AI and Price Fixing

tried and true hustle of elimonade stand. So you got your lemons, you got your sugar, you got your water, you got your ice, you got your paper cups, your sunk around. I don't know ten bucks in your ingredients and supplies all told, and your goal is to make more money than you spent putting it all together. Well, you got some decisions you need to make. Where are you going to locate your lemonade stand? How are you going to get the attention of people going by? And

perhaps most importantly, how much are you going to charge? Now? If you charge too much, no one's going to buy a cup of your lemonade. If you charge too little, you'll end up making less than what you spent on materials. Even if you sell out. So finding the balance between what the market will bear and what you need in order to cover your costs is one of those fundamental

journeys every business takes, whether it's products or services. Now, an agile business operator might adjust things on the fly. Maybe there are some external factors that are playing an effect. Right, Maybe the day is extra hot, you know, maybe there's a block party or a festival nearby, so there's more foot traffic than normal. That would be a good place

to put your lemonade stand. These are the things that are going to affect demand, and they could mean that you could price your eliminade a little higher than you would if it were on a quiet and overcast day. Heck, maybe you realize early on that you price things too high, so you quietly lower the cost of a couple of lemonade and sales pick up. As a result, you have

engaged in what is called dynamic pricing. Now, from a high level, this sort of pricing happens all the time, though it's not always shifting moment to moment where we can see it in action. So you might find that a product in the store costs a little more or a little less since the last time you went shopping at that store. But in your typical store, you're not likely to see the price tag change right before your eyes. But that is changing, and it's largely thanks to AI.

AI powered dynamic pricing is intended to maximize profits for business owners. That's really all there is to it. When you get down to the bottom line. It's meant to make people more money. And by people I mean business owners customers. It's meant to take more of their money. So maximization can get pretty crazy. An aipower dynamic pricing tool could in theory adjust prices on products not just based on local demand, but also as a result of

how the perception of the product changes. Here, I'm going to give you a totally hypothetical example. Susie the TikTok star is brainstorming ideas for content and she decides to create a comedic sassy sketch and it centers around a particular brand of butter. So she shoots her video, she edits it, she uploads it, and the engagement comes pouring in. Now among the countless teeny bopper fans watching her content,

there are also AI bots. One of those identifies the brand of butter in the video and takes note of the engagement that the video is getting. It's saying, huh, lots of people are watching this video. And it does some quick checking on available stats to figure out how popular Susie the TikTok Star is in say a specific

local region. It figures out there's actually a huge fan base in this area, and it concludes that at least some of those fans are likely to rush out and try to recreate Susie the TikTok Stars video and to do that, they're going to want the genuine article. They're going to want the same brand of butter that Susie used.

And so this AI, which happens to be a price strategy AI for a local grocery store chain, makes the call to bump the price on that particular brand of butter because there's about to be a surge in demand. The AI is not waiting to detect the demand is surging and then respond to it. It's actually being proactive. It is predicting the increase in demand and acting on it now instead of later. Now, we've seen dynamic surge pricing for years already, though not necessarily in the grocery market.

Ride haling companies have been making use of surge pricing to respond to demand for years since they launched. You've probably been in a situation in which the cost of hailing, you know, an uber or lyft or whatever is way higher than it normally would be for that part of

town in that time of day. This often happens when there's a big event going on, like a huge music concert or sporting event or conference or whatever, and during that window of time you may be paying twice as much or more for a ride as you would any other day. Now, there are a couple of reasons for this.

One is that demand is exceeding supply. There are more people who are looking for a ride than you have drivers available to give people a ride, So raising prices can discourage folks who don't need a ride as badly as the ones who do. So the folks who really need a ride, presuming they can still afford to pay for it, will stick it out, and yeah, they'll have to pay more than usual, but the point is they'll get a car and they'll be able to get to

where they need to go. The folks who say WHOA, Okay, well, y'all, maybe we shouldn't go to shake Shack. Maybe we should just find a place to eat around here. They'll back off and that helps improve availability. It decreases demand and allows the supply to get to where it's needed. But another reason is that surge pricing alerts drivers to areas of high demand. So cities can be pretty darn big. If you don't live in a city, you might not

