true, there's a car that matches nearly everyone's configuration somewhere and at a reasonably savings. Someone else posted here that after the car was on the lot for 30 days, the sales associated was sort-of surprised that it had an automatic $3600 discount.

Tesla can sell more off the inventory by listing it out in larger volume than they do. Making it seem like there is "light" inventory keeps what they actually do have in inventory quietly silent. And why charge $2000 transit fee for a car that is still in Fremont for someone who actually does pick up one off the inventory list? Ship it for free as if it were a custom order. I suspect the web site changes coming are possibly intended to allow for better inventory searching that will include new and loaner-program units.

In terms of Model X inventory, Jon McNeil's claim for MX inventory may have been the vin #s planned out here - can you spot the spike? two 1000 Vin # blocks done in one day each. 52xxx and 53xxx on 5/20-5/21. Otherwise, average weekly vins following that have been 700wk. Waiting to see when 58999 comes about.

 

[vgrinshpun]

So what is the "correct" number of cars for Tesla to hold in inventory for display/test drive, service loaners and purchases "off the lot" instead of custom orders? Let's try some first principles analysis.

Looking at the Tesla website, they have 84 service centers in the US and 82 internationally, for a total of 166. Each service center would require 10-50 service loaners, depending on size. I know of one particular service center has a goal of completing 30 vehicle services per day. So let's say an average of 25 loaners per location which would account for 4000 cars. Another way of getting at this is with a fleet size of 250,000 cars, if an average vehicle requires 1 day for an annual service and 2 days per year for some other issue with 250 work days that would require 3000 cars in the service loaner fleet at 100% utilization. Conclusion: 3000-4000 vehicles in the service loaner fleet.

There are 110 US sales locations, so I will assume the same ratio for international for a total of 220. I initially thought that each sales location would require 4-20 cars for display and test drives. At an average of 10, that accounts for 2000+ cars. However, it was pointed out that Jon McNeil said they were adding 1000 Model X's to the demo fleet. Since they already have some X's out there and a full complement of S demo cars I now believe there are 3000 cars in the demo fleet. The reason for this is that Tesla is using "overnight test drives" for qualified prospects as a way to convert them into buyers. An overnight test drive car only services one prospect in a day, whereas a normal test drive car might serve 8-10 prospects daily.

What about brand new cars to be purchased from inventory? Tesla sells about 25,000 cars per quarter. If we assume that all cars were purchased from inventory, 2 week delivery from factory to sales location and 2 weeks that the vehicle is on the lot, that would require 8000 cars. We know that there is some split between custom orders and purchases from inventory, but don't know what that split percentage is. If we assume 50%, that would mean Tesla needs to carry 4000 new cars in inventory. This number is the most speculative because we don't know the % split or the average days on the lot.

So adding these together Tesla needs a minimum of 3500 service loaner + 3000 demo + 4000 new inventory = 10,500 cars to be carried on the books at Tesla's current level of sales and fleet size. And that assumes a perfect distribution by location, vehicle configuration, etc.

I did similar analysis couple of days ago, complete with the spreadsheet of the count of global stores and service locations by country, but did not have time to organize and post it, - plan to do it later this week. One thing that is missing in your very good take on this, however, is the pipeline. With certain rate of cars sold every quarter and a certain time cars spend in transit, there inevitably a sizable chunk of the "produced but not sold" cars that are actually in the pipeline. These are not addressed by Tesla press releases as they cover only cars in transit to customers that placed an order, not cars in transit to replenish the sold out units from the demo/loaner/inventory fleet.

Other than that you are absolutely correct: throwing around seemingly large quantity of "produced but not sold" cars could be misleading without analyzing the actual reasonable needs for demo/loaner/inventory fleet based on the sales and service locations. I know you actually did not state it in this way, but this is the real reason for your post, I believe .

