Because a non-trivial segment of the auto-buying public doesn’t pay much attention to the news. You (and I and all of us) are completely non-representative of the public - we’re all actively participating in a text-based financial message board, for goodness’ sake. Our level of awareness of Tesla - and recent changes in Tesla’s pricing structure - is going to be vastly higher than that of the average car buyer.
Lowering your price by $1K on 1.8 million vehicles costs about $1.8 billion in revenue. You can buy a lot of advertising for that budget. At some point, it’s not “asinine” to suggest that pulling the advertising lever is a smarter move for Tesla than pulling the price-reduction lever in order to stimulate demand, if such is necessary.
I know many folks who don’t participate on discussion boards but follow the news. Many conversations with those folks start with: “Well Elon is at it again.” And then the Musk/Tesla news is discussed. For them EV vehicle is synonymous with Tesla. Yeah they know that there are some other automakers out there making EVs . But they likely couldn’t name one.
Any person who is the market for an EV likely would check pricing of a Tesla first then see how the competition stacks out against the premier brand.
The price cutting recks havoc upon the competition and builds market share.
Further up thread Brian Gu vice-chair of Xpeng argues the market share is to dominate The world’s car industry will shrink to only 10 companies over the coming decade, a Chinese rival to Elon Musk’s Tesla has said, as intense competition in China’s electric vehicle market spills on to the global stage. Brian Gu, vice-chair of Guangzhou-headquartered Xpeng, said for Chinese companies to be among the last carmakers standing, they would need to have annual sales of at least 3mn vehicles, underpinned by global exports. The world’s largest carmaker Toyota sold 10.5mn cars in 2022, while Tesla sold 1.3mn.
What if lowering the price by $1k on 1.7 million vehicles causes 1.8 million to be sold?
I wouldn’t be surprised if Tesla is doing this roughly in real-time. When they have a huge backlog of orders, they raise prices or keep them the same. As soon as the order backlog starts slipping, they lower prices. If it keeps slipping, they lower the prices some more. Once the order backlog is where they want it, they keep prices stable for a while.
And it is definitely not asinine to suggest that someday, maybe someday soon, Tesla will [want to] start advertising.
But again, you’re talking about people who follow the news. Many people don’t. And not everyone who you could sell an EV to is someone who is “in the market” for an EV. They might not even really consider the possibility of an EV, thinking that “Teslas are expensive” - not realizing that the price has dropped and subsidies are back.
That’s the goal. But lowering the price by $1k on 1.7 million vehicles still costs you about $1.7 billion in foregone revenue. It is reasonable to ask whether an advertising budget of, say, a billion dollars might have also increased sales by 0.1K - for lower expense.
I can only use my wife as an example. She is interested in the Lyriq and has ZERO knowledge about anything Tesla is doing.
I often educate her on EV credits, new EV products etc. A $1k price reduction does NOTHING to make her either interested or even aware of Tesla.
I think the vast majority of those that are shopping for an EV, where Tesla is not already at the top of their list, have a good chance of having no awareness of any price reductions.
What if that’s not the only goal? What if the goal isn’t only to simply maximize sales and minimize expenses. Maybe there are multiple goals?
What if another goal is to maximize the number of EVs out there replacing ICEs while maintaining a reasonable profit level?
What if another goal is to qualify for certain government tax credits that are limited by vehicle price? (As I recall, one of the price reductions earlier this year was specifically for this purpose.)
What if another goal is to force competitors to also lower their prices by $1k to reduce the amount of capital they have coming in to use in their effort to compete?
Of those goals, only the one qualifying for subsidies is really defensible.
Again, advertising can help you sell more cars. So even if your goal is to maximize the number of EVs out there, you need to evaluate whether you get more bang for your buck by cutting your prices or by spending that amount of money (or less!) on advertising. Cutting prices only affects the decision-making process of people who are well-enough informed to know that you’ve cut prices; advertising can reach people who are not engaged at all. And of course, you can do both.
Lowering your prices doesn’t “force” competitors to also lower their prices. I mean, obviously they don’t have to - but it’s not even the price that would encourage them to lower their prices. It’s losing customers. If Tesla cuts their prices by $1K, but my customers at Albaby Motors aren’t really aware of that price cut, then I’m not losing sales - so I’m not going to cut my prices. Similarly, if Tesla doesn’t cut their prices by $1K but instead starts running a billion dollars worth of ads that suddenly pull customers out of my dealership, I might have to cut my prices by $1K to try to keep them.
It’s not about whether Tesla can and should have a strategy that tries to attract customers - it’s about whether cutting prices is necessarily the optimal lever compared to other ways that companies try to attract customers, like advertising. Virtually every company in every industry advertises - even ones with dominant tech and engineering, like Apple. At some point, deliberately eschewing advertising starts to look more like a CEO indulging an idiosyncrasy rather than a smart business plan…
@albaby1 I wasn’t arguing against advertising, I was simply explaining that there may be multiple goals that lead to choosing to reduce price. Reducing price isn’t only used as an alternative to advertising.
