Hacker News Reader

Comment by partingshots | Hacker News Reader

So... Let me see if I have this straight:

1. Uber is unprofitable and the only way it can become profitable is to get SDC's

2. Uber is significantly (years) behind Waymo in the SDC space.

3. Waymo will launch SDC taxi services first meaning:

- When it puts in an order for SDC components no one else is going to be buying in bulk and thus it can have effectively 100% of capacity of these specialized equipment makers

- It is going to be competing with other taxi/ride share services with all the cost advantages of SDC vehicles while its competitors are paying human drivers (and have basically no fat to cut from their current pricing)

- It will be able to improve its services so when someone else does launch their service will be inferior.

4. Uber expects that its users will stick to it over the course of years in the face of significantly cheaper competition.

5. Uber expects that it is going to be able to continue to use human drivers even while it competes against those same people with its SDC's (i.e. when your employer hires your replacement but expects you to train them).

I simply can't imagine how Uber is worth anything at the moment.

View on HN

> can't imagine how Uber is worth anything

If the operation doesn't make an economic profit, and if there isn't sufficient moat to defend the operation until it can become profitable, then I agree with you it's not worth anything.

But in fact, Uber does have a massive moat -- its network of drivers and riders in thousands of cities around the world. No competitor is even close.

It takes tremendous capital expenditure to build that network, and it will take tremendous capital expenditure to neutralize it.

It remains to be seen if Uber can make an economic profit, but with an operating loss of just 10% of gross revenue, it's not at all inconceivable.

> It takes tremendous capital expenditure to build that network, and it will take tremendous capital expenditure to neutralize it.

Does it, though? Every driver I've talked to basically drives for "all of them". Several told me they try to find whatever apps are local for driving and they setup those alone with Uber and Lyft and simply take whatever pays the most / is more frequent.

Uber's software and network are solid but I don't think there is a whole lot of loyalty there. Anyone who can drive a wedge into a spot in their business could slowly erode it, IMO.

Now, they're too big to die quickly or anything like that. But it doesn't seem impossible at all to me.

> literally drives for "all of them".

But how many have they _stopped_ driving for?

None, they just drive less of A and more of B until they barely have any customer for A anymore until A dies.

it's economically unreasonable to drive for more than one apps simultaneously. Now, it could be economically reasonable to drive for more than one apps at many points in the past.

If Lyft and Uber are so easily exchangeable, why is Lyft is still a minority in the US while spending more money?

There's something more interesting here.

It's not unreasonable, people do it all the time. Look at the window stickers in a rideshare, if they've got one they've got four, all around Seattle.

like i said it's reasonable to drive for all at different points in the past but it's unreasonable to drive for all 4 at the same time. Think about it, incentive-wise, if you complete 50 trips you got $x , if you complete 100 trips you got $2x. If you only got 50 trips in you, why are you splitting them between two apps 25 trips each and got $0 incentive?

Almost every driver I ride with is switching between the Lyft and Uber apps to see which one is offering better rates/destinations. The incentives appear to be insufficient to offset rate disparity in this market - unclear if that holds true everywhere.

I'm not at all well-versed in statistics, but it seems that without enough extra cash to consistently one-up the competition in terms of incentive structures, a ride share company could not make the driver's choice of which app to use a non-random event. So yes, a driver wouldn't necessarily choose all four at once, but they also wouldn't consistently pick one.

It's a well known fact that the biggest factor in taxi revenues is how much paid miles/km you can do per day. The single biggest factor to increase your revenues is increasing that number.

So unless the incentives are massive, a driver will always look for the next ride by any means necessary. It's nearly never worth it to "wait for a better option".

How many paid miles you do a day. Deadheading is not good.

Agreed - Facebook has a network effect/moat because a new service that's better in every way is useless if your friends aren't on it. Getting the network to migrate is hard and when this does happen FB is quick to buy the threat (Instagram, WhatsApp) or compete and kill them (Snap).

Nobody holds an allegiance to Lyft or Uber - they pick whichever is currently a better deal for both driving and riding. Uber's recent loyalty program is the first thing that slightly moves this in favor of continuing to use Uber, but it's only a little incentive. Large swings in price still favor switching.

Their only moat comes from being able to undercut competitors with VC money in any market they appear in until the competitor is dead and Uber can then raise prices to be profitable.

It's probably a winner take all market. A SDC competitor would be a threat, but I'd be both impressed and surprised if Google could pull that off.

FWIW, I have a loyalty to Lyft and will use its service even if Uber is a bit cheaper, but that's because I'm a techie that is aware of all the shady shit Uber has done.

I do this as a rider, but I suspect it's different when you're a driver and your livelihood depends on it.

Uber actually has one moat: their international presence. Other networks don't have it.

If you're a business traveler, you can Uber out of most airports without having to install the local application, and business customers are less price sensitive than local customers in general.

Definitely agree with your point about nobody having an allegiance for local rides, and it's not clear if Uber can defend against hyper local upstarts at scale.

Except it isn't everywhere, so their moat isn't complete and is eroded by the fact that you need to rely on local transport options in a lot of areas - I don't think this really counts much as a strength of theirs, and I think their current attempts to operate in so many markets and their reliance on a lack of enforcement/laws that can change may make this more of a liability for the company.

Quick note that Facebook, unfortunately, has no real competitor for the sheer number of things it does. As soon as one comes out built for non-technical users the way Facebook is - and hopefully is a nonprofit - I’ll go running to it and encourage my friends, too.

I’d want one that’s just an online account with a single picture, name, address, phone number that lets you have friends. No posting - basically just a contact list that’s always up to date because people update themselves.

Keybase could be this.