be used to it. But if there's a large demand that arises in one particular part of the city, drivers will see it on their apps and they can actually choose to reroute toward that part of town in order to take advantage of the surge. So that means that you get more drivers coming into the area, and that means the supply for drivers increases. This in turn affects demands and eventually the surge subsides. So this is one

way to dynamically respond to demand spikes. It's giving an incentive to drivers to say, yes, there are probably people around my area right now who are looking for a ride, but way more people are looking for a ride across town, and I will benefit from the fact that surge pricing is in effect, So if I can get there in time to pick up a ride, I can end up getting more money for that ride, so it actually makes sense for me to leave the area I'm in right

now and head across town now. Of course, the other reason that surge pricing exists is that it creates opportunities for the company to make that cheddar right because the companies that are employing oh wait, I'm sorry, I'm sorry, the companies that are contracting with the drivers, they certainly want their cut of the pie as well. So surge pricing means making more money from the same trips as

you would any other day. Though this doesn't last forever because once supply and demand will reach parity, then surge pricing will decrease. Otherwise you run the risk of alienating your customers. So surge pricing is something that can address a sudden increase in demand. And I don't think it's a perfect solution, but at the same time like it. It is a way to solve that problem. It kind of stinks if you're one of the people who really needs a ride. That means that you're going to be

paying extra. If you're able to wait, then eventually that serge is going to drive enough drivers there to bring the pricing down. But then on the flip side, if you're a driver and you get to an area just as surge pricing is on the decline, kind of stinks for you because you missed out on an opportunity. So like, it's not a perfect system by any stretch of the imagination.

But let me do you one better. What if the AI knew how much money the customers were making, and therefore the AI could adjust the price on products and services in order to kind of give it that little personal touch. And by personal touch, I mean adjust ca so that they're as much, but not more than the customer would be willing to pay, perhaps more than what the person ahead of that customer or behind that customer is paying, but just the right amount to get the

maximum money from that customer. Does that seem fair? Imagine walking into a grocery store and finding out that the carton of oat milk you've picked up it's actually going to cost you forty cents more than it costs the person who is behind you. That probably wouldn't feel really good that. Keep in mind, this dynamic pricing is meant to cater to income levels, so being charged less means that you would be in a lower earning household, at

least assuming that the system is working as intended. So in some ways you can look at this as being kind right, bringing the cost of products to within the budget of lower earning customers. But that's not really what we're talking about here. We're talking about finding the price that you can set as your maximum before going to a point where people will just not buy anything at all. You want to find that spot where someone is going

to hand over money. It might hurt, but they're gonna do it because they need the thing, And it's essentially price gouging, right. And when you flip it over, when you say this is charging more for the same stuff because the person who's doing the shopping is a big earner, that feels wrong too. And yet this is something that dynamic pricing could end up doing. In fact, it's more than could end up doing. I'll explain more, but first

let's take a quick break. So before the break, I was talking about this sort of situation in which a store could use dynamic pricing to potentially personalize the shopping experience so that certain individuals would end up having to pay more for specific items versus others. Like you could have all the people in the store at the same time, and they'd each have an individual experience of how much

each particular item would cost. Here in the United States, the grocery store chain Kroger is facing some scrutiny from legislators like Senators Elizabeth Warren and Bob Casey over the implementation of dynamic pricing that is at least in part driven by AI. So the chain makes use of digital price labels, and around eighteen percent of its stores I think it's like five hundred locations across the United States.

They own and operate like twenty seven hundred and fifty, so five hundred is around a little less than twenty percent of their stores. But they have these digital price tags, and of course with digital price tags, you could change the price of things instantly. When the store makes use of proper labels like paper labels, I guess I shouldn't say proper. That's being really biased. Paper labels, well, there's not really the opportunity to switch things up so quickly.

