Hacker News new | comments | show | ask | jobs | submit login
Uber's "Route-Based Pricing" Predicts How Much You're Willing to Pay (bloomberg.com)
62 points by pdog 9 months ago | hide | past | web | favorite | 55 comments

> “Society is more willing to accept wealthy people paying higher fares,” said Chris Knittel, a business professor at the Massachusetts Institute of Technology. “But if the repercussion of lower fares in lower-income places is longer wait times, that’s probably what they want to keep an eye on."

I can't help but recall all the times Uber or its defenders claimed that its supposedly uniform fare scale meant that it could serve all neighborhoods equally, unlike the taxi companies. It will be interesting to see if this policy reproduces the same behavior while laundering it through machine learning.

Interesting, I hadn't heard that defense before. It seems like Uber's app-based hailing is the reason why poor neighborhoods are able to get serviced.

Previously, no taxi driver would drive around in those areas - because of opportunity cost of higher fares in richer areas, and fears of getting robbed when there was tons of cash in the car.

> Previously, no taxi driver would drive around in those areas - because of opportunity cost of higher fares in richer areas

Well, if Uber is using dynamic pricing to ensure fares from richer areas are more expensive than those from poorer ones won't that have the exact same effect?

No, because Uber publishes all of the ride requests, regardless of fare. As long as there is one driver willing to go there, the ride gets fulfilled.

If you are leaving drivers to their own devices to try to find hot spots, they have no way to know someone ten blocks over needs a cab right now from a poorer area.

Gypsy cabs exist for those reasons.

the flip side is that they are redistributing the wealth such that it seems the wealthy passengers subsidize the cost for poorer people.

Its funny. When I first started using Uber, every time I took a ride and paid the fair its was almost a joyous experience to pay what seemed like so little. It made me want to use Uber for all sorts of trips and tell everyone else to do the same.

Now every time I open it, it feels like just a liiittle too much to use for anything but a very occasional trip.

They get more from me for each trip I now almost never take. Win the battle, lose the war.

There should be a name for this kind of algorithm. How bout "Pyrrhic Optimizers"?

They were losing money on you before by subsidizing your ride. Now their price better reflect the actual economic costs, and it's not surprising that this means you don't take a taxi as often.

Doesn't seem like we can know that for sure without knowing their complete costs per ride. Even at the higher price that might still be true. It might not.

What we do know is that the algo isn't based on their cost, its trying to figure out the max it can charge me and "get away with it". I'm suggesting that its time horizon is much too small. It was enough to chase me away. That might be what they wanted. It may not be.

Uber has been profitable in most developed markets* for over a year, this is just meant to squeeze them further so they can subsidize new ones.

We are going to live in a bold new world, where getting picked up at Whole Foods costs X% more than a Grocery Outlet across the street. It's really amazing how loose they are with ethics and how willing to court controversy.

*Uber points to profits in all developed markets https://www.ft.com/content/afa1d4de-33bc-11e6-bda0-04585c31b...

Wouldn't this suggest that they aren't subsidizing rides to compete with Lyft except when Lyft subsidizes rides as well?

I always see people saying that Uber subsidizes rides now and that rates will rise later if they win a market, but the evidence doesn't support those claims. If they just want to gouge people as everyone speculates they would already do so in markets where there is no "competition".

Instead, it looks like even absent a direct competitor, that they still have competition in the form of car ownership. We should be celebrating their ability to keep driving costs down until it's cheaper to use Uber than to own and operate a car.

To me the variability is too much. I can't tell at 1pm whether I a 3:30pm trip to my doctor office will cost me $8 or will there be rain and will it cost me $30.

Since it's not reliable I prefer it much less. I'd rather it just cost $16 both times.

Can't you schedule in advance? Does that not come with a price at the time of scheduling?

We are building smarter and smarter machines that nickel and dime from us at scale. How does this end?

It leads to consumer surplus being further and further converted into supplier surplus. Ultimately, this trend will approach 100% conversion so that there is no more consumer surplus and all surplus is realized by producers.

It is an interesting sort of economic reality to speculate about...but I don't really have enough background to even credibly start.

You're ignoring the existence of competition...

Not really, competition only checks this effect if the price-discrimination advantage isn't so compelling such that:

1) All firms start doing it...because what Producer wants to leave money on the table?

