Comment by JumpCrisscross | Hacker News Reader

Comment by JumpCrisscross | Hacker News Reader

Parent Full thread

> Given the sheer volume of demand for rides, the relative ease of copying this service, and the number of cars owned by Americans, I don't see how their large revenues can be anything other than temporary. I think the investors are playing the short game.

The long game is in capitalizing on (a) network effects and (b) the effects of changing technology on consumer behavior and market structure.

Start with network effects. I live in New York. There are many transportation options. Like my tendency to default to a "top of my wallet" credit card, I tend to fire up Uber first. Other options are considered if Uber disappoints. It seldom does; Uber remains at the top of my "stack". This leads to me demanding several rides a day more from Uber than from its competitors. That, in turn, makes Uber marginally more attractive to drivers. The technology is easy to copy, but the ecosystem is not: network effects.

Let's now consider the effects of changing technology on consumer behaviour. I can walk out of the aeroport in San Francisco, Paris, Austin or Abu Dhabi and have an Uber fueled and ready. Uber has been the driving factor behind me abandoning my driver's license, and with it expectations of ever owning a car.

Here's another technology: driverless cars. Driverless cars reduce fleet costs and risks, and increase scalability on the supply-side (i.e. scale is no longer limited by driver availability and quality). If I were writing Uber's strategy, everything until the availability of fully automated automobiles would be geared towards establishing a strong opening for driverless cars. The tactical phase deploys in force only after the supply side of the scaling problem has been kicked away.

But neither network effects nor technology will change overall number of people who need rides (demand) and the fact that driving cars is unskilled labor.

If Uber pays more, taxi drivers will make rational decisions and become Uber drivers. Or Lyft drivers. Or whatever other competitor.

If Uber has nicer cars or more efficient service, taxi services will adapt or go out of business. The drivers will go elsewhere, not change industries.

And so on.

>But neither network effects nor technology will change overall number of people who need rides (demand) and the fact that driving cars is unskilled labor.

This may be unrelated but I don't see how technology changes doesn't increase demand. If Uber is convenient/cheap enough that people forgo car ownership and decide to use Uber, doesn't that increase demand. Even if, in the long term, car ownership was cheaper, surely there are people who would rather spend $6 today than sign a loan for $6000.

> But neither network effects nor technology will change overall number of people who need rides

It did with me (granted, N=1). I reduced my own driving to zero and replaced it mostly with Uber. That isn't just a shift in my quantity of rides demanded, it's a shift in my demand curve.

> driving cars is unskilled labor

I would not have given up driving if the alternative were yellow cabs. New York taxi drivers are not knowledgeable about routes. Further, it is not uncommon for a trip to end with a debate over whether the credit card reader works. I will pay extra for even a minimally professional Uber driver over an unprofessional yellow taxi.

The correlation between driver skill and rider experience is one of the problems Uber is running into when it considers scaling. That limitation is one that would be mitigated by driverless cars.

> The drivers will go elsewhere, not find other jobs

Network effects limit the viability of alternatives: there will be less demand on those platforms. As mentioned above, I do not believe most yellow taxi drivers would do well as Uber drivers–the supply is not fungible.

If Uber significantly affects demand, then we might see seismic shifts in the demographics of drivers.

As for the existing drivers, my point is that the effects you're touting are temporary. Taxi companies that provide a lesser service will die off and be replaced by others that follow the example of Uber. The overall economics of the thing will not change unless there is a significant change in demand. I don't think this is going to happen due to Uber any more than it did due to ZipCar.

> Taxi companies that provide a lesser service will die off and be replaced by others that follow the example of Uber.

I don't think Uber in its present form will shift the demand curve dramatically. Driverless cars will. A world with driverless cars could follow two models. Let's call them Apple and IBM.

Apple: each household owns a driverless car. Perhaps it can be occasionally rideshared via UberX, Lyft, etc. In this world, you are right: Uber tomorrow will be Uber today.

IBM: Uber (or Google or Tesla...) owns (and finances and repairs and replenishes) a fleet of driverless cars. Households pay it use fees. In this world, Uber's strategic autonomy is dramatically improved over today. If I were Uber, this is the long game I'd be playing for. I would be (a) training riders to depend on hiring cars with minimal personal contact, (b) encouraging the development of driverless cars, and (c) building a cash pile for rapidly acquiring a vast fleet of driverless cars the moment they become viable.

You are far more optimistic than me about the future of driverless cars. Given N driverless cars, some fraction p will eventually run over and kill small children. Stand back and watch the hysteria and the end of driverless cars.

