4 ms·

> that booking a ride costs Uber pennies. Their share of the ride revenues is dollars. How is that not going to be super profitable once new market development

> that booking a ride costs Uber pennies. Their share of the ride revenues is dollars. How is that not going to be super profitable once new market development costs ramp down? Because Uber has no business "barrier to entry" for the ride booking: Other people can write smartphone apps that talk to the cloud or whatever also. If the money involved in the app is dollars from pennies, then others will be highly motivated to write the app. So, the taxi business can return to small, local operators who also use a smartphone app. Then the dollars Uber is getting stay with the driver and/or their local company. Uh, mostly the taxi business is a local business. To me, for a centralized company to hope to make big bucks, or even any bucks at all, in the taxi business is wack-o. For self-driving taxis, okay, for Level 3 automation, but that doesn't help much. What might help would be Level 5 automation, but IMHO that would require essentially full, real AI, and no one knows how to do that. Thus, to me, the idea that the taxi business will be affected by self-driving cars is just nonsense hype -- the media likes to have hype to get clicks.

valuearb 9y ago
The Taxi business used to be a local business. It used to be insular and protected from competition by cozy relationships with regulators. That’s no longer true. Uber is on hundreds of millions of phones. It’s incredibly expensive to get your apps on phones nowadays. And saying other people can write apps that talk to the cloud is a massive oversimplification, building a system that tracks tend of thousands of simultaneous transactions while giving real-time updates to drivers and riders, with 99.xx% uptime, and web/android/ios front ends, is a massively expensive effort. Lastly even if you spend hundreds of millions getting your apps on phones and building real-time infrastructure, no one is going to use your service until you acquire a massive amount of drivers in their market. Uber’s brand, which includes that massive bases of installed apps and active drivers, is a huge competitive advantage. It’s the reason 6x more riders choose Uber over Lyft every day, despite Lyft having app installs, drivers, real time software, and despite Uber itself spending the last year or so shutting on their own brand with their tawdry behavior. Even if you target Uber in a single market or small set of markets to reduce your investment costs for acquiring app installs and drivers, but now your infrastructure costs can’t be amortized over as large a base, making you a high cost competitor.
- graycat 9y ago
  > Lastly even if you spend hundreds of millions getting your apps on phones and building real-time infrastructure, no one is going to use your service until you acquire a massive amount of drivers in their market. I don't get it: The taxi business is older than cars! Before cars we had essentially taxis pulled by horses. In parts of Asia, there were taxis pulled by people, on foot or bicycle. The times I used a taxi, I just called a phone number. Now with smartphones, people can call such a phone number while walking down a street. Answering the phone can be a person and/or computer. For the branding, etc., if you are right about the difficulty of having a computer answer the call from a smartphone, then have some company write a generic app that all taxis in the world can use. The difficulty of getting an app installed? Depends on the app! I'm still having a super tough time seeing any money in a central taxi business, computers or not, smartphones or not, etc. E.g., people have done serious, well funded, polished research on just how to do dial-a-ride for 40+ years.
  - valuearb 9y ago
    Dial a ride? That’s something to research 40 years ago. The new world is app driven, no one need answer a phone call. The Taxi business has always been about two things, dispatch and monopoly rights. The drivers made a pittance while the profits accrued to dispatchers and medallion owners. Ride sharing has (mostly) broken the monopoly, so the profits are going to accrue to the most successful dispatchers. I agree with you that it seems existing Yellow Cab/Green Cab/etc dispatchers could invest to build a similar dispatch system and jump in quickly to compete with the hundreds of thousands of existing cars/drivers they have. But in reality it’s a lot harder than that. They have to be willing to compete on price and service while throwing away their existing business and angering their existing medallion owning stakeholders. It almost never happens that an existing business torches itself to change everything to compete using a new business model.
