Every time this topic comes up, the discussion always revolves around whether or not Uber is properly licensed, or if it's fair for them to operate. What's often overlooked is the very reason so many of these licenses became as valuable as they did in the first place, and like so many economic issues it boils down to one thing: trust. For example, anyone who's seen more of NYC than just JFK airport and Times Square k…
And that's exactly why I like Get Taxi and Uber: they are most trustworthy. In a usual system, taxi parks have too little power and influence, each of them individually, to make these guarantees, and government has the power, but doesn't really have an incentive to do a really good job. Uber/Lyft/Gett, on the other hand, are big enough to have the power of control, but at the same time they still have competition whi…
Let's imagine that there's a bad driver — nothing horrible enough to be newsworthy, but just bad manners, bad routes, stuff like that. Let's imagine that he operated at an individual taxi park. This means that to provide a feedback to authority, the user would have to spend his time to call the company, to "snitch" about the driver. Would someone do it just to say that "the driver wasn't especially bad, but I would prefer better in the future"? Actually, I think yes, they would, but the probability of this is much slower than a user casually clicking 3 stars icon after the ride. And when you have less data points (customer support calls vs star rating actions), your information is worse has more noise in it. So, your decision making about hiring/firing drivers is worse. This is an argument about size.
And the argument about motivation — I think it's more or less obvious. Let's say your taxi driver has a city-issued beautiful plate in his cab. Do you even know who to call after he loads you off at a wrong street? Do you think they will be working at 11pm on Friday? Do you think they will be responsive and appreciative of the situation? Well, I assume that you interacted with bureaucrats at least once in your life, so the questions are really rhetorical.
These are examples about only one component of the trust issue, customer support and detection of bad performers, but mechanisms like this would work in other situations too.