Also, before driverless cars become widespread, they will need to be first proven and then regulated. Imagine the fantastic force of the driving lobby that will oppose that from happening: taxi drivers, truck drivers, fleet drivers.

Why do trains and subways still have a human sitting in them?

You chose to make your point with probability, which is actually what will most likely make people accept driverless cars in the end.

It's amusing to read the same reasoning after striking away "driverless":

"You are far more optimistic than me about the future of cars. Given N cars, some fraction p will eventually run over and kill small children. Stand back and watch the hysteria and the end of cars."

or with elevators:

"You are far more optimistic than me about the future of automatic elevators. Given N elevators, some fraction p will eventually malfunction and kill small children. Stand back and watch the hysteria and the end of elevators."

If the accident probabilities can be brought to acceptable levels, why wouldn't we adopt them?

PS: With trains and subways we have 1 driver per dozens of passengers, where the savings gained from reducing it to 0 may not outweigh the risk increase/loss of safety feeling.

Good point. I totally agree that if the risk is low, we should adopt the technologies.

I was addressing how I believe these sorts of accidents will portrayed in the news and media, especially if there is an organization with an agenda behind the coverage.

Yes, I think (automatic) car makers will have to be extremely careful. If we have dramatic accidents to soon, it may be hard to sway the public opinion with statistics (as is the current case with nuclear power).

    “If you look at newspapers from American cities in the 1910s
    and ’20s, you’ll find a lot of anger at cars and drivers, 
    really an incredible amount,” says Peter Norton, the author 
    of Fighting Traffic: The Dawn of the Motor Age in the 
    American City. “My impression is that you’d find more 
    caricatures of the Grim Reaper driving a car over innocent 
    children than you would images of Uncle Sam.”

Same thing happened with the horseless carriage when it first showed up. Didn't stop it.

http://www.collectorsweekly.com/articles/murder-machines/

That's the amazing thing about Uber: it does increase demand. By making it easier and simpler to ask for a cab, and actually providing an enjoyable experience in a market where expectations are very low, they managed to attract people who didn't really "need a ride" before.

This does come up in every thread that touches Uber, posted from different accounts. But do you have any idea if this is actually the case? It sounds a bit implausible that measurably more people would suddently get a burning desire to call a cab just because the app is so cool. Numbers, or any facts at all, could certainly help make your case.

Even if the only difference between Uber and taxis was that Uber seamlessly takes credit cards, that's enough to increase demand vs. walking / public transit for the growing population that doesn't carry cash. Add in that an uber is much more likely to show up when called than a taxi, and all of a sudden it seems worthwhile compared to an airport shuttle or wheedling a ride from a friend. Finally, knowing you can get home after a late night without relying on the uncertainty of the taxi system may make you more likely to go out in the first place, again, increasing demand.

Just an anecdote, I sold my car after moving to Seattle because public transit + bike, Uber, Lyft, and Flywheel (yes, normal taxi) + Car2Go + car rentals, was cheaper than owning (including opportunity cost), storing, and paying insurance on my car.

I started using Uber and Lyft a few months after selling my car. Flywheel was a good enough and cheap enough alternative for my needs.

I highly doubt more people are calling cabs based on the technology, but I would definitely believe fewer people are buying cars (or people are selling cars) because alternatives are easy and plentiful.

Anecdotally: yes, it has for me and many people I know.

It's not because the app is cool. It's because it eliminates the biggest points of friction of taking a cab (namely, ensuring that one arrives soonish, and payment).

It would be shocking if that didn't increase demand in the margin. Who knows what the magnitude increase is, though (I believe it is significant).

I can only offer anec-data from the Atlanta area, but I know multiple people (including myself) that either sold a car or removed buy a car from their plans solely due to Uber. I've heard a lot of people state they used to risk a DUI driving home now take Uber instead.

It's nothing to do with the app being "so cool", but being able to get a clean car that actually shows up and being able to pay with a credit card. If you call a taxi in ATL, you expect at least a 15-30 minute wait if they show up at all. 90% of the time the driver will be blaring music or talking on their cell the entire ride. Some times they will refuse to take you to places, or refuse to run the meter but demand $xx for the trip. (Happened multiple times to me, all my friends have similar stories.) If the meter is running you have to carefully watch that they don't drive longer routes or even literally in circles (happened to me). And at the end of the trip, it's guaranteed they will only want cash and to take a CC will likely result in a screaming match and attempts to charge extra fees.