You're not going to have a shelf stocker just walking directly ahead of a shopper, paying attention to what the shoppers looking at and then slapping a new label on in order to change the price. That's not going to happen. But in these Kroger stores with the digital labels, it's entirely possible to change the prices for specific items at a moment's notice. Now, let's just say that you have

these digital labels. You pair this with cameras that have things like facial recognition technology, and you gather information about customers. Maybe you know a lot about this customer already because they're in your loyalty program. Maybe they're using your app, and the app is communicating in real time what you're doing, Like if you're using it to scan products, it's sending

that information to the home base. Well, Kroger could dynamically change prices as you were actually moving down the aisles, and before you can actually look at the digital price, the amount has changed from the person ahead of you. That amount will be reflected ultimately when you check out of the store. So what's more, Kroger has been gathering customer information for years because those customer loyalty tags that

you can get at stores like grocery stores. Kroger has one, These entitle you to stuff like discounts, right well, it's also a tag that lets the store keep an ongoing tallly of everything you've ever purchased there. Through years of data gathering, Kroger knows the brands that specific customers gravitate toward, So the store knows if a specific customer is more likely to go for, you know, whatever brand is on

sale versus a premium brand. Right, some people have brand loyalty and they'll buy that brand no matter if the price is up or not. Others are more pragmatic and they will go with whichever brand is the cheapest at the time, at least cheapest by volume or whatever. And Kroger knows this. It also knows like if you prefer you're salsa, spicy or mild. You know, it knows what kind of so does you drink, and the snacks you crave and what impulse buys you're inclined to indulge in

when you get up there to the cashier. And sure, there are ways to use that information that can be of benefit both to the store and to the customer. But the concern is that through this system Kroger could engage in profiling and price gouging. That Kroger could dynamically change prices based on factors like race, age or sex. Now to be clear, the company has said that the idea is to quote lower prices more for customers where it matters most end quote, but skeptics remain well skeptical.

So in an open letter to Kroger's CEO, Senators Warren and Casey ask for more information about the digital price tags and Kroger's plans, and expressed concerns that the store could use customer and world information combined to quote calibrate price increases to a distract maximum profits at a time when the amount of American's income spent on food is at a thirty year high. End quote. The senators pointed out that Kroger system could let the store take advantage

of local conditions. So, for example, let's say that there is a big storm that's forecast to hit the area. Or let's say you live in a place like my hometown of Atlanta, Georgia, where the summer we have had numerous water main breaks happening all over the place. Well, you might find that a trip to the store means that you'll have to pay a premium if you want to get that bottled water you're gonna need while there's a boil water advisory in place and you have no

power at home. It's like being kicked when you're down. Now, am I saying Kroger is currently doing this, that it is proactively raising prices when there's a prediction that demands going to surge. I am not saying that. I'm saying that the system allows for it to happen and to do it quickly. And that's before you get to the potential issue of personalized pricing. I mean, the whole purpose

of the system ultimately is to maximize profit. That might mean that you're doing it in service of the customer, but it's not through an altruistic desire to help the customer. It's because that's the way you're going to end up making more money. And if it turns out that that's not the way you're going to make more money, then it's not going to happen. The letter is a really good read, and it isn't just a letter. The senators

cite specific reference materials. In fact, there are certain pages where the references take up more space on the page than the letter does, but they really show that they've done their homework like this isn't just them pontificating and theorizing that there's an issue. They're citing sources. They point out that it's not just Kroger that's rolling out a

system that's like this. Retailers like Walmart are doing it too. Meanwhile, several companies are really eager to find ways to let JI to figure out just how much money they can charge for their goods and services without taking off the

customer base and driving them to competitors. Forbes magazine has an article titled Harnessing AI for Dynamic Pricing for Your Business, and it talks about this, and it was written by a writer and business advisor named Neil Sahota, who explains that AI boosts the concept of dynamic pricing quote by analyzing vast data sets to predict demand fluctuations, understand customer price sensitivity, and identify the optimal price point that maximizes

revenue or market share end quote. Now, if we examine that statement closely, it can start to get sinister is probably not the right word, but certainly predatory could be a word I could use. Because you're talking about understanding customer price sensitivity. That phrase just means knowing how much you can charge before your customers decide they've had enough. Right, Like,

the wording kind of falls into that. Let's couch this in a way so it doesn't sounds as predatory as it is, but it's really all about how much can I charge you before you walk away from me? So yeah, it's a little scary. But in other words, the AI might also be taking into account, you know, like I said, how much people are willing to pay before they just

get fed up with the company and move on. And on one hand, all I'm really talking about here is a more real time approach to something that businesses have been doing forever. It's not like there's some sort of oracle out there that dictates what the price should be for a gallon of milk or a vacuum cleaner or

a new car. These prices are determined by the companies that sell the things, and the companies are taking into account lots of factors that contribute to the cost that they incur in providing those goods and services in the first place. Right, Like, nothing is free, a company has to take that into accoun out or else the cost of doing business will be greater than what revenue it pulls in, and ultimately the business will fail. So you

can't put all the blame on companies. You can't just you know, reduce them to greedy capitalists or something along those lines, although that can if they could turn into that, certainly, but that's the way business works. You gotta make money or else there's no reason to do the thing anymore.

But I think it's the immediacy of AI power dynamic pricing that really gives me the willies, as well as the ability to respond to or even predict conditions that will lead to increased demand, allowing for price gouging opportunities. But don't worry, it gets worse. Okay, I'll explain how, but first let's take another quick break. We're back, and one other element that really comes into play when we are talking about AI and price strategies and dynamic pricing

is the concept of competition. Now, if I own a grocery store and I decide I really want to squeeze every penny I possibly can from every shopper that I possibly can, and if I'm the only game in town, well I guess I'll get away with it, or I'll get run out of town one of the two, but presuming there are at least a few other grocers in the area, my customers will have alternatives to my store, and if they feel that my prices are unfair, they

can move to the competition, So that creates an incentive for me to lower my prices. However, let's imagine we have a group of business owners and all of them are employing aipower dynamic pricing. They're not talking with each other, but they're all using the same or similar tools in

order to figure out how much to price stuff. Now, all the competition is on a similar level, and it's possible that one business, by calculating thee at the competitors have increased prices, might actually lower prices on their items

or to lure people away from the competition. But it's more likely that the AI will effectively set prices across the board so that no matter where a customer goes, they're going to face higher prices than they would like, as high as the market will allow, and you end up with an AI powered cartel. In other words, this is not hypothetical. We have seen this play out in certain businesses in certain parts of the world. For instance, there are people who allege that this is happening in

rental housing. We turn to the tail of a software package called real Page from the company with the same name. So the company Real Page formed out of various acquisitions over several years, but its main product is property management systems for landlords, and one thing the software does is suggest rental fees for the landlord's properties. Now, the landlord first has to feed a lot of data into the system, including data that would not be publicly available, at least

not to the general public. So that would include stuff like how many units the landlord oversees, how many of those are vacant, how much is being charged in rent at the moment, etc. Now, Real Page's power comes from the fact that it's an incredibly popular piece of software and lots of landlords depend upon it, which means real

Page has access to a ton of data. So not only is real Page suggesting a rental price based off one landlord's specific situation, it's taking into account other rental properties managed by other landlords in the area and what those rental fees amount to. And this can mean that collectively Real Page starts to nudge markets into a sort of price fixed cartel. See. The whole goal of the software is to find that sweet spot in which landlords will maximize profits. So how much can you charge for

your rental property before folks just pass it over? You don't want to underprice your property. You don't want to leave money on the table, But you also don't want to price it so high that no one's going to rent it and it just remains vacant. So real Page's purpose is to arrive at the magic number that's going to result in the most profit for its customers. The landlords now applied to a single location, like a single

apartment complex. Arguably, this might not be so bad as long as people living in the rental property feel they're getting value for their money, it's not a huge issue. But when you start to apply this across a region, you reach a point where the rental market has little to no competition and people will struggle to afford a place to live. There will be no alternatives they can turn to. They'll just be looking at kind of a

landscape of high prices across the board. Critics of real Page have argued that the third party software has led to an anti competitive marketplace and that it turns landlords into a monopoly, even if the landlords themselves are unaware of this, and some critics are from really important positions

and authority. For example, back in twenty seventeen, a member of the Federal Trade Commission named Maureen Ohlhausen argued that algorithmically driven price strategies applied across multiple landlords is really no different from a human being collecting all this information and then telling all the competitors how to fix their prices so that they can maximize profits and not worry

about competition. Now, that would be against the law if a human being did it, and she was saying, maybe we should consider something similar for algorithmically driven strategies, because technically that's not accounted for in the laws. Real Page is a defendant named in several lawsuits that have accused the company of anti competitive practices. J Karma has a great piece on this titled We're entering an AI price fixing dystopia in the Atlantic, and that explains the Atlantic

