[lsc]

I'm... not sure you'd need a lot of network effects? I mean, you need a critical mass of drivers and riders, but they can be all in a small area and you can still provide a service. I mean, for the occasional user, sure, brand recognition matters; If I'm going from a strange airport to a hotel, I'm probably going to use the brand i know.

But usually? I switch from uber to lyft on my daily commute based on saving a few bucks almost every day. (right now, Lyft has some scheme where you can buy a 30-pack of $15 off coupons for $300. But they're used every time you ride lyft, so if you want to go somewhere for less than $15, you use Uber)

I mean, 90+% of my rides are areas I ride in a lot, where it would be worth my time to put a fair bit of effort into discovery.

I think that on the driver side, there's a lot of drivers with very strong locational preferences, too; If you are in SF and try to get a ride to San Jose, my experience is that 3 out of 4 uber drivers will kick you out of the car when they find out, so I think you might be able to recruit drivers that way, too.

[djakjxnanjak]

Sounds like Uber and Lyft are both subsidizing your rides - but for a new service to get off the ground would require even more subsidies. There’s a cost to idling and waiting for someone to come along and pay you for a ride. The higher the density of riders and drivers, the lower this cost and the less subsidy required for a given price level.

[icebraining]

You only need to idle if you're using a single app. Many drivers use multiple apps at the same time.

[djakjxnanjak]

True, but I believe this is approach only works well in certain cases. For shared/pooled rides, drivers are automatically assigned a chain of rides. In these circumstances probably few drivers would bother to keep another app open that hardly ever gives them any rides. It would take major subsidies to overcome that. You would need a lot of capital, and I don’t know who’s going to provide that other than investors.

[ak2347]

The network effects are key here because you can't have competitive ETAs without having a lot of supply density and without enough demand you can't build the supply density, so aside from the incentives you are seeing Lyft and Uber using now, there are even more subsidies to build the marketplace and you have to do that city by city and neighborhood by neighborhood in the case of larger markets like NYC. There are also heavy ops costs to get a driver through the funnel and to keep them on the platform. Not saying it couldn't be done but it is not so easy for someone to come in and take marketshare.

[lsc]

sure, city by city and neighborhood by neighborhood... but my point is that you have a viable business, even if you only cover one city or even one neighborhood. Most of my traffic is to work and back.

Assuming that you have cheap/open source software (a big assumption) you could even setup federation. Like you and your crew setup a co-op in one area, and get a lot of riders through personal contacts. I and my friends setup a similar drivers co-op in a nearby area, and similarly use our contacts to build up a very local critical mass.

With federation, you could set it up so that if one of my drivers went to your area, say to drop off a customer, they could optionally be put on your network, either to work there, or maybe just to get customers on your network who wanted to go to my area.

I mean, you still have the huge problems of bad behavior, but I don't think uber/lyft have really addressed those problems very well, either. I think those problems are really hard, and maybe those problems would be easier to solve with a bunch of smaller but federated companies than with two large companies? Maybe not. I don't know.

I'm just saying, I think a lot of the critical mass issues could be pushed off to local leaders.