The TMC forum has an excellent Google spreadsheet that summarizes Tesla delivery and schedule information. I keep thinking that there should be a way to condense the data into a VIN delivery timeline, but the scatter in the delivery dates makes it challenging. Now I’ve figured out a way to do it. I took the data and sorted by VIN number. Then I grouped the VIN’s into blocks of 100. Then I calculated the average delivery date for each block. For some data, there was only a delivery range and not a specific date. In that case, I used the middle of the range as the delivery date. As an example, the 2200 block had 17 data points with an average delivery date of 12/25. So we can expect that all 100 in the block had an average delivery date of 12/25. Then we can look at how many blocks had an average delivery date in December and how many in January. And we can estimate deliveries by month.
Results: December deliveries include blocks 1200 thru 2900, or 18 blocks of 100. After subtracting the Post-Signature gaps at 1200-1249 and 2120-2199, that makes about 1670 deliveries in December. Very close to 400/wk. January deliveries come in somewhat lower because of a week off and changeover to accommodate 60 kWh battery packs. Blocks 3000 thru 3800 had average deliveries in January. So 900 January deliveries. As most of this is loaded late in the month, it looks very much like they are back up to their 400/wk rate. Looking ahead, blocks 3900 and 4000 are both showing a delivery average date of Feb 5.
Interesting approach. Wouldn't it be more revealing to average the VINs around each date (using every VIN from say 20 days prior and 20 days after each day) and then assume that the average VIN shipped on that day.
Then you could calculate assumed daily deliveries based on the increase in VIN deliverd on any day.
Plotting these assumed daily deliveries would start to tell you how accurate the numbers are or aren't. Example if the numbers are generally positive then you may be onto to something. If the plot shows negative deliveries (where average VIN is less than from previosy dates) then you know the numbers are sketchy.
You shoudl also see a dip in assumed deliveries due to Chrismas shutdown.
PS: This is all a side show. Elon claimed the break even point was just 8,000 annual sales. Everyone mroe or less agrees they shipped 2,000 cars in Q4 ... so they should be profitable but they aren't.
My prediction: No one will ask him why.
KB, Initially I tried averaging as you suggested, but the data was so chunky (feast and famine in the VIN reporting) that I needed to average a large group of points to have any confidence. Then it was hard to see short term trends. By breaking it into "hundred blocks", I feel more confident about the build rate and the monthly totals, and I can see short term trends more easily.
The Dec and Jan delivery numbers should be correct within 100 cars. That puts my Q4 estimate between 2500 and 2600. I'm not sure how sales get reported officially, as many customers made their final payments weeks before delivery.
K_V, thank you for the encouraging words. Your discussion on this board has been valuable.
Comstock, it's very easy to extrapolate delivery rate into earnings and revenue. Revenue is a direct calculation based on average sales price. For earnings, you also need to estimate gross margin and estimate operating expenses. People who have crunched the numbers (including investment analysts) feel that they can just about break even at this build rate. As for the price of TSLA, there's still upside room as long as investors don't demand profit. So far, that seems to be the case.
Many thanks for making this data useful to everyone here! This will make pre-release disussions much more reasonable. Perhaps I should say it could make us all less snippy and more thoughtful.