is where the article's from. We're not entering an AI price fixing dystopia in the Atlantic anyway. This article explains how plaintifs I've argued that Real Page is essentially the lynch pen that's holding together a monopoly among landlords, at least in certain regions. And among the bits that Karma reveals is that Real Page allegedly strong arms its clients, so the landlords, in other words, to adopt the price suggestions that Real Page it submits, or else those landlords

risk getting the boot from the company. Karma has quoted an antitrust lawyer for the American Economic Liberties Project named Lee Hepner, who says, quote enforced compliance is the hallmark feature of any cartel end quote. The company I feel I should add disputes these claims and says that they are baseless. Karma also points out that in the United States, winning a case against a company that is allegedly responsible for anti competitive collusion is really tough because antitrust laws

do not take into account this particular situation. In your good old fashioned price fixing arrangement, you've got two or more entities that meet and they agree upon pricing, they remove competition from a region, and they milk the consumer for all their worth. But you have to prove that there's collusion between these two or more parties. Right. However, in this case, we're not talking about two entities that

are in direct communication with one another. It's possible they never met at all, they don't even know each other. They just both happen to use the same software package. So it's the software doing all the work of collusion, and you don't have any actual, traditional direct collusion between the parties themselves. And since the law doesn't account for this in most places, there's really no recourse unless laws are amended or new laws proposed to deal with the situation.

We're kind of in a bit of a fix. But wait, it gets worse. We don't even need a situation in which everyone is using the same software package. With Real Page. There's a dispute between how many landlords actually use it. University of Tennessee law professor Maurice Stuck says that more than forty markets in the US sees thirty to sixty percent of landlords of multifamily apartment units relying upon real

Page for their pricing strategies. Now, the company released a statement arguing that it's only used in seven percent of rental units. Karma, however, points out that this is largely because Real Page is talking about all rental properties, every single one, whereas what Stuck was saying was specifically multifamily apartment buildings. But even if we're talking at the low end of usage, if other landlords are using other property management software packages with price strategy driven by AI, you

can arrive at the same outcome. So Maurice Stuck also so studied the case in Germany in which two different gas stations, each using a different pricing algorithm, saw their margins increase by thirty eight percent once both of them were using these algorithms. Now earlier, when just one of them had employed a pricing algorithm, its price margin didn't change at all. It was only when each of them were on their own pricing algorithms that the situation changed.

So clearly, if you're a business owner, the lesson is that AI driven pricing strategies make you more money. They are more profitable, so the incentive is clear it is best to get on board. For customers, it's much worse news because it means you're being hit by the effects of collusion, even if what's actually going on isn't collusion. From the traditional legal perspective, the legal thing is likely to be an issue for a while because it takes time for laws to change, especially at a big scale.

All our regions have already taken steps. However, San Francisco recently passed an ordinance that forbids companies from using price strategy software that makes use of non public data among competitors in an effort to arrive at prices. So that's good news for the folks of San Francisco, a region where it is monstrously expensive, like the cost of living

in San Francisco is truly out of this world. Now, I'm sure we're going to see regulations and laws follow at some point to protect the free market, if not the customer, unless, of course, the company's profiting off these practices lobby hard enough to prevent any opposition to form within the government. But come on, how likely is that? Hi,

for one, welcome our robot overgrossers. Okay, that's it about how AI can help business owners, but in the same breath create an environment of collusion that hurts the consumer. And I don't think this is always going to be the case, but I think I think if we are depending upon AI for pricing strategy, it's more likely to happen than not, because if the mandate is find the way where I'm going to make the most money. Ultimately, this is where we arrive, right, There's no there's no

other destination but the effects of collusion. That's the only place you can go to. And the only alternative would be businesses that don't get on the AI train, that do try to compete by offering lower prices. And whether

or not that works, I can't say. I suppose in the short term it could force the AI driven strategies to adjust so that they're not being completely drained of customers by the competition, But I imagine there'd be a great deal of pressure to get that other business to get on board, you know, be a team player, even though it's legal to be in a team like that. I hope all of you out there are doing well, and I will talk to you again really soon. Tech

Stuff is an iHeartRadio production. For more podcasts from iHeartRadio, visit the iHeartRadio app, Apple Podcasts, or wherever you listen to your favorite shows.