2) The firm that does it best gains a significant advantage over competitors thus leading towards non-competitive market conditions.

And, from a purely speculative perspective, if the price discrimination advantage wasn't strong enough to result in one of the two above situations, then why would a firm invest in this sort of tech anyway?

But, like I said, I am no where near educated enough in the subject matter to really credibly start speculating too far. And it is certainly conceivable that competition could check this trend...but I don't believe it is a sure thing.

Let me try to demystify what we're talking about, and show competition works fine even in this scenario.

Let's say a ride is worth $20 for the consumer and $12 for the producer, for a total surplus of $8. If the price they agree on is $14, the consumer surplus is $6 and the producer surplus is $2.

If the producer can be sure the consumer values the ride at $20, they can charge $19 and take home most of the surplus.

Now lets introduce a competing producer, which is equally competent and can also figure out all this.

By your "what Producer wants to leave money on the table?" theory, they would also charge $19, and maybe get half the market that way.

But obviously they could also charge $18, get all of the market, and make a lot more money. Producer 1 will obviously retaliate and prices will go down.

You're describing a "race to the bottom" those 2 competing producers would participate in. But we all know that when there's a quasi-monopoly with only 2/3 big players, they would "cooperate" to maintain the high prices.

Phone providers are a good example of this: In France for many years there was phone providers with extremely high prices: 50, 60, or even 70 euros per month for unlimited plans. They were all telling how their prices were fair, that they need infrastructure investments, etc etc. Then came a disrupting player: 'Free', who started charging 20 euros per month for the unlimited plans. Immediately, the 3 other players started offering 20-25 euros unlimited plans as well. The infrastructure hasn't suffered from this, 4G came as expected, and Free is now one of the major players.

What would happen in your example, is, unless a 'Free' equivalent shows up, is that they would all charge something around $19.

> But we all know that when there's a quasi-monopoly with only 2/3 big players, they would "cooperate" to maintain the high prices.

> In France for many years there was phone providers with extremely high prices

It's a lot easier to spin up a new taxi app that undercuts Uber/Lyft than it is to build new telecom infrastructure.

Nothing's stopping Newber from showing up with a shiny new app and undercutting both, if consumers are upset enough.

Not a big fan of the "but we all know that..." argument by anecdote.

Economists have studied this stuff extensively for a very long time, and their consensus is that normally markets are quite effective at keeping prices down, as long as they're kept open.

There are certainly exceptions and corner cases where it works differently, but that is what they are.

You're right the telecom has this particularity of not being a "open market" with these big government regulations and bids for frequencies etc.

As the sibling comment points out : "Nothing's stopping Newber from showing up with a shiny new app and undercutting both". As long as this statement stays true (i.e Lyft/Uber don't abuse their dominant position to prevent competitors to enter the market), then the market should be healthy. I'm not convinced they would though.

I really...like really am not the person to speculate credibly in this area...but so long as people want to have a discussion I do enjoy hypothetical speculations :)

But...let's remember first that my intent is not to make claims...I am merely continuing a mental exercise in speculation/critical thinking.

Forgive me if I am over simplifying, but the example you forward is a traditional take on perfect competition.

But I don't think that we should discount the possibility of a lack of competition in this theoretical market. First, Price Discrimination itself requires that the firm doing the discriminating has some degree of monopolistic control/power in the market [1]. The existence of any price discrimination precludes the presence of effective competition in a market.

I think the Uber market is one that already exhibits signs of imperfect competition given the plethora of competitors that have struggled to find a footing:

* https://techcrunch.com/2016/11/08/uber-competitor-karhoo-shu...

* http://austininno.streetwise.co/2016/11/30/scoopme-an-uber-c...

* https://qz.com/583498/uber-competitor-sidecar-is-shutting-do...

Let's further suppose that Uber chooses not to pocket the revenue from this price discrimination immediately and instead re-invests it in the short term to establishing a Network Externality [2].

> Network effects become significant after a certain subscription percentage has been achieved, called critical mass. At the critical mass point, the value obtained from the good or service is greater than or equal to the price paid for the good or service.

If they keep it up long enough they should be able to slowly grind competitors out of the market. And while they might actually succeed in creating a monopoly, they don't have to in order to start rent-seeking [3]. Arguably, they already are doing so by increasing producer surplus with this price discrimination and widening the gap between rider cost and driver pay.