 

[bonaire]

The Produced but not Sold carries weight because those are the base-case metrics of a production company. Pipeline is meaningless if the units being sold by a producer are not sold. Three metrics are now the public info - Production count, Delivery count and "in transit". The in transit is also somewhat misleading (maybe intentionally?) because there are always cars in transit. Why are they in transit while during the quarter inventory cars were being built? Why not build all custom deliveries first and then build the inventory? It does sound like there is intention to have cars in transit at the end of the quarter because they are "locked in by contract" and guaranteed sales while custom orders are pipe replenishment. It may be better for the aggregate sales model to build inventory cars ahead of some customers' orders - but it does make them wait longer and you end up with the final week of quarter "push" to get sales done. It could be smoother.

 

[Krugerrand]

Someone posted here that the lot at Charlotte had appx 300 cars. But there are off-site storage fleet sites too, like Lathrop.

Lathrop is not an off-site storage fleet location. It's a castings factory with parking for employees. Nothing more, nothing less.

 

[dennis]

Lathrop is not an off-site storage fleet location. It's a castings factory with parking for employees. Nothing more, nothing less.

There are also a lot of Model S's in the parking lot at the Fremont factory. I wonder if Tesla is hiding inventory there?

 

[Krugerrand]

There are also a lot of Model S's in the parking lot at the Fremont factory. I wonder if Tesla is hiding inventory there?

Clearly that's why there's been major parking issues for employees.

 

[schonelucht]

Why are they in transit while during the quarter inventory cars were being built? Why not build all custom deliveries first and then build the inventory? It does sound like there is intention to have cars in transit at the end of the quarter because they are "locked in by contract" and guaranteed sales while custom orders are pipe replenishment. It may be better for the aggregate sales model to build inventory cars ahead of some customers' orders - but it does make them wait longer and you end up with the final week of quarter "push" to get sales done. It could be smoother.

This does not make sense. When you build all custom deliveries at the beginning of the quarter only, you will also make some customers wait longer than necessary : those that order right as you switch over to inventory. Also I don't understand why having cars in transit over a quarter would be a bad thing?

Best is simply to produce an ordered car as soon as it is confirmed and let the pipeline grow to whatever it was. At some point Tesla tried to move to this model, but finally abandonded it.

I did similar analysis couple of days ago, complete with the spreadsheet of the count of global stores and service locations by country, but did not have time to organize and post it, - plan to do it later this week.

Interesting, please mention me in the post so that I don't miss it! Also, did you try to estimate the number of CPO cars used as loaners? I believe this is a substantial minority of the number of loaners in use.

One thing that is missing in your very good take on this, however, is the pipeline. With certain rate of cars sold every quarter and a certain time cars spend in transit, there inevitably a sizable chunk of the "produced but not sold" cars that are actually in the pipeline. These are not addressed by Tesla press releases as they cover only cars in transit to customers that placed an order, not cars in transit to replenish the sold out units from the demo/loaner/inventory fleet.

True. It happens quite often that Europen inventory pops up on the website for sale while they are actually still in transit. Have you tried to estimate how many cars are in this category? Curious how to get a handle on this number.

 

[Starno]

When the next economic downturn will happen, demand will probably fall, and pick it up after 18 months - 2 years. I hope Tesla will be well prepared to manage a 2 years weak demand.
It will probably happen by end of 2019 - early 2021.

 

[neroden]

So what is the "correct" number of cars for Tesla to hold in inventory for display/test drive, service loaners and purchases "off the lot" instead of custom orders? Let's try some first principles analysis.

Looking at the Tesla website, they have 84 service centers in the US and 82 internationally, for a total of 166. Each service center would require 10-50 service loaners, depending on size. I know of one particular service center has a goal of completing 30 vehicle services per day. So let's say an average of 25 loaners per location which would account for 4000 cars. Another way of getting at this is with a fleet size of 250,000 cars, if an average vehicle requires 1 day for an annual service and 2 days per year for some other issue with 250 work days that would require 3000 cars in the service loaner fleet at 100% utilization. Conclusion: 3000-4000 vehicles in the service loaner fleet.

