It's pretty clear at this point that Uber's surge pricing model has been met with mixed reactions and in many cases outright derision by the customer base. The pricing model instituted in some large cities at the end of 2015 allows customers to pay more for the luxury of having an Uber driver arrive in a timely fashion when demand is high.
At first this sounds like a very good idea, Uber simply keys up the price of the fair percentage doled out to the driver until drivers swarm an area where demand is high, this gets the drivers a larger payout per fair but also ensures that the customers in high demand areas also get picked up faster...so what's the problem?
The problem is that surge pricing can't be accurately given a price estimate like non surge pricing calculations are given and often people being picked up in high demand areas are simply focused on one factor, getting picked up ....often under inebriated circumstances, when they sober up after the revelry is done they then notice a larger than usual Uber transaction fee that pisses them off.
The way Uber can fix the problem is to simply allow customers to set a priority on which factor is more important to them and then have the driver selection algorithm use that customer cue to modulate the driver selection so that their desires are not exceeded in real time.
For example, Uber can have the app. specify User convenience metrics in the settings, a first one could be a maximum time to wait (mttw) in all situations...this would allow high demand area calculations to allow users who are willing to wait, to do so and be connected with a more distal car for a cheaper potential final fair cost rather than be connected to a proximal car that is participating in the surge pricing enabled in a given area but pay the unknown higher price.
Another metric that the User should be given control of is simply a maximum price (mp) setting....if the surge fairs exceed that price then the algorithm should bias for drivers outside of the surge area with the trade off of the longer wait time...specifying to the User that their settings are biasing the selection of more speedy options for less costly ones. When all cars exceed the Users set maximum price the app. would simply indicate that no such drivers exist and return to them the lowest fair (which likely is a very distal one and thus likely to take a longer time to arrive).
By engaging this limited form of social oversight as a tool to inform the users of the available landscape of drivers and prices in a seamless way Uber can mitigate against any surprise pricing while also improving the trip experience by allowing communication with the customer as the trip is being facilitated such that a demonstration of customer care (in allowing them to a degree to select for wait time over price) is first and foremost.
This would eliminate the surprise that people have voiced experiencing after getting dropped home by Uber drivers called in an area under surge pricing and then having their card charged an unknown surge price for the trip. I am actually a bit surprised that Uber didn't implement this type of feedback control for Users of the app. (in general) as if it were present it would have fixed the issues of "surge pricing" automatically...since Users would have been given these types of controls as a default aspect of the ride service.
That said, better late than never....it should relatively easy for their engineers to implement the modifications to the application and the driver selection algorithm to accommodate this User bound metric sub feature...mainly because it works as a filter against the Users set mttw and mp values primarily (only when driver count falls to low or zero numbers should the algorithm need to be recalculated to consider drivers outside of the surge area that under the current system are not engaged at all).