In this dystopian market state, Uber need not fear competitors because they could reduce their rent-seeking percentage to match or beat new entrants. They could also spread costs of undercutting local competitors to their entire user base (including raising prices on users it identifies using this price discrimination technique that would be willing to stomach the increase) which would allow them to potentially operate at margins unsustainable/unattainable for new competition. This would be a classic monopoly [4].

Now, there is nothing intrinsically bad about markets that have Monopolies...in fact there are some examples where it is more efficient to have a single supplier. But that is a different speculative discussion. I am just trying to explain why competition may not be an effective check to this practice.

TL;DR Uber's market likely already suffers from imperfect competition casting into doubt the effectiveness of competition to check this trend. Moreover, a firm might use price discrimination to carve out network externalities in amenable markets.

[1] http://www.economicsonline.co.uk/Business_economics/Price_di...

[2] https://en.wikipedia.org/wiki/Network_effect

[3] https://en.wikipedia.org/wiki/Rent-seeking

[4] https://en.wikipedia.org/wiki/Monopoly

The only mystifying part is your abstract analogy that never seems to happen in reality

Yeah, except... all the time. Sibling comment has an example about telecom in France. Google fiber. Low fare airlines which drove down incumbents' prices. This list is as long as the history of economics.

Competition doesn't always work. It isn't magic. It may even need assistance. But to say it never happens is just... not.

Of course it rarely happens. Let's say I say "go out and make a bus company that captures this market opportunity". How would you even begin to do it?

> 1) All firms start doing it...because what Producer wants to leave money on the table?

That seems to just skirt around the issue. The basic (Econ 101) point of competition is that if all existing firms are making 100 units of profit, and a new firm can make 50 units of profit by offering an equivalent product at a lower price, then that new firm will take business from the existing firms.

Isn't the way they gain advantage by having the lower price?

(so riders with options will not choose the fair with a larger supplier surplus)

The reason they invest in the technology is because they are able to utilize it across enormous numbers of fares.

Cartels and monopolies easily evade competition. And evasion of anti-trust is now just another business expense.

You're ignoring duopolies.

Competition is an empty suit of armor at this point.

It's something we'll try to explain to our grandchildren and they'll try to understand it like we try to understand how people fought duels in the 17th century.

I have trouble seeing how this is supposed to work.

In the short run, my willingness to pay Uber is roughly "whatever Lyft is charging for the same route."

In the long run, if Uber and Lyft somehow collude to increase prices a lot, I buy another car and use both services much less. My short-run behavior won't tell them where this threshold lies, and once it has happened it is too late for them to get my business back for a few years.

Price discrimination works if you have real market power, or if marginal and average costs are hugely different (see: airlines, software). Uber's business doesn't seem to resemble either situation.

I believe Uber's tactic is to be the "last man standing". This seems to be a money pit, and the ones who fund it continuously know about it. It's a game of patience. When taxis will be nullified, and competition will have ran out of money, then and only then they will be the crowned. A Pyrrhic victory if you may.

As for the ethics of Uber, it is like the $1 billion in my bank account. It doesn't exist :)

Even a hypothetical taxi monopoly faces close substitutes: owning a car for locals, renting a car for travelers. I still don't see that much room for price discrimination.

The "Standard oil strategy" of losing money to drive your competitors out of business doesn't really seem to work. If Uber declares victory and jacks up prices, presumably they will have a dozen new competitors the next week. How long did it take substitutes to pop up when Uber and Lyft left Austin?

What an actually evil Uber does at this stage is climb into bed with the regulators, suggesting all kinds of new regulations to kick the ladder down behind them. Uber doesn't seem to be going down that route so far.

It's every company's goal to be the "last man standing". In this case though, the other "men" on the playing field includes car ownership. Even absent Lyft, taxis and other competitors, the big prize isn't winning the current market for livery services. No, the big prize is winning the market for transportation. To win that market, Uber needs to be faster, cheaper and better than Lyft, taxis, public transportation, car rentals and car ownership. If they succeed, consumers win.

Today I had an Uber driver rant to me about a passenger who was charged $110, and he only got $45.

Charging more for trips based on supply and demand makes sense. Why wouldn't you charge more for passengers who are willing to pay more? But it seems completely ridiculous for Uber to treat that money differently and take 100% of it.