    - graycat 9y ago
      I'm getting attacked here, but I'll be rational. > Dial a ride? That’s something to research 40 years ago. The new world is app driven, no one need answer a phone call. Look, you've got it in spite of some of what you write: The "dispatching" part is 40 years old. So, that's nothing new. When I looked at dispatching as an optimization problem, it looked so ill-defined that any optimization looked not worth the bother. Or, just assign the next call to the closest taxi or some such. I have a quite good background in optimization and a year or so communicated with a guy who claimed to be high in Uber and working on optimization. I sent him some outlines of some approaches, some questions to clarify the real problem, etc. and never got back anything at all meaningful. My guess was that he knew next to nothing about optimization and had made no real progress on optimization at all. I doubt that Uber's dispatching optimization amounts to much. For answering the phone, for a solo business, let the driver answer his own phone. Or let his wife or kids at home do it. Or have 10 taxi guys form a little informal group and rotate the phone answering one day at a time among themselves or their wives. For a simple server to answer the phone, that can't be very hard to do. Then for the app, the smartphone just uses a standard Web browser, no mobile app required. > The new world is app driven, no one need answer a phone call. Don't need an app; on the client side, a Web browser, including on a smartphone, is enough. Besides, having a human answer the phone stands to be a big, huge advantage. Nearly any taxi customer would much rather talk to the nice voice of the wife or daughter of the taxi driver than some computer. Out where I am, when I called for a taxi, that's what I got, a wife or daughter. Where I live, 70 miles north of Wall Street, the taxi business is just mom and pop operators. There's no way for Uber to make any money here; nothing like Uber is an advantage to anyone or needed or wanted. You have talked about medallions, monopolies, sweet-heart deals with governments, etc., but out where I am my impression is that the situation is fairly simple, that anyone who wants to be a taxi service need only say so or nearly so. On my computer I have lists of phone numbers of local taxi services. That's all I need. I don't have or need a smartphone (a friend gave me a cell phone, but I've never bothered to learn to use it; I just don't need it), but if I did and was out someplace and needed to call a taxi, then I'd have the phone numbers of the local taxi services on my smartphone. When my startup is farther along, I'll do more driving, get a smartphone, carry it in my car just in case, and to use it will just make a phone call to a local taxi service. Here's one: Just set up a Web site with names, addresses, phone numbers, and user rankings of taxi services, for the world. In the US, include GPS coordinates, Zip code, and area code. Link to maps at Google. Start in a medium sized US city where there are lots of small, local taxi services and expand from there. Client side? Sure, just a Web browser. A smartphone user could just paste in his GPS coordinates, and the Web server would give back the 10 closest taxi services. Giving back the 10 closest is common at Web sites now. At least in all but two dozen or so of the largest US cities, anyone who uses a local taxi service has a good shot at being a well-known, favorite customer of a local taxi service, that is, with "the human touch", and get good service -- a good situation for both the taxi and the customer. It would be harder but not impossible for a big company to do some such. Sure, if I suddenly had to travel to some medium sized city in Europe for the first time, maybe I'd be glad to use the brand name of Uber. But so far I've never been in Europe even once! Maybe when my startup is doing well!

