This week we talk about surge pricing, Walmart, and the Robinson-Patman Act.

We also discuss personal data, AC settings, and Delta’s earnings call.

Recommended Book: How the World Became Rich by Mark Koyama and Jared Rubin

Transcript

The US Robinson-Patman Act of 1936 is also called the Anti-Price Discrimination Act, and it was passed to make it illegal for a product supplier to charge different prices to different customers.

So a company that makes candy bars wouldn’t be allowed to charge one price to most of their customers, all the smaller and mid-sized convenience stores and mom-and-pop grocery stores, for instance, and then a lower price to the big stores, the Walmarts and Amazons of the world.

The concern was that these larger players, which at the time this law was passed were burgeoning grocery stores like A&P, would be able to achieve a monopolistic position in the market for these goods, these slightly lower prices giving them one more advantage over their smaller competitors.

During the four decades or so of this Act’s enforcement, small grocery stores has prices that were, on average, about 1% higher than those offered by their large competitors, and the eight largest grocery store chains only captured about 25% of all grocery sales in the US—essentially every city and town of any size had at least one small grocery store, and most had several of them, during this period. It was a very competitive market.

During the Reagan administration in the 80s, though, enforcement was abandoned, as the folks in charge of that enforcement were convinced this Act was holding back growth; they saw it as a handout to small businesses at the expense of big business, so while it technically remained on the books, they just stopped enforcing it, and the big businesses in these spaces got the message pretty quickly.

Walmart was the first big business to really lean into the new powers afforded them by this fresh governmental stance, and that led to it becoming the country’s largest grocery store chain by 2001, and other big grocery brands, like Kroger and Safeway, began to do the same, consolidating all their buying so they could put in huge orders like Walmart was able to put in, and that allowed them to demand lower prices, which in turn allowed them to dramatically increase profits and gobble up their smaller competition.

All of which led to the emergence of food deserts across the country, a term that was coined in 1995 to refer to areas where there are simply no grocery stores within a reasonable distance of relatively large populations of people, because smaller grocery stores can no longer compete, even when they’re the only player in town; folks have to travel to the larger chain stores, and have no real options closer to home, which can result in food precariousness, and situations in which the only nearby food options are unhealthy ones—the snacks at gas stations, for instance.

This same general pattern played out across all retail spaces, including pharmacies and bookstores and athletic supply stores, and between 1982 and 2017, the total market share of independent retailers in the US dropped from 53% to 22%.

Which in some ways is great at the federal level, as—and this is what the Reagan administration seemed to want, back in the 80s—big businesses can grow a lot faster and bigger than small businesses, and that can lead to outsized GDP numbers, and other such macro-scale figures.

Unfortunately, while independent retailers tend to keep nearly half of the revenue they pull in within their local community, major chains only keep something like 14% in the local community—so the shift from independent to chain retailers has had a deleterious impact on communities across the US, in the sense of having less competition, having food and other sorts of product deserts, and in terms of tax revenues and overall economic wealth being sapped from these areas and moved to other places, creating some relatively few winners and a whole lot of losers, in the process.

What I’d like to talk about today is another type of variable pricing, this one more directly aimed at consumers, and enabled, at least in its modern incarnation, by big data and the devices we use every day.

—

Dynamic pricing refers to changing the price of goods or services based on all sorts of variables.

Demand or surge pricing, for instance, might see the price of a bus ticket or rideshare ride with Uber cost more during rush-hour, the idea being that there are only so many bus seats and only so many available rideshare rides to go around, and when everyone’s either trying to get to work or get home from work, there will be a lot more people wanting these finite number of seats and rides than there are seats and rides available.

Upping the prices, then, is a means of determining who wants these things the most, because they’re willing to pay at times massively inflated prices for something that would cost far less in an hour or two, once the rush has subsided.

Similar price-inflation occurs during peak energy-use periods, and energy companies usually explain this price-bump by suggesting that it encourages their customers to use more energy when it’s abundant and cheap, and to use less of it when it’s scarce and expensive.