Also, for the marginal buyers, price and features have a huge impact. Sure there are plenty of car buyers that are wedded to one make (for example, my parents only buy Subaru, and I mean ONLY since 1985). And sure, there are plenty of car buyers that buy only one type of car (for example, pickup trucks). But the marginal car buyer, and the new-to-the-market car buyers will do some research before buying a car. They may visit a bunch of dealers (and nowadays, there is often a district with 4, 5, or 6 dealerships all in a row on the same street) and check out what they have. Sometimes they will say “hey, I like that red one over there” and then buy it on the spot. Or they may do their research online (this is becoming VERY popular as people have become more comfortable with internet connected devices, and websites/apps exist to do so relatively easily). And they will figure out what they can afford, and which features they want, and then they will choose a color, and buy the car. So when it comes to affordability, $1k could easily make the difference to many buyers.
What does “defensible” mean in this context? Was it “defensible” for Ford to price the model T low such that many people could buy it and flood the roads with them?
Oh, sorry - meaning as an explanation of why price cuts should be adopted to the exclusion of advertising (implicit in regarding the idea of considering advertising as “asinine”). Getting below a subsidy price threshold can only be accomplished by changing your pricing. But putting competitive pressure on other manufacturers and trying to maximize your EV sales can be furthered by either or both cutting prices and by advertising.
Apparently you’ve never heard of shadow mode. Since about 2017, Tesla vehicles run two autopilot/fsd programs simultaneously. One can control the vehicle when requested, the other runs in the background even when not engaged. The shadow program compares its driving decisions with that of the driver. When there is a disagreement it gets recorded and the data is available to Tesla.
Tesla has far more real world data, including edge events than anyone else. The below link describes how the shadow network was used to teach FSD how to better deal with bicycles.
I had a slightly different understanding of how it works. When you do not have FSD/AP engaged it is, effectively, still running, but in shadow mode. When you engage FSD/AP shadow mode goes away because it is actually operating the car. This is also why you can just engage it and it is able to near instantly take over.
But I think you are overstating what shadow mode is capable of. It really only gets one chance (at a time)- to disagree with the driver, because everything is different after that because the FSD did NOT take the action it planned to take. So we don’t know if the NEXT decision would have been good or bad because the driver did something else.
I’m aware of shadow mode…but that doesn’t tell you what the car would do in the spots that humans won’t get into in the first place, but that the car might if it was “unchaperoned.” The humans are always keeping the car from encountering those situations, limiting the scenarios that the car will encounter. Cruise and Waymo don’t have that safety net - they get to see what happens when the car wanders into a situation that a human would never let it get into.
That seems like a minor deficiency. The program is still potentially learning from every disagreement it does record. I’m not sure how your concern is limiting that learning experience.
To my knowledge, robotaxis aren’t allowed to work in rain, fog, or snow. That represents a lot of edge events they are not experiencing.
Here is a common example. When stopping at a red light or stop sign and then making a right turn. When in FSD the car is very tentative and cautious. It often stops, creeps up a bit, stops, creeps up a bit sometimes 3 or 4 times. Even with perfect visibility and no cars or pedestrians anywhere near. So if there is anyone behind me I’ll tap on the accelerator so to not annoy them. Whatever the neural net was doing after the first stop or deciding what to do next is skipped.
The third variant is the same situation but in shadow mode (FSD off). This is just normal driving. Pull up and stop (or slow to ~1 mph) and go. That same 3, or 4 or 5 seconds of neural net processing and creeping is not being done, so we don’t know if it was going to pull out in front of car or was seeing a phantom object or whatever. The only decision that the shadow mode got to validate was to stop before doing a right turn on red.
Got it, thanks. In this example it seems that the program will recognize that it was over-ridden. This can potentially be flagged and looked at by human programmers who can give this issue greater focus, for example through simulations.
Depends on the sophistication of the shadow mode. Will it register your moving before shadow mode decided it was safe as a disagreement event to be flagged?
In any case, I agree with your larger point that there are limitations to this method of learning, just as there are to any method. But the shadow program should still be able to detect problem situations where its response is consistently slower than that of the driver. While it may not be able to learn from these events for the reasons you mention, it should be able to alert Tesla of a problem area that needs to be targeted.
BTW, with respect to an earlier post, it appears that shadow mode runs independent of whatever legacy autopilot/FSD program is available for active use.
“As the labeling network became better, it was deployed in “shadow mode,” which means it is installed in consumer vehicles and run silently without issuing commands to the car. The network’s output is compared to that of the legacy network, the radar, and the driver’s behavior.” Tesla AI chief explains why self-driving cars don’t need lidar - TechTalks
The link is a nice nontechnical summary of how Tesla processes all this driving data.
Revenue declined 23.1 per cent to 120.2 billion yuan quarter on quarter.
“BYD suffered from weak demand for vehicles as consumers were expecting further price cuts after huge discounts on both petrol and electric cars were offered [by various carmakers],” said Gao Shen, an independent analyst in Shanghai.
Some people were upset because they bought prior to the price cuts. And apparently EV sales paused because people have waited to see if further price cuts happens. How long will this behavior last? How much will EV sales be affected?
BYD delivered a total of 508,706 cars in the first three months, down 25.6 per cent from the previous quarter.
It dethroned Tesla as the world’s largest electric car builder last year, bolstered by strong sales of its cars in China, the world’s largest EV market.
Since October, 2022, thousands of drivers drifted towards BYD’s cars that are priced below 200,000 yuan – about 30 per cent cheaper than premium EVs assembled by Tesla and its Chinese rivals such as Nioand Xpeng – as they were unnerved by worries about job prospects and incomes.
