Hacker News Reader

My own impression taking Uber is that same trips would have cost me more than twice if I were to take a regular cab. The reported $470M operating loss on $410 revenue seems to be somewhat in same ballpark. I don't believe current pricing is sustainable. I don't see any cost advantage for Uber other than paying drivers less than the drivers-owners of regular cab service.

IMO, Uber may be hoping that once regular cab services are put out of business, Uber can raise the prices. The sustainable long term value prop seems to be ability to call up a cab using mobile app on your phone (request and dispatch portion of regular cab business). Everything else appears to contribute to loss.

I also don't believe time has come for self-driving cars or will be in next 5 years at least. If another technology bust happens before self-driving cars become popular, these projects will go on the shelf if not cancelled (like the pet projects of founders and executives in 2000 dot bomb).

Besides I don't see why taxi companies can't offer the same value (ordering taxi on a mobile phone app). Here in Moldova we already have that. You can order a taxi from several services at once with a mobile app, and the app is made by the company/companies (I'm not really sure it's only one, but they certainly do business together) behind these services.

Maybe their plan is just to subsidise drivers until self-driving cars arrive, by which time they will have brand dominance and can carry on merrily.

This is always my thought whenever I see a conversation like this come up about Uber. I understand they're all important issues, but why care so much about whether their drivers are employees or contractors when they're intending to lay them all off ASAP? They're just buying time and taking market share until self-driving cars arrive.

Don't remember the link, but Kalanick was recently quoted about how he'd buy a bunch of Tesla self-driving cars in 2020. Whether it will be Teslas or not, I believe it's clear that their whole strategy is to move to self-driving cars as immediately as feasible.

Because that employment aspect is important to those people now. Let Uber do it and what argument do you have against others doing the same? "Soon they'll be robots anyway."

I don't know why people think self-driving cards benefit Uber uniquely. Everyone else would get to use them too, after all, so Lyft's costs also go down.

It's like saying, "whoa, tomato prices have skyrocketed -- I bet grocery stores are making record profits!"

To be sure, there might be some differential advantage, but not enough to justify subsidizing everyone in the hopes they won't jump ship after competition comes in from other brands without these "legacy costs".

Well, to the argument I was trying to set up (whether cities should regularize Uber or fix taxis), one important change is that Uber now has the capital to make self-driving cars happen, and will continue to have the capital to make them continue to improve, and the taxi industry and transportation agencies don't. That is, I'm not drawing a huge distinction between Uber and Lyft, since both are questionably legal, but between Uber/Lyft/etc. and taxis. You could argue that letting Uber run self-driving cars is better than trying to have government-run self-driving cars.

Depends a lot on whether Uber themselves are making this happen, or we're looking to Tesla, Google, etc. If self-driving cars are going to happen anyway, then there's an argument from the other direction that permitting Uber has no long-term benefits.

Uber would have an entrenched brand worldwide and benefits of scale and purchasing power. Would you tackle McDonalds with an international effort to roll out burgers? You have access to the same teenage workers, same buns and potatoes and so on, right?

Ignoring other car services, the public would have the option to buy their own self-driving car, but we'll get to the point where the majority won't. You'll be choosing between:

Own car - larger cost, garaging, cleaning, insurance, power, depreciation, fixed format (sedan, van, truck), etc.

Their car - pay per use, or subscription to use as required, hire the type of vehicle you need when you need it.

It seem highly unlikely that they're planning on waiting 20+ years it will take before fully autonomous vehicles are able to reliably navigate busy city streets and are licensed to do so without a driver behind the wheel.

It'll be less than a decade, people are already talking about replacing trucks with self-driving trucks and there's been lots of research and experimentation with self-driving cars and synchronization. The pieces are slowly falling into place and the key things remaining are the laws need to accommodate autonomous vehicles and larger scale tests need to happen.

When the law catches up, insurance companies will have something in place.

When there's a large scale test, production/distribution have to ramp up.

Even if Uber can replace just 1% of their drivers with self-driving cars it would be worthwhile.

I don't think it will be 20+ years in their initial, primary markets.

What makes Uber a competent owner/operator of a car fleet, compared to someone like Avis?

They solved the "hail a taxi by pressing a button on your smartphone problem". It's not really subsidized, if they weren't setting piles of money on fire trying to grow really fast(and taking moonshots on autonomous cars), they would definitely be profitable.

Having used the apps that literally hail a taxi through a smartphone button, I feel compelled to note that Uber has solved a lot more than that. They've also solved the utterly miserable taxi experience that follows pressing the button.