On very hot days when everyone is using their air conditioners to stay cool, then, inflated energy prices might encourage them to be less aggressive with their AC settings, keeping their indoor temperatures at a more reasonable level, which in turn ensures there’s more energy available for everyone and less risk of brownouts or blackouts.

This pricing strategy is often seen by those on the receiving end of such price-bumps, as price gouging, which refers to companies taking advantage of temporary variables to massively inflate their prices, at times to abusive levels that they can justify by pointing at those variables and a desire to moderate supply and demand.

So if there’s a big convention in town, local hotels can argue that they’re doubling or tripling their prices because there are not enough rooms for everyone who wants rooms on those days, but this could also be construed as a money-grab, these hotel companies knowing that some people won’t be able to avoid paying for a place to stay during the convention they have to attend, so they’re taking advantage of customers who have no choice but to pay up.

We saw similar dynamics play out globally during the height of the Covid-19 pandemic, when folks who had high-quality masks on hand were able to charge incredible sums for those masks because production hadn’t yet scaled up, so they were relatively scarce and thus precious, and these people and companies with the right product at the right time knew they could get away with charging many times the actual sticker-price of that product, because some people would feel they had no choice but to pay it.

Each situation of this kind will feel reasonable and suitable for the supply-demand situation to some, and completely unreasonable and abusive to others, and it’s possible to have a bit of both in many such situations—the companies in question actually want to manage a scarce supply of something, but are also keen to make as much money as possible while doing it.

Dynamic pricing has become even more common in online marketplaces like Amazon, where it’s not just holidays or events or the sudden emergence of global pandemics that can impact demand and thus, the prices retailers can get away with charging would-be customers.

Amazon has algorithms that keep track of what competitors are charging for the goods they offer, what sort of demand the market is seeing for said goods, what inventory looks like—if they have a lot or very few of something available to sell—and all sorts of other factors that might reasonably impact the price of a product, even a little bit.

As of 2024, the price of a product listed on Amazon changes several times a day, in some cases every 10 minutes, and they make about 2.5 million prices changes every single day, adjusting for those aforementioned micro-scale variables, on a product-by-product basis, but also adjusting their entire catalog so that relatively uncommon goods have higher prices, but common goods have lower prices, which means customers shopping around will tend to see Amazon’s lower-priced goods more often than the higher-priced ones, which in turn can adjust their perception of the company and its marketplace in a favorable, lower-price direction.

Amazon also has access to just a silly amount of data about their customers, some of it scooped up while we surf their sites, and some bought from other data-aggregators. And this allows Amazon, just like most tech companies and retailers, these days to track our behavior, watching what we click on, how long we linger on different products or product types, noticing our searches and contextualizing all of it with where we live, what we’ve purchased in the past, and so on.

The company isn’t very transparent about how it uses all this personal data, but while it’s been been speculated that they might adjust prices based on our individual profiles, most evidence suggests they mostly use it to determine what we’re shown—what products are promoted to us, basically, as opposed to setting prices based on what it thinks we’ll pay, as individuals.

The same generally seems to be true of other retailers right now, though there are concerns that this might change at some point in the near-future, as new technologies, some based on AI, enable the more-rapid and sophisticated crunching of data, and the consequent individualization of prices, even in person.

US airline Delta, for instance, recently announced that it would be using AI to help it boost profits by charging different customers different prices for the same airline seat.

These prices would be based on their customer profile, which means all the data scooped up by Delta from various sources, including things like past purchases, regular flight schedules, and how much money their systems think each customer makes and has available to spend.

The president of the company said on a recent earnings call that they’ve been running a pilot project for this approach that resulted in about 3% of ticket sales being sold based on this model over the past 6 months, and by the end of the year, their goal is to increase that to 20% of tickets.