Of course every company and industry wants to price services this way, but few are able to. Why is that? There are two different ways that I'm thinking about this.

Monopolistic Power:

The ability to price a product/service based on a customer's willingness to pay and not the product's cost is the key sign of a company's monopolistic power. Most companies and industries do not have the privilege of pricing this way. But a select few can charge customers their maximum willingness to pay, and reap huge rewards. Think Google's ad bidding model.

People who have been saying that Uber, Lyft and other rideshare companies are perfect substitutes might have an issue explaining this monopolistic power. The high-competition model for the rideshare industry would mean rides are priced on cost, not on value.

Allocation of Fixed, Perishable Inventory:

Some industries, for example airlines and hotels, have ~100% fixed, perishable inventory. A hotel room at the Hilton in Chicago has different for every night, and it's impossible to sell a room for a date in the past. It's also very hard for them to grow or shrink their airline seats/hotel rooms for a given period of demand.

Similarly, when an Uber driver is looking for work and an Uber customer is looking for a ride - the Uber driver is in a specific place at a specific time, and the Uber customer's demand is for a ride at that given time from that given place. If the price isn't satisfactory for both parties, the ride doesn't occur, and both lose out.

Even if there were 100 ride companies with intense competition, you could imagine the pricing changing based on perceived value. Airlines have immense competition (and historically limited profits), but still price based on value. Also worth noting, Uber not only can charge customers value-based pricing, but can also estimate the driver's willingness to work for a given price, and pay based on that amount.

The flip side is that they can also predict what's the minimum pay an individual driver is willing to take.

The airline industry has been doing personalized pricing for years - if you don't believe me, redo a flight search in incognito mode. It's interesting how Uber's other transgressions have place a magnifying class over all their other activity.

I've just done a few Google searches for flights (SEA to SFO, LAX to NYC) while logged into two different google accounts, as well as Incognito mode. All the quotes were the same.

I saw the same result on google.ca, and google.co.uk (When adjusted for exchange rates.)

The airline personalized pricing is (or used to be) based on 'purchase city' - they may be using a method to identify it that incognito mode doesn't break

The sector I'm most curious about for this is the rental car industry.

Even more opaque than airlines.

Amazon and other retailers do this too.

Uber doesn't appear to be doing anything here that isn't already par for the course.

So they looked at the value of my house before quoting me? Maybe that is why I was surprised to see them want $50 more than normal a few weeks ago to get me home.

Just to clarify, this isn't personalized pricing based on how much a user is willing to pay (user history), but rather pricing based on the "net-worth" of trip's origin or destination?

Knowing Uber's reputation, I doubt that they would use the trip origin and destination as only input variables. I'm sure they know a lot about the user from their history, credit card, social media accounts, etc.

Why would Uber stop at the route and not charge by estimated user income, their job, age or maybe even the reason why they are traveling?

From what I read, I also understood it to be based on the route (start location and end location).

"Uber is a company filled with over-optimizers, who will continue to futz with prices and hope to find equilibrium."

Ha ha! Yeah. Drivers tell me they have little consistency in their take; there was even a short time period where Uber stopped paying bonuses. (that didn't last long)

Just creates a war of escalation. Someone will release the "Kayak for Uber/Lyft/Etc" if it doesn't exist already.

Let's not forget that the ride is a commodity. So wealthy people -- who may use ride services more frequently -- move to Lyft. That is, unless, drivers turn of the service outside of rich areas and Uber can ensure better supply.

Surely wealthy people would be more concerned with getting a ride faster than cheaper?

Worth pointing out also that Lyft only operates in the US. Uber is in 81 countries.

Once a friend of mine had famously quipped: Pick one profession and any country in world - rules of that business kind of remain same. Uber is a taxi company and they are acting like one.

Should not be the way forward for Uber-like (read other 'push a button') services should be decentralized / p2p (sure i've watched Silicon Valley this week) in a way such that service provider and payer are not screwed like this. (Next Richard Hendricks out there - are you listening to this mate ;)

That would be beneficial for drivers, and riders, but not very beneficial for investors.

It's still possible that taxis are not an investment-friendly industry, just on the fundamentals.

Was that a $3.33 fare the video? The minimum fare in Sydney, Australia is $8AUD ($6USD).

do they have uber pool? pool is usually around 3 or less

Applications are open for YC Summer 2018

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact