We Take A Deep Dive Into Bloomberg’s Tesla Model 3 Tracker


It’s been around for awhile now, but understanding how it works is still a mystery to some

Here we try to flush it out in simple terms so that everyone understands what’s going on.

For Tesla, achieving the production numbers set by its management will be key to the company’s long-term success. When Tesla set out to build the Model 3 – a relatively affordable and stylish electric car – people lined up to buy it. It was the next best thing since honey flavored Cheerios.

Furthermore, Tesla managed to do what no other carmaker in the history of the car world: open the sales with over 400,000 preorders for their entry-level electric car.

Unlike some other car makes, sitting on top of large swaths of unsold inventory, Tesla’s got a completely opposite problem; they can’t produce enough of them, though production is now finally hitting high levels. Ever since the Model 3 was launched in July 2017, future owners were confronted with ever-changing delivery times, production delays and overall, a sluggish output of the Model 3, which holds true to a much lesser degree today.

To make matters worse for Tesla, customers were not the only ones raising concern. The production issues also confounded that of Wall Street and brought some financial qualms to the company as well. In reality, nobody really knew the exact figure that is the weekly output of the Model 3. However, there’s been a pretty good tool that can help us estimate the current production rates of the entry-level machine.

The tool is built by Bloomberg and addresses the problem in a rather sophisticated manner. It estimates the number of Model 3s rolling out of the factory in Fremont, California, by utilizing the Vehicle Identification Number (VIN) database – a set of unique strings of digits displayed on every new car sold in the U.S.

Bloomberg’s estimation tool uses two methods: the first one sends a digital request to the National Highway Traffic Safety Administration website. With this, you get a good picture of how many VINs have been registered for Model 3s at any given time.  The tool tracks the number of VIN’s between each batch registration, using the number of intervening days to estimate a production rate.  Currently, their latest estimates put the overall production at 66,260 Model 3s so far, meaning that Tesla is allegedly building approximately 5,824 of those vehicles per week.

The second method uses a more work-intensive approach. Bloomberg scoured the internet for VINs posted by people on social media—Reddit, Tesla Motors Club and other websites frequented by Tesla customers and fans. Furthermore, a layer of self-reported VINs submitted directly to Bloomberg was also accounted for, combined with the VIN data gathered from social media and a trending graph was created. While not 100% correct (naturally), this does give us a pretty good showcase of the Model 3’s build rate.

The tool gives us a pretty good estimate for the production levels currently, but also, what the future brings. Pushing production numbers forward is key to Tesla.

Thanks to Musk’s far-too aggressive goals, Tesla has repeatedly fallen short of its own manufacturing targets. Hopefully, that will change in the future and we finally get to see that entry-level, $35,000 Model 3.

Source: Bloomberg

Categories: Tesla

Tags: , , ,

Leave a Reply

19 Comments on "We Take A Deep Dive Into Bloomberg’s Tesla Model 3 Tracker"

newest oldest most voted

From article: “…Thanks to Musk’s far-too aggressive goals, Tesla has repeatedly fallen short of its own manufacturing targets. Hopefully, that will change in the future and we finally get to see that entry-level, $35,000 Model 3.”

Musk challenging his team by setting hyper aggressive goals ( aka “far-too aggressive”) has been Musk’s management style from day 1 (both at Tesla & SpaceX) and will hopefully not change anytime soon because it’s thus far served as a not so secret ingredient to Tesla’s achieving the #1 EV market share pososition by a wide margin in North America and Western Europe… and SpaceX being the #1 HO launch provider.

Tesla ultimately always accomplishes Musk’s declared manufacturing targets as demonstrated on Bloomberg’s Model 3 tracker…

and *YES* Tesla delivering on those targets are known to be subject to time-shifting (aka “Elon Time”) which by now should not be a surprise to anyone able to apply reasonable context skills to Musk’s stated goal targets.