The main way to get people in would be a focus on event scheduling relying on twilio and sms for including users that don’t have accounts.

Make it cheap but not free ($1/month for people with more than 25 contacts) and advertise no ads or tracking ever.

What would motivate everyone you know to use this service?

Convenient way to do event scheduling - some friends I know are not on FB so it’s a pain currently.

Also having up to date contacts to stay in touch. Phone numbers and addresses change and it’s hard to know when that happens (and currently most people don’t put this on FB).

Could LinkedIn go this way...bridge from its business network to casual social network also?

I doubt it - linked in is a mess. They also have too much tied up with recruiters being their real users.

I’d want something simple.

> non-technical users

what did you mean?

I didn’t write that, but I’d guess easy to use like FB (grandma can use it).

Not setting up your own locally hosted diaspora node.

It's amazing how FB has gotten so easy to use now. As one of the early users I remember FB was more cooler I'm those days because majority found it too difficult to use.

FWIW I deleted fb 3 years ago.

This is indeed what I meant.

> Nobody holds an allegiance to Lyft or Uber - they pick whichever is currently a better deal for both driving and riding.

That's not exactly true: I use Lyft exclusively because have anti-loyalty against Uber because of their greater level of past unscrupulous behavior and scandals.

Yeah - I should strike 'nobody' from my vocabulary in most cases. I meant most people.

Here in Brazil my experience has been the same: drivers use all apps.

I just don't see how Uber can maintain said network on the long run and still be profitable, not something I'd put my money in.

What Uber is doing would generally be described as predatory pricing on other sectors. On sufficiently unregulated markets such is Uber's it's extremely risky, no anti-trust legislation is needed to trump it, just a continuously low entry barrier, including lax regulation. For a more concrete, though only related example (rather than speculation as to Uber's future) see Dow's history.

> Does it, though?

It absolutely takes tremendous capital expenditure to build the network. Uber has an accumulated deficit of $20 billion precisely because they subsidized both drivers and riders to build the two-sided network.

It doesn't have to be winner-take-all. A duopoly in most medium to large cities is perfectly viable. But you won't see dozens of competitors because the drivers and riders would wait too long for a match in a sparse network.

Unless the ride-share commodity can be differentiated some other way, it would take large subsidies for a long time to make much of a dent in Uber.

> It absolutely takes tremendous capital expenditure to build > the network.

I always think some Uber drivers shall copy Uber and run a non-profit version of Uber to compete on pricing. What can prevent this?

Have you seen RideAustin? https://rideaustin.com

Love it. That's how it should be. Drivers should get max possible cut of the fares while the network operator works as a non-profit.

100B pump/dump is a scam. Think about all the indices-tracking-ETFs/Funds that will pick up this overpriced stock(our hard earned 401ks). Sigh. Silicon Valley should not have pumped up what was originally a good idea. This has a high chance of ending badly

What funds are those specifically? Certainly not the S&P 500. The Russell 3000, but they don’t have a choice.

The point of investing is deciding what you believe in (financially speaking)! If you don’t like Uber, make some money on it! You can’t short in an IRA, or sell naked options, but you can sell futures or buy put options. All of these would express a bearish sentiment on Uber. So don’t feel like you sre being screwed, go make some alpha!

For the record I agree with you, and will short Uber the day it IPOs. I’ve already made some good money on shorting Lyft and expect I’ll do well on the same strat with Uber.

Uber is a non-profit version of Uber. They made a 10% loss last year. So a non-profit (assuming they intend to at least break even) would actually need to pay the drivers less.

That’s an amusing proposition but a true non-profit could have more advantages like tax breaks and volunteer staff.

I think the intent of the parent's question was to say why can't a group of drivers get together and create a ride sharing app that is jointly shared by all the drivers and just gives all the money to the drivers, hence taking no profit (rather than the more formal term 'non-profit' which has the tax implications like you say.)

Besides, companies only pay taxes on profits, so loss making companies don't pay any tax either. (And usually have some kind of ability to carry forward losses to reduce future tax bills)

Why would anyone work voluntarily for what would essentially just be a company that gave all it's profits to the drivers but expected them to work for free. Volunteers tend to work for causes they believe in for some ethical or moral reason, and while I sympathize with the drivers like I'm sure many do I doubt many would work for free so someone else can earn more.

> volunteer staff

Why would some people work for free so that drivers can earn more or riders could pay less?

Uber racked up a $20 billion cumulative deficit by subsidizing both drivers and riders.

Even if you cloned the Uber app and had people volunteer to do customer support, you'd still wind up with a service where riders pay more, or drivers earn less, or both.

That works for BlaBlaCar [0] that's true carpooling, without drivers trying to make a living on the platform. Riders pay less and drivers get paid less.

But if Uber is losing $4 billion a year at the current prices, it's hard to see how there would be a sustainable market where drivers make more and riders pay less.

[0] https://en.wikipedia.org/wiki/BlaBlaCar

You don’t pay tax if you don’t make money.

> It absolutely takes tremendous capital expenditure to build the network.

It's like breaking trail: It's a whole lot of work for the person in front, but it's damn easy for everyone else.

Uber, being first mover, had to spend great heaping piles of money convincing people to become rideshare drivers in the first place. Because they had to sell people on the idea of driving folks for a living, and then get them to buy new cars that meet their standards, and then train them on how to be a rideshare driver, and all that fun stuff.

Everyone else just needs to convince a group of people who have already bought the car and learned how to drive it that it's in their interest to hedge their bets. Which, given how Uber has historically treated its drivers, is something that requires approximately zero convincing.

"Every driver I've talked to basically drives for "all of them". Several told me they try to find whatever apps are local for driving and they setup those alone with Uber and Lyft and simply take whatever pays the most / is more frequent."