When Uber hit Atlanta, it was awesome. A car that actually shows up in 5 minutes? Clean, quiet, with a friendly driver? Map of your gps tracked trip emailed in receipt and card on file charged? The reasons people avoided cabs were removed. I personally take far more uber rides than I ever did taxis, and I know many friends for who it's the same.

It's not that demand for rides wasn't there, it's just there was no supply of quality rides before Uber. Taxis in ATL do not provide an acceptable experience, and are used only if there are no other options. People who before might have crashed on a couch, drove home after drinking, not attended an event with limited parking, asked a friend for a ride, etc. are using Uber because it's easy and economical. I have no idea how big of a percent this is nationally, but among the city dwellers in Atlanta I know of, Uber has been immensely popular and used for trips people would previously avoid with taxis.

Definitely true for me, I take more cabs in Providence since Uber became available because the local taxis are so inconvenient/bad I never considered it a real option, and I take more cabs in Boston because Uber is a lot cheaper than the yellow cabs.

Not sure about this. I would have taken public transit more often before. At UberX's price point, I ask for a ride more often than before.

I think you're misunderstanding what "network effects" means. A product or service benefits from the network effect when it becomes more useful the more users it has.

For example, a telephone by itself is completely worthless. A telephone when all your friends also have telephones is great.

Uber doesn't benefit from this effect in any meaningful way. If my friends all use Uber, that doesn't make Uber inherently more useful to me than Lyft (or a cab). Its utility is divorced from the number of other people using it.

What you're talking about is more properly called "mindshare" (much as I hate that term) or simply "consumer awareness".

Well, if Uber isn't used then there aren't any cabs when I want one. Lots of people use UBer => I can always get a ride. SO that example might not be a good one.

You need to look at it the other way - it's really a two-headed network effect: there are two classes of users, drivers and riders.

The more drivers they have the more useful it is to riders. There more riders they have the more useful it is to drivers.

Whether this is sufficiently far from the original meaning of "network effect" that we need another name for it, maybe. Given how obvious this effect is, is there an existing name already for it?

This isn't particularly new nor unique to technology. This is the entrenching effect you see in all exchanges - the stock exchange, eBay, etc... any clearing house for sellers and buyers to meet will benefit from this effect.

I don't think "mindshare" is the right word for this.

> Given how obvious this effect is, is there an existing name already for it?

It's a marketplace.

As a seller, and a buyer, you generally benefit from going to a larger marketplace. But that analogy breaks down for ride-sharing because there's nothing preventing drivers from being simultaneously available on multiple ridesharing services, and it's trivially easy to switch between them. So there isn't any true isolation between different services, and it makes more sense to think of them as part of a single, larger ride-sharing marketplace.

>Uber doesn't benefit from this effect in any meaningful way. If my friends all use Uber, that doesn't make Uber inherently more useful to me than Lyft (or a cab). Its utility is totally divorced from the number of other people using it.

True for the general model, but IIRC they're adding (or have added) features that do exploit network effects:

- The fare-splitting feature [1] gets more useful as more of your friends have Uber.

- The carpool service [2] will be more economical the more people you can find on the same route.

[1] http://blog.uber.com/2013/07/15/faresplit/

[2] http://blog.uber.com/uberpool

Those are much better examples of where network effects would come into play.

> Uber doesn't benefit from this effect in any meaningful way. If my friends all use Uber, that doesn't make Uber inherently more useful to me than Lyft (or a cab). Its utility is totally divorced from the number of other people using it.

The larger the rider base, the more drivers the service will have. Higher driver density certainly leads to a better product.

Ebay experienced something similar.

Driverless cars is exactly the play. Once driverless cars arrive, their margins will jump from 20% to something like 50-80% overnight.

I sincerely doubt that. I believe once driverless cars come around, there will be nothing to differentiate Uber from any other app from an established fleet management company like Avis, Hertz, or Enterprise.

If they started to have a 50-80% unit margin, why wouldn't people compete with them? Why wouldn't, say, Amazon compete with them? It doesn't sound like a tough business to get into for Amazon -- just buy a bunch of driverless cars. And Amazon excels at competing at close to 0% margins -- why would we imagine that Uber would be successful at competing against Amazon in this scenario?

Their costs would fall by that magnitude, sure. But so would everyone else's, so that doesn't immediately translate into higher margins.

I think it would benefit Uber more than rent-a-cars and other such services because of their greater effectiveness at finding rides to match with unused (driverless) cars, but that figure overstates it.

Reply on news.ycombinator.com