There are 110 US sales locations, so I will assume the same ratio for international for a total of 220. I initially thought that each sales location would require 4-20 cars for display and test drives. At an average of 10, that accounts for 2000+ cars. However, it was pointed out that Jon McNeil said they were adding 1000 Model X's to the demo fleet. Since they already have some X's out there and a full complement of S demo cars I now believe there are 3000 cars in the demo fleet. The reason for this is that Tesla is using "overnight test drives" for qualified prospects as a way to convert them into buyers. An overnight test drive car only services one prospect in a day, whereas a normal test drive car might serve 8-10 prospects daily.

What about brand new cars to be purchased from inventory? Tesla sells about 25,000 cars per quarter. If we assume that all cars were purchased from inventory, 2 week delivery from factory to sales location and 2 weeks that the vehicle is on the lot, that would require 8000 cars. We know that there is some split between custom orders and purchases from inventory, but don't know what that split percentage is. If we assume 50%, that would mean Tesla needs to carry 4000 new cars in inventory. This number is the most speculative because we don't know the % split or the average days on the lot.

So adding these together Tesla needs a minimum of 3500 service loaner + 3000 demo + 4000 new inventory = 10,500 cars to be carried on the books at Tesla's current level of sales and fleet size. And that assumes a perfect distribution by location, vehicle configuration, etc.

Good work. I think you have to estimate that each service ends up requiring two days of loaner -- even if you drop the car off and pick it up the next day. If you estimate that each car has 1 annual service and 1 additional issue per year (which feels about right based on my experience) that's 4 days of loaner per year; I think 4000 loaners is more likely (though we know they've been short on loaners). However, worth noting, the total fleet on the road keeps increasing, so they need even more loaners. So we're getting to a number which roughly equals the measurable inventory. (Since they list the demo cars and loaners in "inventory" as well as listing the deliberate for-sale inventory.)

 

[neroden]

. They're able to sell more so the uniqueness drops as well so I don't get why someone wants to pay full price when buying "their own" custom order other than "they made it for me!" novelty.

Some of us are very, very picky. I demanded flat white paint, with *white* roof (no sunroof), light colored interior, tech package, dual chargers, fog lights, etc. etc. I actually still haven't seen an inventory car which exactly matches mine. Now that's why you'd get a custom order.

After experiencing this I am never buying an off-the-lot car again. I'm tired of compromising. I will get the interior color I want, the exterior color I want, and the exact list of options I want.

I can state with absolute certainly that amongst the inventory, the specific collection of options I want will almost always be absent.

Now, on Model 3 with practically no options, that's going to be less true. But on any car with a lot of options they are NOT going to have all possible configurations in inventory: it's a combinatoric explosion, too many configurations. It won't happen.

 

[vgrinshpun]

Some of us are very, very picky. I demanded flat white paint, with *white* roof (no sunroof), light colored interior, tech package, dual chargers, fog lights, etc. etc. I actually still haven't seen an inventory car which exactly matches mine. Now that's why you'd get a custom order.

After experiencing this I am never buying an off-the-lot car again. I'm tired of compromising. I will get the interior color I want, the exterior color I want, and the exact list of options I want.

I can state with absolute certainly that amongst the inventory, the specific collection of options I want will almost always be absent.

Now, on Model 3 with practically no options, that's going to be less true. But on any car with a lot of options they are NOT going to have all possible configurations in inventory: it's a combinatoric explosion, too many configurations. It won't happen.

Ditto here.

 

[vgrinshpun]

Interesting, please mention me in the post so that I don't miss it! Also, did you try to estimate the number of CPO cars used as loaners? I believe this is a substantial minority of the number of loaners in use.