I heard the same from drivers. It's the ultimate replaceable commodity. Drivers jump ship on an instant and so do I as user. I don't care if it's Lyft, Uber or anybody else.

For the drivers, Uber (and Lyft) tie some hurdle rates (rides/week or hours driven etc) to cash bonuses reward their loyalty. It seems like they have some status levels as well (e.g. Uber Diamond Pro) that I guess probably leads to better economics / bargaining power for the driver. So on margin drivers will specialize in one type of service especially in consistently high demand areas (i.e. cities). Unclear which one is better but on surface it seems Uber has more of these in place than Lyft IMO.

For the riders, Uber Rewards and Uber Cash (that actually forces you to reload if you have below whatever your current ride costs) also make the product quite sticky and make the rider less inclined to take other options on margin.

Small economic tools like these that exploit our psychological biases (e.g. loss aversion, gamification) may go a long way in protecting the moat in this commoditized, competitive environment, and Uber seems to be ahead so far.

(Of course, whichever player comes out with SDCs wins the whole pie)

Is there a valid comparison between Uber and a company like Amazon? Amazon's loss leader strategy paid off and they're doing well now, so can Uber "become profitable whenever they want to" and be fine?

Huge barriers to entry to compete with Amazon (warehouses, tech). Thus no serious direct competitor. Few barriers to entry in ride business (just another app)

Amazon was CF positive and invested it's cash heavily in capital assets (warehouse, etc.). Uber invested (burned) its money on subsidies. Not really a comparable.

Amazon burned less than $100m... it was funded off of positive cashflow... inventory was shipped and paid for by customers, before Amazon had to pay the vendor. It was/is a marvellously capital efficient business.

Grandparent post is assuming self-driving cars, which take one side of the network (the drivers) out of the equation entirely.

> it will take tremendous capital expenditure to neutralize it.

It doesn't take tremendous concentrated capital, though. A bunch of independent regional competitors can fight for each city / country separately. Uber has a leg-up with short-term tourists, who already have their app installed, but for regular users, anyone with a decent app and some ad money can be a competitor.

At least on Android, for ~75% of mobile users globally, Google Maps is more important than Uber for getting from place to place. Every Android user has Google Maps. Virtually every Android user uses Google Maps. The same is not true for Uber.

Uber has a shallow moat around an ugly castle.

and they were genius to acquire waze et al before anyone noticed they were going for the monopoly end game.

I have a strong feeling that Uber will be launching their standalone mapping service soon

> But in fact, Uber does have a massive moat -- its network of drivers and riders in thousands of cities around the world. No competitor is even close.

Their platform is a commodity to both drivers and riders. I haven't met a driver who only drives for Uber. Uber provides no incentives for loyalty to drivers, like actual employment as employees. As a user, I have no loyalty to Uber, either.

Neither the drivers nor the riders are capital items.

> Uber does have a massive moat -- its network of drivers and riders

The network of drivers would be rendered obsolete by SDC

Not necessarily, the capital costs of SDCs means that Uber has time to continue development since their costs for cars is essentially zero.

>But in fact, Uber does have a massive moat -- its network of drivers and riders in thousands of cities around the world. No competitor is even close.

Merely being ahead of competitors isn't a moat, it would have to be something that makes "entering into serious competition" difficult for any new entrants. Uber already lacks the monopoly power to raise prices significantly because of competition from Lyft and local providers.

TNCs do experience network effects, but it's nothing like e.g. operating systems or social networks.

> Merely being ahead of competitors isn't a moat

No competitor can catch up without heavily subsidizing both drivers and riders. If no competitor is willing to spend billions to steal 20% of the market, then that's a pretty effective moat.

In VC parlance, a "moat" means something much more than just needing money to enter the market, but an insurmountable barrier even with a lot of money.

The network effects are the moat, and it would take a lot of money to disrupt that.

Any given market can't sustain more than a 2 or 3 rideshare services because splitting the network makes it too sparse for drivers and riders to match within a few seconds or minutes.

Uber has the #1 or #2 spot in virtually every market. That's an economic moat if I ever saw one.

https://www.investopedia.com/terms/e/economicmoat.asp

A moat has nothing to do with the venture’s current rank, and network effects (as above) are much weaker, what with drivers able to drive for multiple services, the ability to start a service in niche markets (eg wingz and airport rides), and the ability of customers to get a new app and probe multiple wait lists. Austin was able to sustain an extremely fragmented market when Uber left.

I hope your read that link, which concurs that being number 1 by a large margin is not itself a moat. A patent on the key technology is a moat. The unwillingness of everyone to switch is a moat. “They’re number 1” is not a moat.

And the instant they become profitable someone will undercut them. Thanks VCs for subsidizing my rides!

The driver network is worth precisely zero as a competitive moat if (and only if) you are competing against a company with self driving cars. The customer network is a competitive edge but it is also far less sticky than having a robust network of drivers that riders can depend on.

> its network of drivers and riders in thousands of cities around the world

Whose switching costs are virtually zero. smh

What moat? All the driver needs to do is load another app on their phone.

What moat? All the user has to do is change the search engine to Bing from Google.

Yet not many people switch to Bing but so many drivers drive for both Uber and Lyft and riders who easily take whichever is cheaper between Uber and Lyft.

That's right: Google has to constantly be better then Bing or else lose searches. Remember Ask Jeves?

>It takes tremendous capital expenditure to build that network

What I don’t understand with the gig economy is why the gig workers organize and cut out the platforms.

Do the drivers need uber/Lyft? Do renters need Airbnb? Let these companies take on VC build the tech platform, launch, verify the market...then fuck them, leave them holding their own bag while the workers ride off in the sunset.