valuearb 9y ago

The Taxi business used to be a local business. It used to be insular and protected from competition by cozy relationships with regulators. That’s no longer true. Uber is on hundreds of millions of phones. It’s incredibly expensive to get your apps on phones nowadays. And saying other people can write apps that talk to the cloud is a massive oversimplification, building a system that tracks tend of thousands of simultaneous transactions while giving real-time updates to drivers and riders, with 99.xx% uptime, and web/android/ios front ends, is a massively expensive effort. Lastly even if you spend hundreds of millions getting your apps on phones and building real-time infrastructure, no one is going to use your service until you acquire a massive amount of drivers in their market. Uber’s brand, which includes that massive bases of installed apps and active drivers, is a huge competitive advantage. It’s the reason 6x more riders choose Uber over Lyft every day, despite Lyft having app installs, drivers, real time software, and despite Uber itself spending the last year or so shutting on their own brand with their tawdry behavior. Even if you target Uber in a single market or small set of markets to reduce your investment costs for acquiring app installs and drivers, but now your infrastructure costs can’t be amortized over as large a base, making you a high cost competitor.

graycat 9y ago
> Lastly even if you spend hundreds of millions getting your apps on phones and building real-time infrastructure, no one is going to use your service until you acquire a massive amount of drivers in their market. I don't get it: The taxi business is older than cars! Before cars we had essentially taxis pulled by horses. In parts of Asia, there were taxis pulled by people, on foot or bicycle. The times I used a taxi, I just called a phone number. Now with smartphones, people can call such a phone number while walking down a street. Answering the phone can be a person and/or computer. For the branding, etc., if you are right about the difficulty of having a computer answer the call from a smartphone, then have some company write a generic app that all taxis in the world can use. The difficulty of getting an app installed? Depends on the app! I'm still having a super tough time seeing any money in a central taxi business, computers or not, smartphones or not, etc. E.g., people have done serious, well funded, polished research on just how to do dial-a-ride for 40+ years.
- valuearb 9y ago
  Dial a ride? That’s something to research 40 years ago. The new world is app driven, no one need answer a phone call. The Taxi business has always been about two things, dispatch and monopoly rights. The drivers made a pittance while the profits accrued to dispatchers and medallion owners. Ride sharing has (mostly) broken the monopoly, so the profits are going to accrue to the most successful dispatchers. I agree with you that it seems existing Yellow Cab/Green Cab/etc dispatchers could invest to build a similar dispatch system and jump in quickly to compete with the hundreds of thousands of existing cars/drivers they have. But in reality it’s a lot harder than that. They have to be willing to compete on price and service while throwing away their existing business and angering their existing medallion owning stakeholders. It almost never happens that an existing business torches itself to change everything to compete using a new business model.
  - graycat 9y ago
    I'm getting attacked here, but I'll be rational. > Dial a ride? That’s something to research 40 years ago. The new world is app driven, no one need answer a phone call. Look, you've got it in spite of some of what you write: The "dispatching" part is 40 years old. So, that's nothing new. When I looked at dispatching as an optimization problem, it looked so ill-defined that any optimization looked not worth the bother. Or, just assign the next call to the closest taxi or some such. I have a quite good background in optimization and a year or so communicated with a guy who claimed to be high in Uber and working on optimization. I sent him some outlines of some approaches, some questions to clarify the real problem, etc. and never got back anything at all meaningful. My guess was that he knew next to nothing about optimization and had made no real progress on optimization at all. I doubt that Uber's dispatching optimization amounts to much. For answering the phone, for a solo business, let the driver answer his own phone. Or let his wife or kids at home do it. Or have 10 taxi guys form a little informal group and rotate the phone answering one day at a time among themselves or their wives. For a simple server to answer the phone, that can't be very hard to do. Then for the app, the smartphone just uses a standard Web browser, no mobile app required. > The new world is app driven, no one need answer a phone call. Don't need an app; on the client side, a Web browser, including on a smartphone, is enough. Besides, having a human answer the phone stands to be a big, huge advantage. Nearly any taxi customer would much rather talk to the nice voice of the wife or daughter of the taxi driver than some computer. Out where I am, when I called for a taxi, that's what I got, a wife or daughter. Where I live, 70 miles north of Wall Street, the taxi business is just mom and pop operators. There's no way for Uber to make any money here; nothing like Uber is an advantage to anyone or needed or wanted. You have talked about medallions, monopolies, sweet-heart deals with governments, etc., but out where I am my impression is that the situation is fairly simple, that anyone who wants to be a taxi service need only say so or nearly so. On my computer I have lists of phone numbers of local taxi services. That's all I need. I don't have or need a smartphone (a friend gave me a cell phone, but I've never bothered to learn to use it; I just don't need it), but if I did and was out someplace and needed to call a taxi, then I'd have the phone numbers of the local taxi services on my smartphone. When my startup is farther along, I'll do more driving, get a smartphone, carry it in my car just in case, and to use it will just make a phone call to a local taxi service. Here's one: Just set up a Web site with names, addresses, phone numbers, and user rankings of taxi services, for the world. In the US, include GPS coordinates, Zip code, and area code. Link to maps at Google. Start in a medium sized US city where there are lots of small, local taxi services and expand from there. Client side? Sure, just a Web browser. A smartphone user could just paste in his GPS coordinates, and the Web server would give back the 10 closest taxi services. Giving back the 10 closest is common at Web sites now. At least in all but two dozen or so of the largest US cities, anyone who uses a local taxi service has a good shot at being a well-known, favorite customer of a local taxi service, that is, with "the human touch", and get good service -- a good situation for both the taxi and the customer. It would be harder but not impossible for a big company to do some such. Sure, if I suddenly had to travel to some medium sized city in Europe for the first time, maybe I'd be glad to use the brand name of Uber. But so far I've never been in Europe even once! Maybe when my startup is doing well!