Net new orders

[bonaire]

I'd like to start a discussion of the term "Net new orders".

I was doing some math and research into this and the only way I can come up with the math behind the term used in the letter:

• Despite not having the refreshed Model S in stores for the full quarter, and not until June for international markets, Model S orders increased year over year. With the addition of Model X orders, total Q2 net new vehicle orders rose 67% from a year ago.

Taking the fact that no confirmations for Model X occurred in Q2 2015.
And that Tesla does not define the internals making up what a "Net new order" is.
And that the Model S Vin # issues in Q2 of 2016 are just about 7% higher than Q2 of 2015. See graph below of Model S Vin # issues per month.

12315 MS Vins in Q2 2015 - 13184 MS Vins in Q2 2016 = 7%

The fascia refresh cars inclusive as a small subset of Q2 2016.

I then add in about 7100 Model X Confirmations done to customer reservations during Q2 2016. Zero for 2015.

The total becomes (13184 + 7100) / 12315 = 65% increase.

So, is it agreed that a "Net new order" is a reservation that has become a CONFIRMED order which locks in the refundable deposit into a true order?

Terms do change often - and so is an order just a reservation until it is confirmed. Then it is an order "for sure"? And does that make it a "Net order"? And is a Net New Order really just a set of confirmed reservations done per unit time?

Help the analysts know what a Net New order is. I don't think they know even after the Q2 ER call last night.

From that, one must review the vin #s that occur now in Q3 for Model X and Model S production. Presumably that lines up with the Q3 "Net new orders" which is the basis of the H2 sales profile.

I consider a Confirmation done for a Model X on 6/30 to be a "Net new order for Q2" even if Vin # is issued on 7/9/16. And thus, Confirmations for Model X done July 1 onward and Vinned soon after are Q3 Net new orders. These are Vin #s starting in the 145xx range. We are at roughly 169xx as of 8/4/16. Perhaps 2400 "Net new orders" for Model X first 33 days of Q3. Model S confirms of July 1 started with Vin # 1502xx and are now at 1550xx. 4800 there. Total net new orders so far in Q3 are 7200. Annualized ~ 80,000 Or perhaps 40,000 Half-annual. My only glitch in the data is a jump in MX Vin #s on July 3-4 which doesn't make sense. Between 134xxx and 141xx . Could be a clean Vin # used at 14000 for some reason.

 

[schonelucht]

My only glitch in the data is a jump in MX Vin #s on July 3-4 which doesn't make sense. Between 134xxx and 141xx . Could be a clean Vin # used at 14000 for some reason.

I am tracking this progression as well and that's a very solid catch. I hadn't yet seen it, but it is indeed a strange jump. Two explanations : a) a consecutive block of VINs got assigned for building demo/inventory/loaners b) a block of VINs got assigned to the backlog of a certain region, maybe GB/JP/AU cars? Although I feel we should have heard about such a sudden large allocations on TMC. We heard about other large waves of customer VIN assignments too.

 

[dandurston]

My $0.02:

"Net new orders" is simply new orders minus cancelled orders. Reservations are not orders, so they wouldn't figure in here.

Your math seems right. Model S sales up about 7% and then Model X coming in with an order rate about 1/2 that of the Model S.

 

[bonaire]

My $0.02:

"Net new orders" is simply new orders minus cancelled orders. Reservations are not orders, so they wouldn't figure in here.

Your math seems right. Model S sales up about 7% and then Model X coming in with an order rate about 1/2 that of the Model S.

Thanks. Someone going online and doing a 7-day reservation pre-confirmation I guess is not an order but a reservation. There are a lot of terms and since they use the term Reservations when dealing with PowerWalls last year - but the press says "Orders". See this link here: Tesla Energy Has 100,000 Orders For Battery Energy Storage System

But when Tesla claims "Net new orders" - we kind of need an explanation of that.

I really do believe it simply is a "confirmed, contracted reservation that has turned into a non-refundable situation". Sure, customers can walk away from the deposit and not accept delivery. But I believe it goes like this:

Reservation - a name and an optional refundable value (ie. 40K Signature, 5000/2500 Production, 1000 Model 3, $0 Powerwall)
Order - a confirmed order where a reservation is turned into a contracted order
Net New Order - fancy way to proclaim a confirmed order - perhaps means "a reservation that became an order this quarter"

Are there Gross New Orders?
Are there Old Orders?

No where in this set of proclamations is it mentioned named customers. The Q1 "Net New Model S Orders up 45%" more than likely, if not definitely, contain the many old-fascia P90D and P90DL, which we see at stores now in the USA and the fascia re-stocking Vin # configurations that were entered into the system on 3/30/16 just prior to the end of quarter. The new fascia were destined for Europe. That is my claim and I need someone to productively refute that. Those are the 1348xx to 13698x series. Could be up to 2000 pre-configured store-stocking orders done in Q1. I would bet a dozen donuts that was included in the "Net new orders for MS" in Q1. There are blocks of similar vins in Q2 as well, but nothing like the big block that ended Q1.

 

[dandurston]

Gross New Orders would be new orders without subtracting cancelled orders.

Old Orders is orders placed prior to the period being discussed (e.g Q2).

There's no reason to think Tesla is using anything other than the obvious meaning of these terms. Tracking new orders, cancelled orders and net new orders is a worthwhile endeavour. Thinking too hard about what these terms mean isn't.

 

[neroden]

Gross New Orders are new orders for cars in production (probably whether refundable or not, honestly, but I don't really care, it's not important). Some people will cancel their orders; some will walk away from their deposits. These cancellations are subtracted to get Net New Orders.

Old Orders are orders from last quarter (or earlier) where they haven't managed to deliver the car yet.

Old Orders + Net New Orders == Backlog, which is a standard term in certain industries.

 

[dennis]

I'm sure Tesla stated Net New Orders because if they just said New Orders the shorts would be proclaiming that Tesla wasn't taking into account the myriad of orders that had been canceled.

 

[bonaire]

Gross New Orders would be new orders without subtracting cancelled orders.

Old Orders is orders placed prior to the period being discussed (e.g Q2).

There's no reason to think Tesla is using anything other than the obvious meaning of these terms. Tracking new orders, cancelled orders and net new orders is a worthwhile endeavour. Thinking too hard about what these terms mean isn't.

Detailed order tracking is the basis of Tesla's financial health and future. So, that's why some do it.
Firms like Fidelity, Bailley Gifford and TRowePrice should be doing it every day. When you hold 10 Million or more shares, you are dealing with Billions of your customers' money and have to do very careful due dilligence and not just attend quarterly Q&A calls to listen to see if they are saying the right things.
These analysts and some other "analysts" on wall street who appear as experts on CNBC appear to do no level of deep diligence and just repeat what the company says. Ronald Regan once said "Trust, but Verify".

"It's 11 o'clock - do you know where your money is?"
- in other words, are you investing with mutual funds who dive headlong into data or are they just throwing money around in a gambling fashion trusting other analysts who trust hype and circumstance?

Sometimes, I think these fund managers are not able to keep up with all the companies they invest in and simply "buy the market" and hope that the law of averages doesn't hurt them too badly if something goes bad. Example - Vanguard's Health Care fund. Went down .3% yesterday because Bristol Myers Squib went down 16%. - it is 6.5% of the fund's holdings. When a stock becomes homogenized out into mutual funds, it has a lot of surrounding protection among the mass of investments, as Jack Bogle designed when he invented mutual funds in the first place. Stocks become protected like "Bubble Boy" because everyone is buying them and placing them into vast pools of stocks and the continued buying and holding helps isolate shares from day traders - which also helps create short squeezes when worked the right way. Before mutual funds were created, companies like Tesla could never have used the stock market the way it does today. Public companies go on "road shows" to entice mutual fund managers to buy their equity - basically asking them to participate in fund-accumulation of their stock. Until market corrections occur, and liquidation of funds to customers begins, there is little reason for these folks to sell. (queue the final scene in Raiders of the Lost Ark - warehousing any and all things).

Anyway - the whole point is ... "This stuff matters". More so to some than others.

And I think "Net New Orders" simply is "reservations which were confirmed specifically during this quarter - inclusive of Tesla store restocking orders because these cars will eventually be sold at some point". The binding of a non-binding contract.

 

[doctoxics]

I consider a Confirmation done for a Model X on 6/30 to be a "Net new order for Q2" even if Vin # is issued on 7/9/16. And thus, Confirmations for Model X done July 1 onward and Vinned soon after are Q3 Net new orders. These are Vin #s starting in the 145xx range. We are at roughly 169xx as of 8/4/16. Perhaps 2400 "Net new orders" for Model X first 33 days of Q3. Model S confirms of July 1 started with Vin # 1502xx and are now at 1550xx. 4800 there. Total net new orders so far in Q3 are 7200. Annualized ~ 80,000 Or perhaps 40,000 Half-annual. My only glitch in the data is a jump in MX Vin #s on July 3-4 which doesn't make sense. Between 134xxx and 141xx . Could be a clean Vin # used at 14000 for some reason.

Very helpful analysis. However, Tesla's statement that production of S and X at the end of Q2 was nearly 2,000/week is not consistent with the VIN assignments. Tesla also mentioned reaching 2,200/week by the end of Q3 so a slow ramp is expected to continue in Q3. At 2,000/week, we would expect 9,000 S and X by August 3 (4 1/2 weeks X 2,000/week). Why don't the VIN assignments support 9,000 rather than 7,200?

The difference could mean 20,000 or 27,000 S and X produced in Q3.

I agree with the idea expressed in this thread that estimates of production are highly critical to predicting Tesla stock price movements. All efforts to arrive at useful estimates are welcome and encouraged.

 

[neroden]

Here's one possibility: Geographic batching of VIN number assignments. Suppose you're missing data on an entire country and it's getting the higher-number VINs.

 

[bonaire]

Here's one possibility: Geographic batching of VIN number assignments. Suppose you're missing data on an entire country and it's getting the higher-number VINs.

neroden - that hasn't been the case historically. China, Hong Kong, Europe, US, Canada - all come off the same string of Vins and are interspersed.

The graph I have up in the OP is simply the general Model S vin assignment over time. There have been peaks but it is relatively easy to see the rate of increase. The peak most visible is early September, 2015 - which coincides with the volume of cars that ended up being sold to Denmark in Q4.

In the attachment, about 13,000 Vin #s were "given out" (or skipped) versus production. Data is data. It is not emotional. The data below is collected from public postings of Vin #s. It is not official Tesla data. What is official data is the numbers of produced and sold cars. So, why so many extra Vin #s? And where are those inventory cars? Have they over-stated production at times? Scrapped some? Lots of employee cars?

 

[schonelucht]

> In the attachment, about 13,000 Vin #s were "given out" (or skipped) versus production.

Sometimes a car doesn't pass quality control and they rebuild (with a new VIN). It happens, but I believe it is rare. I think the VINs skipped is related to refresh cycles. Seems to be especially prominent in the quarters when they dumped the 85kWh battery, when they went dual drive, when they had the refreshed nosecone. Makes sense that they'd estimate hey we go up to (let's say 134371 with the pre-refresh and start with 135000 for post refresh?)

 

[pz1975]

> In the attachment, about 13,000 Vin #s were "given out" (or skipped) versus production.

Sometimes a car doesn't pass quality control and they rebuild (with a new VIN). It happens, but I believe it is rare. I think the VINs skipped is related to refresh cycles. Seems to be especially prominent in the quarters when they dumped the 85kWh battery, when they went dual drive, when they had the refreshed nosecone. Makes sense that they'd estimate hey we go up to (let's say 134371 with the pre-refresh and start with 135000 for post refresh?)

Maybe the skipped VINs represent the changeover to new AP hardware onboard.

 

[Spidy]

I am tracking this progression as well and that's a very solid catch. I hadn't yet seen it, but it is indeed a strange jump. Two explanations : a) a consecutive block of VINs got assigned for building demo/inventory/loaners b) a block of VINs got assigned to the backlog of a certain region, maybe GB/JP/AU cars? Although I feel we should have heard about such a sudden large allocations on TMC. We heard about other large waves of customer VIN assignments too.

I also thinks it's just Tesla getting some demo/service cars out there. Just to be clear there is one x too much on the one VIN so we are talking about 500 cars. Looking at how many service centres Tesla has it's really just like 35 per SC, then consider showrooms and/or other countries.and it's even less. Well, it will be easy to figure out when we see the inventory cars in a few weeks.

 

[bonaire]

They've had gaps in Vin #s going back to early 2013. So, it's not always aligned with anything. Biggest bucket of Vin bytes (skipped) were the Vins that were assigned to RWD in mid 2014 which presumably got a new Vin when offered to upgrade to a D model. Those would have been "assigned but abandoned". But then what about "actual production" builds that have not sold yet? Those keep growing. Now, the idea that a car built but failing a QA test could still be counted for Produced but not ever Sold. But 5000+ of them? I think more were used by employees. I also believe miles driven under AP by employees can qualify as R&D efforts and expensed as R&D costs.

 

[bonaire]

Can someone explain this Vin # in any way? MS with 168668 (yet they are giving out only 155xxx to customers. This translates to an MS with P90D or P90DL battery, so nothing "special" other than very high Vin #. Built May 2016. Already crashed out of service at 200 mi. Hmm. As they say, it is "off the charts". What games with Vin #s could be going on?

 

[TMSE]

Can someone explain this Vin # in any way? MS with 168668 (yet they are giving out only 155xxx to customers. This translates to an MS with P90D or P90DL battery, so nothing "special" other than very high Vin #. Built May 2016. Already crashed out of service at 200 mi. Hmm. As they say, it is "off the charts". What games with Vin #s could be going on?

No, it isn't gameplay, but a wake up call to you and other VIN counters that your exercise is futile. You can as well stop wasting your time now.

 

[Spidy]

No, it isn't gameplay, but a wake up call to you and other VIN counters that your exercise is futile. You can as well stop wasting your time now.

1 strange VIN on a 200 miles crashed car vs. 300 perfectly logical VINs from cars delivered to customers in the tracker in the same time.

 

[ggr]

I see, the strange, out-of-sequence VIN made it crash!