Besides the Founders and VCs, who wouldn’t love to wake up tomorrow morning and read about all the uber drivers organizing, launching their own platform and getting equity?

People who like low prices might not like it, because the main reason Uber, Lyft, etc are cheap is that they use VC money to subsidize rides.

People vastly underestimate the work it takes to build Lyft/Uber because it's simple to use.

They would need someone to do all the legal/compliance work, validate drivers, handle marketing, handle payments, detect and combat fraud, resolve disputes, and the million other things you don't think about when you use the service.

It would be incredibly cool if each of these pieces could be subscribed to as a service. Especially if there was a way to expose your business surface area, domain knowledge, etc. in a way that it could "seamlessly" plug in.

I think it would make more sense for them to unionize, rather than recreate the entire platform from scratch.l

That is so not true. There is no 'moat' whatsoever.

No Uber drivers are exclusive. In the US I see tons of cars with Uber AND Lyft stickers on. In SEA it was extremely common the same drivers had the Uber AND Grab app installed (before Uber gave up).

> No Uber drivers are exclusive

Bing is just a click away from Google.

Hulu is just a click away from Netflix.

My local grocery store is just one block away from a different grocery store.

Exclusivity isn't required to defend a large network. People are creatures of habit, and it would take a big innovation in the service or massive subsidies to dismantle that network.

Except myself and apparently many others in this thread switch easily between Uber and Lyft, etc, and anecdotally the drivers do too.

I agree, which is why I believe Uber's biggest threat is Waymo. Waymo is uniquely positioned to absolutely demolish Uber almost overnight. They have the Google brand and financial backing. Their network of users is already in place as well (Google Maps).

None of those describe a moat. Google is better than Bing. Hulu has a different offering from Netflix (and a lot of people use both). No idea what different distances of grocery stores have to do with anything.

None of those describe a 'moat'. Something like Facebook is a moat. Instagram. A messaging app. Anything where your participation increases the value of the whole offering and hence you'll be locked in, else you lose a lot of value (e.g. from all the contacts you made in the aforementioned networks)

The cost of driving multiple platforms is pretty low. If the maps in my city are accurate, literally the same drivers are active on both platforms at once...

If there is significant regulation, like airport partnerships, that could be a good moat. Just like the taxis had

>But in fact, Uber does have a massive moat -- its network of drivers and riders in thousands of cities around the world. No competitor is even close.

Ola in India, Didi in China... both are expanding internationally.

If Lyft somehow manages to offer riders consistently cheaper fares (say, 50% off the Uber fare for 1 month) people will switch in droves.

> 50% off the Uber fare for 1 month

That's exactly what it would take.

A tremendous capital expenditure, sustained over time, to dismantle Uber's network of drivers and riders.

Lyft can survive 1 month of that no problem...

25% off would probably do the trick.

it is not that hard to gain foot in any market. just advertise lower rates to customers and higher pay to drivers and within a week u will have more than half of all uber drivers registered on your platform too. 90% of people driving for uber also have lyft account if lyft is available in their area. uber has 0 brand loyalty. it has failed to establish itself as a safe option. its just as shitty as the next option in eyes of most riders.

> But in fact, Uber does have a massive moat -- its network of drivers and riders in thousands of cities around the world. No competitor is even close.

Given Uber's financials, this still reads to me as, "Nobody in the world can burn money anywhere near as fast as we can."

It's not inconceivable that they could get into the black, but a year-over-year adjusted EBIDTA loss that's getting larger instead of smaller, and a revenue as a percentage of gross bookings that's doing the same, says that they've got no better idea than I do of how to make that actually happen.

Let’s say Waymo get to legit, truly market ready SDC taxi tech 3 years before Uber, and they start expanding into Uber’s markets at lower prices. This is bad for Uber, but hardly the end of their business. People won’t immediately trust SDCs, so Uber will lose market share only gradually even if their prices are higher. Plus Waymo have to scale up massively to significantly compete with Uber - that’ll take time.

If Waymo don’t make much of a dent in Uber’s market share in 3 years, Uber gets their SDC tech to market and problem solved. If they DO start making a big dent, Uber can likely raise enough money to simply drop prices to Waymo levels (with human drivers), and just bleed cash until they get the tech out to be profitable. Uber (and their investors) are well versed in this strategy.

But given that Uber is not profitable now competing with Lyft both with human drivers, why would it be profitable tomorrow competing with Waymo both with SDCs?

Even if the profit is no longer negative, it should be near zero since there is not much moat, and it's not even clear whether Uber would be able to compete at all against Waymo (better tech) or even carmakers who own the cars.

The only thing Uber has is brand recognition, but Google Maps actually has even more brand recognition and that's where Waymo is going to be unless antitrust prevents it somehow.

The conventional wisdom is that self-driving cars will be much cheaper than rideshare, and the only question is when they'll reach level-5 autonomy.

But what if they're not actually significantly cheaper?

Let's say an Uber driver who works 60 hours gets paid $1,200 gross, or $20 per hour. Could the self-driving car drive the same trips with $1,200 lower cost?

Not a chance. Because the $1,200 paid to the Uber driver isn't just for the driver's time, it's also for the reduced resale value of the car, and for fuel, and tires, and repairs.

And those costs for the self-driving car will be much higher because the self-driving vehicle will be newer, and larger, and loaded with expensive computers and lidar.

The self-driving car will also require 24/7 realtime monitoring, and customer support, and a depot staffed with technicians, and a field service team.

The IRS rate for deducting business use of a vehicle is $0.58 per mile, not including the driver's time. I'd bet the cost of an autonomous car will be well north of that.