I plan to post later in the week, but to quickly answer your question about the CPO loaners, I dropped off my car for annual and couple other issues yesterday and talked about this with two people at the service center. I was told (and shown a screen depicting it) that this, what I guess slightly smaller than average service center, has 15 loaner cars. About two month ago they were all exclusively CPO cars, but since then Tesla is slowly replacing them with new cars. At this point, I was told, approximately half of loaner cars are CPO, half new cars, mostly MS, with two out of 15 being MX.

 

[dennis]

Currently, Tesla has 12000 new cars in inventory. Not sure how many CPO's on top of that, but at least a couple thousands. What do you expect for this quarter in terms of production/delivery differential?

I expect Tesla to produce 25,000 cars. I also expect them to work to create demand for 25,000 cars. For example, after the Model 3 reveal on 7/28 they could announce the reservations number, that they are sold out until 2019, and offer 2 year leases on S75's.

 

[bonaire]

Lathrop is not an off-site storage fleet location. It's a castings factory with parking for employees. Nothing more, nothing less.

Why do pages like this indicate, in the Source code, a fleet location of Lathrop? Hundreds of them. Do they store cars inside the building? many P100D Model X.

www.tesla.com/new/5YJSA1E26HF204448
and
www.tesla.com/new/5YJXCBE46HF052839
view page source
name="vehicle-fleet-assigned-location" value="NA-US-CA-Lathrop-Murphy Parkway"

 

[bonaire]

Bunch of MS 75 (not D) being made now in the 208xxx series. I am thinking price drop coming in the near-term.
And looks like now using Westside Group Logistics (Torrance?) to locate some of them.

| Tesla
| Tesla

Of course, many also listing out as Lathrop.

 

[schonelucht]

@vgrinshpun mentioned Tesla wants to have a large number of model S's available (possibly at slight discounts) for Model 3 wannabe owners that are tempted to upgrade to avoid the queue. Seems very plausible to me. In fact, I think GM is hoping the same thing with their large inventory buildup of Bolts instead of delivering them to Europe. News out of Norway is that Bolt deliveries (just over 700 so far) will slow to a trickle before a new batch arrives at the earliest in August.

 

[colettimj]

Bunch of MS 75 (not D) being made now in the 208xxx series. I am thinking price drop coming in the near-term.
And looks like now using Westside Group Logistics (Torrance?) to locate some of them.

| Tesla
| Tesla

Of course, many also listing out as Lathrop.

I can't see any Model S 75s when I search "New" from the Tesla site. Odd.
Ev-cpo, however, shows over 200 75s as "New"

 

[HankLloydRight]

Ev-cpo, however, shows over 200 75s as "New"

I think you're looking at both 75 and 75Ds. There are about 82 75's.

 

[mmd]

So what is the "correct" number of cars for Tesla to hold in inventory for display/test drive, service loaners and purchases "off the lot" instead of custom orders? Let's try some first principles analysis.

Looking at the Tesla website, they have 84 service centers in the US and 82 internationally, for a total of 166. Each service center would require 10-50 service loaners, depending on size. I know of one particular service center has a goal of completing 30 vehicle services per day. So let's say an average of 25 loaners per location which would account for 4000 cars. Another way of getting at this is with a fleet size of 250,000 cars, if an average vehicle requires 1 day for an annual service and 2 days per year for some other issue with 250 work days that would require 3000 cars in the service loaner fleet at 100% utilization. Conclusion: 3000-4000 vehicles in the service loaner fleet.

There are 110 US sales locations, so I will assume the same ratio for international for a total of 220. I initially thought that each sales location would require 4-20 cars for display and test drives. At an average of 10, that accounts for 2000+ cars. However, it was pointed out that Jon McNeil said they were adding 1000 Model X's to the demo fleet. Since they already have some X's out there and a full complement of S demo cars I now believe there are 3000 cars in the demo fleet. The reason for this is that Tesla is using "overnight test drives" for qualified prospects as a way to convert them into buyers. An overnight test drive car only services one prospect in a day, whereas a normal test drive car might serve 8-10 prospects daily.