A few years ago I would have agreed with that, but I recently switched to Flywheel in SF because I was tired of the Uber drivers wandering around downtown's one-way streets for half an hour every time I hailed one of them because they don't know the city. I'm not exactly thrilled about going back to taxis, but at least they don't act like it's their first day behind the wheel. (And with Flywheel they can no longer lie and tell me their credit card machine is "broken"...)

I used Flywheel exactly once. It led to a driver taking an incredible series of wrong turns, wasting a bunch of my time, and me spending nearly double what I needed to. All because the driver took a lot of wrong turns, didn't know the area at all, and didn't rely on technology.

In SF.

Drivers not knowing the area, getting lost, etc are not restricted to Uber/Lyft - I've had many taxi rides in the US where the driver has no idea how to get places - in SF/NYC/LA/BOS/DC -- so not even restricted to 1 city.

In many cities uber is paying drivers a guaranteed $40 per hour to drive regardless of fares during peak times.

In many places uber is currently an unsustainable, heavily subsidised operation which would fail in free market conditions, but because of their big stash of VC $$ they are able to survive while eating away at taxi driver market share.

Yup, in NYC, Uber is paying a guaranteed $35K/6 months revenue to new drivers (provided they're online 200 hours per month and accept 90% of trips, but you know). I don't know how much of this they expect to pay out, but taxis work really well in this city (compared to, say, SF), so my guess is that they're competing aggressively. If they raised prices instead of subsidizing the shortfall to $35K, would the popular response to Uber be different?

It would be profitable, but assuming the situation your describe, it would primarily come down to (1) driver churn and (2) costs associated with any adverse legal rulings/judgments.

1. Cost of recruiting drivers is relatively high - probably around $50-$100 per driver. This cost is composed of background check (probably $10-20), the ad that got them in the door ($5-10 for that click), the overhead cost of recruiting staff and onboarding staff and 1099 paperwork ($2-$5). Let's be nice and say $50 is the cost. Again, being generous, if the average ride is $20, and Uber takes 20% ($4), that means that each driver needs to perform 13 rides for Uber to make their $$ back. I don't know for sure, but anecdotally most drivers I've ridden with likely exceeded this, so I'd say they are good on this front.

2. In reading the recent CA administrative decision, the main damages were for reimbursement for mileage, at $0.56 per mile. So this means that IF a state were to determine either that (a) drivers are employees or (b) drivers are not employees but some "third-option hybrid" for whom Uber still needs to reimburse miles, this would great eat into their margins and drag profitability down.

> Cost of recruiting drivers is relatively high - probably around $50-$100 per driver.

Hrm. Instead of working from first principles, we can poke at some data. Zen99 calculates that Uber seems to be spending some $1M/year on Craigslist ads for SF drivers alone [1], and as of the end of last year, they had about 16,000 drivers in SF [2]. Assuming generously that they recruited them all that year, that's about $62 on Craiglist advertising per driver, much higher than your $5-10 estimate.

There are also incentives like an unconditional $500 on your first ride [3]. In NYC, Uber also offers new drivers a free defensive driving class and doctor's examination at their offices [4], because they need those to get on the road. You will not find doctors to do exams for $2-5.

I would guess total driver acquisition cost, including first-trip incentives, is much more likely to be in the $1000-2000 range, at minimum. And we still haven't gotten to long-term subsidies like uberPOOL, which is what I'm really curious about. Even a $10K driver acquisition cost could maybe be sustainable if each driver lasts over a year on average, since it's a one-time cost. But it needs to pay better than taxi wages + cost of car maintenance for the driver, and at least break even for Uber, in the steady state.

[1] https://tryzen99.com/blog_posts/on-demand-is-in-demand

[2] http://www.washingtonpost.com/blogs/wonkblog/wp/2015/01/22/n...

[3] http://www.driveubernyc.com/try

[4] http://www.driveubernyc.com/defensive

you = right. I was thinking on a per unit basis, but if you think on a conversion basis your $1k-$2k makes a lot of sense, even if you start with my #s - if you assume $50-$100 per driver, and that Uber actually only converts 5% (which is generous), that means you get to the $1k-$2k.

It would be crazy profitable.

But it is subsidizing a lot of rides right now, for example "Pool" rides where riders get the discount even if they are the only passenger.

That was the parent's question. If they froze prices with these subsidies, would it continue?

Yup. Alternatively, if they ended the subsidies and raised prices to match costs, would it go back to being a rich people's service?

Don't get me wrong, taxis in many cities (especially SF) aren't serving the public reasonably. But Uber is only part of that conversation because it's price-competitive with taxis. If we thank them for identifying the problem, and then solve it by regularizing Uber's legal status, will the problem come back in three years once uberX and uberPOOL need to be profitable? Should we instead focus legislative effort on making taxis work?