Tesla changing to emulate a more traditional management style likely would result in a more traditional result: low and slow.

“Furthermore, a layer of layer of self-reported VINs…” I think there’s an extra “layer of” in there

Fixed. Thank you.

You guys are journalists and should proof what you write more carefully. If you can’t get the grammar right, how can we trust that you don’t get the facts right? Sloppy grammar looks very unprofessional.

The author actually intended to use Prune. However, based on the comments and our understanding of the intention, it made more sense to change it. We do proofread and edit every article. We also have several writers that are writing in English as a second or third language. Some word use is lost in translation. We’ve made very concerted efforts to provide error-free work, which has been pointed out and applauded recently on numerous occasions. InsideEVs is making great strides with all of our new efforts. We’ll try to do even better in the future. Thank you for pointing it out.

Don’t forget to mention the edit/change above the article so you don’t get accused of un-annotated changes. 😉

Constructive feedback is not equal to snarky feedback or as well received. <– see

I’m a fan of Bloomberg’s tracker, although it is interesting that they don’t always agree with Tesla. For instance, the week Tesla said was their first week of 5000 production, Bloomberg listed considerably fewer

This is actually explained on the Bloomberg page… The nature of the data sources makes it impossible to estimate production at individual weeks precision.

Tracking phantom VINs has more error than the Model 3 tracking spreadsheet, which also tracks outstanding orders and order to VIN times. Use that instead.

Honey Nut Cheerios are delicious!

Interesting to see that Bloomberg says Tesla is producing close to 6,000/week now, which was stated as the upper limit of the current production lines after they were put in place. I thought it would take a few more months of producing 4-5k/week before they’d get it ironed out enough to push near 6,000.

If Bloomberg is right, then we should break 20k Model 3s produced in a month before long, which is pretty impressive.

In their latest blog post, they also said they may have underestimated July production, which means they may end up overestimating August to catch up.

Counterpoint’s post is the key to understanding the weekly Bloomberg numbers. When Bloomberg realized they under-counted in July, they didn’t force a correction. The same for when Tesla closed their factory. They didn’t go back and fix the numbers for the week they were closed. Instead what Bloomberg relies upon is the algorithm’s natural self-correction feature. The math itself will correct past errors that are high or low by introducing new errors that will average out over time.

Some may instantly reply with contempt: “That means the weekly numbers will ALWAYS be wrong as they constantly self-correct!!” Bingo. This is exactly what is happening, and it is a feature not a bug. Bloomberg simply doesn’t care whether any single week’s numbers is wrong. All they care about is the running total over time.

It is like trying to use a bathroom scale to try and measure a few grams of crack, and then saying the scale is broken when it doesn’t give accurate results. But when you step on it and weigh your body weight, it is accurate once again and clearly isn’t broken.

That would be fine if Bloomberg would label the graph accordingly, so that journalists and industry watchers wouldn’t tend to put much faith in any individual week’s numbers… as we see happening practically everywhere except what’s written by InsideEVs’ staff.

The explanation on their site is clear about this; if “journalists” fail to comprehend it, that’s not the tracker’s fault…

Pretty sure they will produce over 20K, but the plug-in sales scoreboard will show less because they have started to deliver to Canada

Not in any significant numbers. IF they deliver 20k then plug-in sales scoreboard will show pretty much 20k. It will be interesting to see the first month above 20k though. Next stop, 40k global deliveries. 🙂

Funny thing is, when I peeked at the tracker yesterday (must have been just hours after this article was written) by pure coincidence, it already showed quite a different estimate for the current week… This just confirms that the weekly numbers are not to be trusted.

What’s more interesting though is that the trend suggests they might indeed already be past 5000 average per week. In fact Elon claimed on the earnings call that they hit 5000/week several times in July — the average is lower because of interruptions. So whether it will indeed end up >5000 in August depends mostly on how many interruptions they have. (The peak rate might actually reach 6000 by the end of August according to earlier announcements.)