I can see this being a case where the second mover has the advantage. Ie let Waymo struggle to find profitability, then hire away their talent and have at it.

In LA, Uber recently reduced the per minute rate from 80 cents to 60 cents. This was after increasing the rates in Sept 18. Their annoucment of the rate decline was preceded by hours until it was effective.

Prior to this week, Uber had a consecutive ride bonus for 3-rides in a row that would add $7 to 11. But the app would malfunction often or not alert you, and break the streak. Almost everyday I complained but Uber never conceded and never gave me the small bonus.

This week, the consecutive ride bonuses are gone. Replaced by a flat $185 bonus for 70 rides in 7-days to $305 for 120 rides. Previously the biggest ride bonuses would require you to drive 11 hours a day consecutively to hit. And if you did 25% or 50% you’d get 10-15% of the bonus.

My current rate is $23/hour minus roughly $5-6 for gas, depreciation, car and insurance. My rent in LA is $3000/month.

Perhaps $17-18 an hour is fair but Uber’s tactics seem incredibly malicious and ever-changing. I feel like they are exploiting drivers unnecessarily although it is nice to have income whenever I need it.

There’s also tons of bad drivers, they encourage you to use the Uber map system which is seriously flawed, doesn’t calculate traffic and is semi-dangerous. They don’t advise drivers how to do basic tips to navigate traffic.

It could be so much better but it’s weird corporate strategies mixed with bad execution. The negatively Uber generates is a serious problem for their brand and I can’t imagine they’ll exist in 10-years based on how they decide things.

I always felt SDC is the death of Uber. If you had 5 billion to invest in SDCs for a taxi service, why would you become a small investor in Uber, when you could probably build the app for $10 million and just roll your own?

Also uber could become profitable if they charged more. I've actually started taking taxi's again, since they are marginally more expensive than Ubers/Lyfts. However, if you pick them up from an airport, you don't have to wait staring at your phone for 5 minutes watching your driver just sit around in their car until they start moving.

I take Taxis from airports and from hotels for that reason (often faster than waiting for the uber to get through traffic).

The exception is SFO where Taxis cost 2x what an UberX costs.

The Taxi experience is still bad though - just yesterday in DC I had to listen to my driver 'joke' about why the Taliban should be able to fight for their country while also repeatedly pointing out and talking about women on the street.

So you prefer being serviced by a machine that records everything you say, do, think, than bearing an interaction with a living human being which may or may not be quite different and interest, question, amuse or revolt you? I must say i don't agree with that, anytime i would prefer a place where people authentically and actually speak their mind and act spontaneously (not considering whether or not i actually think their behavior is stupid, alienated or impolite), instead of being corporate zombies, their soul already exhausted by the system, silent and jaded, having lost faith in any kind of joy in the productive part of life we call "work".

> Men are disturbed not by things, but by the view which they take of them.

Epictetus

Don't worry, many of the taxi\uber drivers record inside their car, too.

So you prefer being serviced by a machine

Yes

Uber drivers are also recording everything I say and do (not think, but that was hyperbole), so the question is whether I want to ride in such a machine with or without some human trying to interact with me. It depends on whether the human is cool or not. In general, no.

I've had plenty of poor but not shocking Uber/Lyft rides. Its nice you can rate them and get them out of the ecosystem. However, as time goes on, the bar on Uber/Lyft's drivers will lower, the cars will get dirtier and less maintained and the costs will get higher.

The stabilization point is probably close to taxi service. I think that Uber/Lyft really solve for SF's taxi problem mostly.

I think the rating system is key and means the lower bound on drivers will be higher than taxis.

> I always felt SDC is the death of Uber.

But aren't SDCs currently something veeery remote (e.g. even Tesla cannot currently drive autonomously even on technically "simple streets" like highways) and especially only as long as no active offensive social behaviours start to emerge (e.g. gang menbers standing in the middle of the street to stop a SDC to rob/kidnap the occupant and/or destroy the car etc...)?

I agree its very far away, except for interstate driving. I can see an SDC-only lane, with a faster top speed. Residential, taxi-cab driving is so far away. Hey, if I step in front of this car, I don't get to sue a fairly poor driver, I get to sue a mega-corp for a tonne of money cause the algorithm should have stopped.

> Hey, if I step in front of this car, I don't get to sue a fairly poor driver, I get to sue a mega-corp for a tonne of money cause the algorithm should have stopped

Haha, that's an interesting grey zone I didn't think about (I'm swiss and usually non-direct/non-physical damages cannot be claimed here if they cannot be demonstrated/proven/tested). Hat off :)

Tesla is years behind Waymo. The predominant issue facing SDCs currently is inclement weather. Lidar (which Tesla does not use) and cameras have a hard time with heavy rain and snow.

Can't Uber become profitable by raising prices and lowering costs?

They have $40B in gross rideshare bookings and $3B in real losses for 2018.

Let's assume 20% of their costs are fixed and 80% are variable.

What might happen if they raise prices by 10%?

Gross bookings drop by 20% / Revenue per booking increase by 10% => $35.2B in gross bookings

Variable costs drop by 30% (due to lower driver acquisition costs since they're oversupplied) => $24B in variable costs

Fixed costs stay the same => $8.6B

That nets $2.6B profit. My numbers are estimations, but it wouldn't surprise me if they could reach profitability without SDC's.

If we're just making up numbers why not assume you'll only drop gross bookings by 0%, or increase bookings even!

Anything can be made profitable if you just invent numbers.