What about brand new cars to be purchased from inventory? Tesla sells about 25,000 cars per quarter. If we assume that all cars were purchased from inventory, 2 week delivery from factory to sales location and 2 weeks that the vehicle is on the lot, that would require 8000 cars. We know that there is some split between custom orders and purchases from inventory, but don't know what that split percentage is. If we assume 50%, that would mean Tesla needs to carry 4000 new cars in inventory. This number is the most speculative because we don't know the % split or the average days on the lot.

So adding these together Tesla needs a minimum of 3500 service loaner + 3000 demo + 4000 new inventory = 10,500 cars to be carried on the books at Tesla's current level of sales and fleet size. And that assumes a perfect distribution by location, vehicle configuration, etc.

The fatal flaw in your calculation is the assumption that every (or most) service visit requires a loaner. I guess you and Tesla have not heard of "customer waiting" or "customer drop off". Those are quite common in traditional dealerships. No one is foolish enough to lend out new cars for routine service that earns them $20-$40 or sometimes $0.
Do you recall Elon's big talk about servicing the cars as fast as in car racing? Then, most services shouldn't require a loaner. In my experience, 90% of the time people just wait 1-2 hours, sip coffee, browse internet/watch TV, then they are out. It should be more true for Tesla services; as they are more sparse, the driving time back and forth is likely much higher than waiting for 1 hour aat the service center.
May be, Tesla service is inherently more complex than service of most other cars. Or, Tesla service is grossly incompetent. Or Tesla cars require a lot more service or more severe service/repairs than other cars. Which one is it, I don't know.
More likely, IMHO, is that Tesla lends out the new cars to keep the excess inventory busy and depreciate them for discounting.
So, I think you folks have it all wrong. The need to lend out cars is because of excess inventory, not the other way around.

About the nightly test drives: A good demand lever will be month long test drives. Want to try out electric cross country driving first? No problem, get a brand new Model X and drive it around for a month to decide. This can justify a 30 fold increase in inventory, and also allow to borrow 30 times more on the ABL line.

Certainly. Tesla has a lot more insight here so it's quite plausible there are some we are both missing. @vgrinshpun mentioned another strong one (federal tax credit expiry).

I do not think this ~7% federal tax credit refund for $100k ASP Model S/X is as big a deal as in Hong Kong or Denmark, where the prices almost doubled. I don't expect much bump in Model S/X demand in US because of this. Elon himself said in the last earnings call, that these tax credits and subsidies are not helping Tesla sales.

 

[dennis]

The fatal flaw in your calculation is the assumption that every (or most) service visit requires a loaner. I guess you and Tesla have not heard of "customer waiting" or "customer drop off". Those are quite common in traditional dealerships. No one is foolish enough to lend out new cars for routine service that earns them $20-$40 or sometimes $0.
Do you recall Elon's big talk about servicing the cars as fast as in car racing? Then, most services shouldn't require a loaner. In my experience, 90% of the time people just wait 1-2 hours, sip coffee, browse internet/watch TV, then they are out. It should be more true for Tesla services; as they are more sparse, the driving time back and forth is likely much higher than waiting for 1 hour aat the service center.
May be, Tesla service is inherently more complex than service of most other cars. Or, Tesla service is grossly incompetent. Or Tesla cars require a lot more service or more severe service/repairs than other cars. Which one is it, I don't know.
More likely, IMHO, is that Tesla lends out the new cars to keep the excess inventory busy and depreciate them for discounting.
So, I think you folks have it all wrong. The need to lend out cars is because of excess inventory, not the other way around.

Really? You live in Silicon Valley? And you think that 90% of the people here wait 1-2 hours at the dealership to have their cars serviced? Instead of dropping off on the way to work, getting a loaner, and picking up at the end of the day or on the way to work the next morning?

I guess you don't work in high tech if you have that much time to waste.

And you must not have read all the posts on TMC from people bitching that they were forced to take an ICE from Enterprise because Tesla did not have a service loaner available.