Only 2% of Tesla Full Self-Driving trial users end up buying it | Hacker News Reader

Only 2% of Tesla Full Self-Driving trial users end up buying it

electrek.co

11 pointsbelter21 comments

I tried it. My Subaru Outback's lane-keeping kit is more reliable. It does less. But what it claims it can do, it does reliably. (Same for Waymo. Does more in a more-limited geography, but does so reliably.) Tesla's FSD would do something bizarre just frequently enough to keep you uncomfortable, but not frequently enough to keep you alert.

Just an anecdote but my friend with a Subaru also says the same thing, it's simple and dependable.

And I live in Western Wyoming, where the thing regularly has to intuit from how other cars are driving lane placement and traction conditions. It just works. Including in near-blizzard conditions. Partly due to having a radar, which Teslas cut to save costs.

All that said, I wouldn't pay $8k or $99 a month for it. Maybe $2k or $20/month for months in which it is used.

With the latest update it really started getting into “this is magic” territory. I tested on street roads and did very well. Didn’t get to try it on a highway though. And yet, it’s not worth $10k or $99 / month. I don’t drive for a living. If they made the subscription significantly cheaper, let’s say $50 or less, I think there would be a lot of uptake. But as it stands it’s just too expensive.

> YipitData accessed credit card data from about 3,500 Tesla owners who participated in the trial and found that only 50 bought or subscribed to FSD after the trial

How does this company get its data? Surely that's not the entire population of trial participants? Or, if it is, why does Tesla give data to this company?

Is this one of those places that buys data from customer budgeting applications or something?

In all likelihood, they're buying that data from the credit card companies. CC companies are very active in selling that stuff.

I'm one of the suckers who bought it for $10k in 2021 and waited around.

It wasn't worth it until the most recent update. It no longer drives like a grandma. It's smooth and useful.

The only real use case of FSD is to show off to your friends, who previously haven't seen the technology. Once you have seen it, the novelty quickly wears off. And beyond that, it's not practical to use for anything else because of all the limitations.

With that said, it is not surprising that people don't convert. Only die hard enthusiasts will convert, the vast majority of people will have worn off the novelty and shown those of their friends and family who are interested and have no further use for it once the free trial ends.

It drives like a terrible driver. However if you have Tesla insurance it can drive like crazy and still not lower the safety score since all it's infractions aren't counted.

> However if you have Tesla insurance it can drive like crazy and still not lower the safety score since all it's infractions aren't counted.

The Tesla insurance thing is interesting. I wonder if consumers have any recourse for Tesla registering false positives for bad driving and increasing insurance rates.

It was fun to use while it was free, but not worth continuing. City driving was terrible and disconnected frequently. Highway driving is pretty nice, but not that much better than the adaptive cruise control that is free or the cruise control on my Subaru.

> or the cruise control on my Subaru

Hold on, is Subaru known for its lane keeping? I sort of used mine's performance as a benchmark for a low-effort system, contrasting it with Tesla's--a seemingly high-effort system--and finding the latter lacking.

Remember that this is an unsolicited trial, so 2% is a very large number.

If somebody signs up for a trial and only 2% of those convert, that's bad. Signing up signals intent and interest.

But nobody signed up for this self-driving trial. It's like if you received a free sample in the mail, and then decided to subscribe. 0.1% would be a good conversion rate for unsolicited "free samples", 2% is really high.

> this is an unsolicited trial, so 2% is a very large number

If valid for the population of Tesla owners, it tanks the company's strategy.

Surely most Tesla owners interested in self-driving already had it, so wouldn't be part of the trial. 2% of owners who weren't previously interested in self-driving sounds like a big number to me.

> most Tesla owners interested in self-driving already had it

Then it isn't the avenue of future growth the company is presently betting it is.

If 2% of people who expressly indicated a preference for not buying self-driving went ahead and bought self-driving after a demo, it indicates to me that the number in the general population is greater than 2%.

Sounds like a big market to me.

> the number in the general population is greater than 2%

This is the 1% of the market fallacy: if we only take 1% of the market we’ll billionaires!

The fraction of the population that might buy this may be significant. But that’s irrelevant. They must first buy a Tesla. For any financially-meaningful horizon, Tesla buyers are the market. If fewer than 2% buy, there isn’t a strategically-viable market.

The market is people buying new cars.

But don't worry, they're pretty confident that a million Robotaxis will be on the road by this time next year. Or maybe 1 year and 3 months or so.

These are much lower than expected.

Reply on news.ycombinator.com