They aren't pulled out of thin air. They're best estimates. If you disagree, what do you think their unit model would transform into if they acquired lyft for $20B in stock and raised prices by 10%? Why?

you cannot really play with it like that, users have an expectation how much something will cost and if they dont feel that its right you can drop a lot. from my own experience price increase if just on its own of 10% can decease sales by more then 20%. and i assume uber had such ideas and has been simulating them, and I'm sure they are testing it out with smaller but not noticeable increases. let's see how that works for them.

> if you disagree, riddle me something entirely different

The given numbers are not ridiculous, so it's just as fair to assume they can reach profitability as to assume they can't.

Becoming profitable at the cost of growth. I think the market would punish that. The story of growth company is what drives high multiples to sales/profit. If that changes to a mature company analysts will start thinking about measuring by P/E.

Your thesis is supported by their financial statements. Their costs are indicative of a company focused on growth. As revenues continue to grow the can scale back marketing and administrative costs as a percentage of revenue to make profits potential. This company has a lot of potential based on the numbers alone... the real problem is the uncertainty in the market for ride and the tech

I'm not sure why you think 20% of costs are fixed and 80% are variable, or that the elasticity figures are what you suggest they are.

Realistically, Uber has a higher cost basis than a taxi because taxis cut down on insurance and fleet expenses through pooling. As such, Uber is not really able to beat them on price. There's no evidence to support the idea that people are willing to pay more than taxis to use Uber or Lyft. Customers have virtually zero loyalty and so do drivers. If Uber raised prices by 10% and Lyft didn't, wouldn't bookings decrease further? Why do you think it'd by just 20%?

Your napkin math makes sense given your numbers but we don't have any data to back them up.

Re: SDCs, didn't a leaked internal study of theirs a few years back indicate that SDCs wouldn't move their margins by more than 5%?

> I'm not sure why you think 20% of costs are fixed and 80% are variable

This was based off the "Core Platform Financial and Operating Model" chart in their S-1, which states 75% of gross revenue is attributable to driver compensation. So, I imagine 80% is the low end.

> I'm not sure why you think the elasticity figures are what you suggest they are

I'm not sure on the elasticity figures. It's a guess based on running many pricing tests for my own business in the consumer services space.

> If Uber raised prices by 10% and Lyft didn't, wouldn't bookings decrease further?

This could be solved by acquiring them for $20B (or less, now).

> Realistically, Uber has a higher cost basis than a taxi because taxis cut down on insurance and fleet expenses through pooling.

I don't have too much insight into the taxi business, but in San Francisco, Uber Pool is substantially cheaper (<50%) than taxi fares.

Again, I'm not saying my numbers are right -- only that it's plausible they can be profitable without SDC's.

The thing about Driver compensation is that it’s not just their take-home pay, it’s what they have to pay to maintain their vehicle and gas. They’re already barely making minimum-wage when you factor in these costs, at some point you can’t get more blood out of the stone. Point taken re pool though it isn’t something everyone will do.

>Realistically, Uber has a higher cost basis than a taxi because taxis cut down on insurance and fleet expenses through pooling.

But Uber's price / taxis' price is wider than Uber's revenue / Uber's costs. Uber addresses various problems with taxis:

- driver certification: the brokenness of taxi certification arguably created Uber, via unreasonable constraints on the number of taxis in many jurisdictions

- availability: Uber's surge pricing may suck, but it's more fair than random price gouging by drivers at odd times in weird places

- rider pooling: you don't have to pass a hailing rider to combine trips; there's an algorithm for that

- verification: fake taxis used to be a serious problem, but Uber (usually) gives you the license plate # of your driver

Also, with regard to fleet expenses, Uber is arguably taking advantage of decaying capital that would have gone to waste. The cost of a car is often quoted in cents per mile, but that's not quite accurate. A car that sits in a garage will eventually fail; the lifespan of a car is limited by time as well as distance. Many people weren't driving their cars as far as they could in the time that they had -- and in my experience, Uber drivers tend to drive in the machine-gentlest way they possibly can (sometimes annoyingly so -- but why wouldn't they?). The use of own cars potentially yields a significant savings over a dedicated fleet. (Likewise, Uber "pools" the cost of insurance on the taxi with the cost of insurance on the own car.)

you sound like a CEO.

this analysis is... insane. your numbers aren't estimations, they're just nonsense. a state of oversupplied drivers is not at equilibrium, the drivers are going to leave until they're not supplied anymore. oops, now you have to acquire drivers again.

so, so much dunning-kruger

Interesting. Is "sound like a CEO" meant to be an insult? On a site meant for startup founders? Weird.

Do you work in SDC development?

There's a difference between making it work for a small suburb area in Phoenix some of the times and making it work everywhere all the time. And the problem with ride-sharing applications is that unless it works everywhere nobody is going to use it.

Making it work everywhere for all circumstances is hard. Like really hard. There's a difference between an intervention-free ride and an actual ride that's both safe and comfortable. You could rack up your miles per safety intervention by programming a paranoid robot that sits at a highway ramp for hours without merging.

Is it possible? Maybe. It's nice that some deep models can magically (with no theoretical explanation) recognize a pedestrian, a vehicle, a bike. But would you trust it with your life? There's no theoretical safety guarantee unlike aviation where you can actually read and understand the proofs. I mean yes it could certainly work somewhere some of the time, but believing that it will eventually work everywhere all of the time is rather naive.

I wouldn't call it naive to believe it will work eventually, if the time scale for eventually is unspecified; after all, we are an existence proof that it's at least possible. Pinning down that timescale is the tricky part though.

Also, your username is hilariously apt.

> 1. Uber is unprofitable and the only way it can become profitable is to get SDC's

That doesn't seem to be true...

> 4. Uber expects that its users will stick to it over the course of years in the face of significantly cheaper competition.

