Surge pricing, dynamic pricing, variable pricing, all types of pricing

I work for a company that does dynamic pricing in the live entertainment market.

Varying or dynamic prices is a big subject in urbanist circles.  Here, for example, is Ezra Klein / Vox analogizing two promising pricing ideas in the urbanist world:

https://twitter.com/ezraklein/status/545657680942346240

To the extent that both Uber’s surge pricing and DC’s variable pricing for parking involve pricing, they are similar.  But as these ideas come more to the forefront, I’d love for folks to understand more about the nuances.

Dynamic vs. Variable

The first big difference between the two is that Uber updates their prices reactively, in more or less real time, as conditions in their market change.  That is, Uber waits until it sees more demand for rides than supply of drivers, then ups the price temporarily.  DCPark on the other hand, uses analytics to attempt to anticipate, rather than observe, periods of high demand, and set the price higher accordingly. This is actually more similar to traditional taxi pricing, where late night rides in cities like Austin have a surcharge.  While DCPark users can read a schedule of prices for months in advance, Uber’s prices can change within a couple minutes.

Supply Effect and Demand Effect

Uber’s surge pricing actually changes two prices at once: the price that customers pay Uber and the price that Uber pays its driver. The idea is that, by raising the payout, Uber will lure more people to drive when they’re needed.  Additionally, some of the people wanting rides at the busiest hour will decline to pay the higher price and cancel or postpone their ride. So there will be both fewer rides desired at the higher prices than the lower prices and more rides supplied.  (Incidentally, there is no iron law that says Uber must move these prices together.  Another company may choose to simply pocket the higher prices.)

DCPark, on the other hand, only changes one price.  There is no labor involved in renting a parking space for a car and DCPark can’t create more parking spaces on demand.  These prices exclusively balance demand with supply because fewer people will want to park in those spots at higher prices.

Why doesn’t everybody do this?

Matt Yglesias offers one explanation why price changing isn’t more common: people really hate it and it hurts your brand. While there’s some truth to this, I don’t think this explains much of the barrier to change. After all, taxis didn’t change their prices until Uber comes along, and suddenly they have. Yglesias cites sports tickets selling out as an example of inefficient pricing, but most major professional sports teams use at least variable pricing (different prices for different events) and a great deal use dynamic pricing (changing prices in response to demand).  I think the following challenges are better explanations for why price-changing has seen slow adoption:

Overhead of price-changing infrastructure

The prices for many (most?) goods and services we buy are literally printed. At a concession stand, the cost of a soda may be painted on a menu. Prices for a parking garage may be printed at the entrance. In order to change prices in all these cases, price displays must be converted to digital. Uber was able to avoid this because it was founded in the smartphone era and used people’s own phones as a price display, but DCPark has had to make sure that all parking meters they install are capable of being updated. And fancier, programmable meters will cost more money to develop. Beyond price display, somebody has to change the prices, either by writing an algorithm or making manual changes. Either way, that overhead is only worthwhile if the old prices are leaving a lot of money on the table.

Communicating prices to customers before decision-making

Most of the time, when a company runs a sale (i.e. temporary price drop), they will advertise it. The point of the sale is to get more customers to come in and buy; if nobody knows it’s happening, then you’re just allowing your regular customers to pay less.  Uber, again, has gotten around this because it was founded in the smartphone era.  Because the only way to summon an Uber driver is through a smartphone app, they can guarantee that customers have seen the current price.  For DCPark, on the other hand, drivers usually make a decision to drive to an area (as opposed to taking the bus, cancelling a trip) before they ever see the price for parking.  Raising the price of parking isn’t intended to have an immediate effect of sending drivers home on that trip. Instead, over time people will get used to the idea that prices are higher at certain times and lower at other times. Or, in SFPark’s case, they’ll get used to the fact that they’re higher on certain streets than others.  Maybe in time customers will be trained to check their smartphone for prices before driving or on-street prices will be incorporated into an app that helps drivers choose between on-street and off-street.  But developing these technologies and consumer behaviors will take time and money.

Pricing solutions are real and they’re beneficial in many cases. But they aren’t uniform nor do they apply uniformly.  They also aren’t magic and they will take time to develop.

2 thoughts on “Surge pricing, dynamic pricing, variable pricing, all types of pricing

  1. I think one of the very interesting results that has come out of things like SFPark is just how much parking demand (and therefore parking price) can vary from one block to the next. And maybe that’s what makes those prices more acceptable: you know that you can just go to a couple blocks over and find cheap parking if parking right in front of your destination is too expensive for you. Uber’s surge pricing, on the other hand, is harder to avoid, since it covers a much larger range of space and time. If you could just wait 10 minutes to avoid the “surge”, I’m sure people would be a lot less upset about it.

Comments are closed.