In theory, this sort of system could be good for some customers some of the time, because it could drop prices on tickets that customers wouldn’t want to, or wouldn’t be able to pay for, otherwise. If I’m considering a trip, but the tickets are more expensive than I want to pay, these systems could theoretically recognize this and offer them to me at a price they can afford to sell them at, and which I can afford. That could lead to more ticket sales, and thus, higher profits.

The evidence on the ground with these sorts of systems usually points at price increases, not decreases, though: the companies using these models to see how much they can get per unit, not using them to sell more units at lower profit margins.

In other words, usually it’s wealthier consumers who get the better deals, as these companies want to keep them coming back, spending larger sums of money on glitzier products and services over time, while poorer consumers have fewer options, and will thus tend to pay whatever they’re told they have to pay.

Delta spent most of July 2025 trying to control the backlash that erupted following that earnings call, and they’re now saying, to the press but also in formal letters to government watchdogs who expressed concerns about what they said they planned to do, that no no no, we misspoke, we’re not using individualized data to set prices, it’s all good, don’t worry about it.

That announcement from Delta came shortly after lawmakers announced they would be pushing to get a new act, the Stop AI Price Gouging and Wage Fixing Act, passed into law, and though some US Senators have said they’ll block such efforts by Delta, other airlines, including Azul, WestJet, Virgin Atlantic, and VivaAerobus are also clients of the Israeli company, Fetcherr, that Delta has been working with to run their AI pricing pilot program—and representatives from Fetcherr have claimed that this pricing model is irresistible to those in charge of these companies, so it will probably take over the airline industry relatively quickly, and they plan to expand into other industries soon.

These sorts of pricing models aren’t typically very popular with customers, and efforts by Walmart and other big grocery chains to remove static in-store pricing labels and replace them with digital versions, or in some extreme cases to remove them entirely and rely on apps on customers’ phone to show prices on goods, raised similar alarm bells, as dynamic pricing can allow the store to more rapidly change their prices based on demand, like Uber’s surge pricing model, but maybe applied to flour or cough medicine instead of rideshare seats, and in-app pricing could allow them to show different prices to different people shopping for the same thing at the same time—again, based on income, buying patterns, and so on.

Walmart and everyone else dabbling in this space has, like Delta, claimed they intend no such dynamism in their pricing, even as their CEOs in some cases continue to brag to investors about the possibilities. As a result, there seems to be a decent chance we’ll see the large-scale deployment of these sorts of models in at least some customer-facing industries within the next year or two, some company deciding to more fully test the regulatory establishment’s appetite for challenging this push into a new pricing paradigm that would, theoretically at least, allow big companies to earn still-higher profits and grow even larger.

Show Notes

https://drive.google.com/file/d/1HQoQhvfVv8p0XmOdDIiWTnmd2YM_za07/view

https://www.businessinsider.com/amazon-price-changes-2018-8

https://en.wikipedia.org/wiki/Algorithmic_pricing

https://en.wikipedia.org/wiki/Dynamic_pricing

https://www.archeraffiliates.com/post/amazon-dynamic-pricing

https://arstechnica.com/tech-policy/2025/08/delta-denies-using-ai-to-come-up-with-inflated-personalized-prices/

https://arstechnica.com/tech-policy/2025/07/will-ai-end-cheap-flights-critics-attack-deltas-predatory-ai-pricing/

https://www.the-sun.com/money/14839597/walmart-kroger-electronic-labels-dynamic-pricing-demand-wendys

https://www.nytimes.com/2024/10/23/business/kroger-walmart-facial-recognition-prices.html

https://www.nerdwallet.com/article/finance/what-is-dynamic-pricing

https://www.theatlantic.com/ideas/archive/2024/12/food-deserts-robinson-patman/680765/

https://www.indieretailermonth.com/statistics

https://en.wikipedia.org/wiki/Robinson%E2%80%93Patman_Act

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit letsknowthings.substack.com/subscribe

Chris Pine

Happy Sad Confused

Introduction

The AI-powered Podcast Player