What significantly cheaper competition?

>What significantly cheaper competition?

The hypothetical Waymo SDC-based 'taxi' service

The assumption that SDCs are going to completely overtake the market is a huge one and is unwarranted. This puts all the cost of building and maintaining a fleet on Waymo, or whoever is first to market. While it cuts out the cost of a human driver, it's not at all clear to me that this is cheaper in net.

Furthermore, a completely robust SDC able to handle all climates and weather in which a human driver is at least nominally competent to drive is still likely a pipe dream. So this model may work well - if it works at all - for the Southern regions of the US, without coverage anywhere else.

While this is true, it's the only proposed path to Uber being profitable. They lose money on every human-driven ride and have no approach to remedy that situation. So if SDCs aren't coming any time soon, or if they aren't any cheaper than human drivers, I'm not sure that's any better for Uber's prospects...

SDC’s are like... minimum 5 years away for something an early adopter would get into. So...

Been hearing that for almost 5 years now:

From 2015: https://www.carscoops.com/2015/01/ford-ceo-says-driverless-c...

Waymo are driving early adopters around Phoenix right now.

No way are we going to be in autonomous vehicles that soon. All of the companies besides Waymo are just doing fancy lane assist and calling it self-driving. And Waymo only works in the most ideal conditions

You're making a very large number of assumptions. In particular #1, #3 and #4. None of these are clear.

You can't imagine any scenario under which an 11 billion dollar money printing machine which is growing and costs 11 billion + 3 billion to run can turn cash flow positive?

When your business is selling $10 for $5 then no, I cannot see a way to turn cash flow positive.

Your example isn't proportionate. It's selling $14 for $11, which is much closer to selling $10 for $8, which is a lot more reasonable than $10 to $5.

Cool, cool. Just gotta raise prices by 25% or cut costs by 20%. Trivial.

They're not doing that. They got $11.27 in revenue for a cost of $9.65. The loss stems from spending an additional $4.67 on advertising and R&D.

1.5 billion on R&D, 3 billion on marketing.

Uber don't get the income from the rides without the marketing so I don't get why you get to ignore that when working out if they are profitable or not.

Does marketing and R&D cost grow linearly with number of rides though?

The “marketing” cost does when it’s driver / rider incentives. It’s literally selling a $10 ride for $8 by chipping in $2 from the marketing budget.

Not true, read the S1. They lose money on every ride due to driver incentives.

Can you be more specific than a 280 page document?

One path is that their network can be leveraged to monetize other services, like Uber Eats or Jump, which could be profitable.

FWIW from conversations with some of the scooter rental app devs, apparently the scooter / ebike on-demand rentals are extremely profitable.

Autonomous cars won’t be proprietary to each manufacturer or company. Once someone has perfected their system, everyone else will drop what they are doing and simply buy it like an OS. It’s all in the software; hardware required (cameras, radar, GPUs) is already cheap and ubiquitous. My money is on Tesla or comma.ai for this.

Most likely the business model for autonomous cars will be a rideshare - instead of selling the software, the company with the best software will buy cars and rent them out for rides via app. Waymo has already gone in this direction.

Also, Tesla Autopilot is a joke compared to what Waymo or even Cruise has.

Why Tesla? They seem to be near the bottom of the pack when it comes to self driving capability.

>They seem to be near the bottom of the pack when it comes to self driving capability.

And yet they are the only company in the world to have shipped a workable level 2 system in mass quantities. Orders of magnitude greater than Cruise or Waymo's fleet size. That puts them miles ahead in terms of data collection, which is the real challenge.

Supposedly adversarial AI networks are solving for the data problem... AI generates the data to train the AI.

My understanding is Tesla and comma.ai have taken a software focused approach with limited hardware. Others like waymo have taken a hardware approach with more sophisticated sensors. What makes you feel Tesla/Comma.ai will win?

Human drivers aren’t equipped with LIDAR.

Because sensors aren’t the hard part. We already have enough hardware to perform superhuman sensing but it’s meaningless without the massive amounts of mapping data and trained models to make sense of it.

http://www.youtube.com/watch?v=OikZ6J1YDlI&t=18m20s

I love how we just assume SDC is just fact. No mention of weather issues, privacy issues, governmental legislation slow down, infrastructure that they drive on.

The human driver still has a minute. It's not like we are talking about Semi Trucks that will primarily have to navigate interstates.

It is going to be competing with other taxi/ride share services with all the cost advantages of SDC vehicles while its competitors are paying human drivers (and have basically no fat to cut from their current pricing)

It's going to be a very long time before self-driving cars are cheaper than cheap cars with human drivers. It's very possible that Uber might reach an acceptable level of self-driving capabilities before Waymo's paid off its investment in its first crop of self-driving cars.

Huh? How much does a full time driver earn? Let's say $30000 for the sake of argument. A car that is run as a taxi or rideshare service full time will last maybe three years. So that eould mean $90000 for the driver plus around $40000 for a decent car. A mass produced self driving taxi car that costs more than $130000 seems very unlikely.

Plus a self driving taxi could run 24/7 while a driver likely only works 40-50 hours a week (at least at the $30k/year price point)

If you want to make up unrealistic numbers, sure, then unrealistically a self-driving car is cheaper.

But with a self-driving car, operational costs like registration, maintenance, fuel/charging, and insurance, would all have to come out of Waymo's pockets instead of the driver's $30,000k. If the car is being driven 24/7, then maintenance costs will be much higher because of the increased wear and tear, and insurance for self-driving cars will be magnitudes higher than it is for human (commercial) drivers. Plus, there's the maintenance costs for the self-driving tech (LIDARs, software systems, etc.) The insurance alone would likely wipe out savings from eliminating human drivers, and the combined maintenance+insurance costs definitely will for the foreseeable future.

The average Uber drives a $30k-40k car (assuming it was purchased new). So you're looking at a $100k difference in costs in acquiring the car (using your $130k price). If the SDCs only last 3 years, that means you only have 3 years to make up the $100k difference, so each SDC would have to make $33k/year more than the comparable human-driven Uber car competitor to break even (since operational costs per car will cost the same or more as paying a human drive would have). $33k/year is roughly 3000+ additional rides/year assuming $10/ride average cost. Assuming 10 rides a day based on media reports, that would require Waymo to at least double the rides each vehicle does every day. And Waymo can't just increase the price on its SDC rides, since if it makes them much more than $10/ride, people will just drive cause it's much cheaper and faster, even including gas and parking.

And remember: in all of the numbers above, Waymo is bringing the costs in-house, except for the vehicles themselves, so there's no markup. Retail costs for similarly specced SDCs cars would be much higher because each respective vendors will want their profits.

So yeah, while I don't believe Uber will achieve self-driving, I think that it's going to be a long time before self-driving cars are cheaper than human-driven cars once all costs are taken into account.

I am not sure if you read my post correctly. Was doubting my parent's point that self driving cars would be too expensive. I was trying to estimate an upper limit below which a self driving car would be cheaper than a car and a driver. So if a self driving car is cheaper than that it should automatically be economical. And I think that a car costing 130k in series is extremely expensive. That gets you a lot of hardware.

anecdotal, so take it for what its worth.

coworker, works his normal shift five days a week. will uber when he gets off work for two to three hours each day and some on weekends. his claim is over six hundred dollars a week plus he claims fifty four cents per mile on his taxes.

he seems seriously happy with the arrangement. he does mention you get into some sketchy parts of town on occasion and will turn off the app after a drop in one and go live when he is in the clear but he has yet to have an issue. he does not expect to drive forever, he is simply piling away a down payment on a home.

At the margin they’ll be cheaper on day one.

> the only way it can become profitable is to get SDC's

This isn’t self-evident. Their services are unit profitable in core markets. And additional revenue streams increase the geographies in which they’re organically profitable.

To keep the math simple: It's worth the amount a greater fool will pay for it.

It's worth it until there is an alternative on the road. Eventually, most if not every company get replaced by a better one. Stock market IPOs are never about predicting what will eventually happen to a company. It is not different than say, Tesla. Everybody keeps repeating that the traditional car manufacturers will catch up with Tesla, but even till then Tesla is gonna do business. And as expected once the novelty of their products (EV and self-driving features) got somewhat out of the way, now it is being judged but the revenue and profit numbers it is pulling.

Uber facilitated 10 billion rides in 2018. I think you are severely underestimating the ease at which another service, benefitting from self-driving capabilities or not, will erode such a massive network. Value of company will be correlated with value of the network, so I can certainly see how it is indeed worth something... and actually quite a lot.

#4 is the most crucial. There's no stickiness. If a competitor builds a better mousetrap, the switching cost is near-zero.

With an accumulated deficit of $7.9 billion in their balance, isn't it worth at least $7.9 billion in tax credits?

Doesn't matter if Uber is actually worth anything - it matters if a lot of people think Uber is worth anything :)

This supposes that Self driving vehicles will be possible in the near future , without a human operator, in every city that Uber operates in. I highly doubt that. At least not in the next 5 years. What is see is Waymo being deployed in select 10 major cities in certain neighborhoods.

Sorry, what’s a SDC?

Self-driving car.

SDC = Self Driving Car

> 1. Uber is unprofitable and the only way it can become profitable is to get SDC's

So I'm not following this closely, but why is Uber unprofitable? I mean the app worked last time I was in the US and surely can't be that expensive to develop and maintain.

Because most of the income from the rider goes directly to the driver.

Worse than that, Uber are taking money from investors and directly giving it to the drivers to subsidise the trip price for the consumer all in an effort to build market share. It's folly to think that customers are going to remain loyal to such a service once the money runs out.

> the only way it can become profitable is to get SDC's

Can't they just get out of markets that are loss leaders and continue to operate in profitable markets? Note that they should but that's always an option if profitability becomes an immediate issue.

it's a very strong brand, as much as it gets a lot of hate here for the last two years of scandals, especially in developing markets noone cares about that. With right change of management and slow but steady rate raise it can be profitable.

Uber needs SDC, alphabet doesn't. Alphabet will not risk a company ending lawsuit due to SDC...uber is more than happy to bet it all on SDC if that is the sole way to become profitable.

Uber vs Waymo will be inetersting for sure.

One company with a lot of capital clears the road in a blue ocean, the others pursuit.

Just like in cycling, fast followers have it a lot easier in getting a % of the market share.

Data. They will sell location data. Mine all transits. Physical location data and daily customer trends will net them billions.

Telcos are already selling that, on more people. Why would Uber inferior dataset would be so valuable...

If Uber is a self driving car company that why do I have to pay for a money losing taxi company, too?

According to the S-1, Uber has a $1bn profit. Where did you see it is unprofitable ?

Only because in Q1 2018 they sold a bunch of their operations to Yandex and Grab for $3B. They effectively lost $2B.

you forgot option B: raise prices, reduce costs

The only issue there is that Uber hasn’t driven out the local livery companies, and Uber’s presence forced many of them to invest in infrastructure. If the rides get too expensive, people will just go back to taxis.

that is exactly what every company in the universe does in the quarter before being offered for sale or filling public offering. always assume the first company filling is the best as possible already.

I concur.

Reply on news.ycombinator.com