This assumes the information is clear and consistent enough across time and distance for arbitrage to happen. Pricing in-app, per customer, changing per day would introduce too much unpredictability for most customers to attempt arbitrage. If people in a group all check their apps, and the person with the best prices orders for everyone, it could work in the context of a shared meal.
But imagine trying to sort out X number of people who each want a different basket of items from, say, the Walmart app. Each of those items fluctuating daily in price for each customer independently makes arbitrage almost prohibitively difficult to coordinate.
The best case scenario is something like Steam sales, where a wishlist function notifies you when items you've "watched" are on sale. There are third parties like, for example, Deku Deals that track this pricing data across time for console games.
But Amazon is already trying to banish external AI agents from any access to its data. And what does a price history graph even mean if prices are specific to each customer and stochastically varied each day to induce impulse purchases?
what stops anyone from creating a third party order book that allows people to submit bids and offers on price discriminated items? It can match buyers and sellers just like a stock exchange.
The vendors who want you to just buy things in their app will treat any such exchange adversarially, and will ultimately always have the upper hand.
They can respond with litigation, as Amazon already is against third-party LLM agents accessing their marketplace. They can respond by banning accounts for violating the terms of service, making examples out of those who profit the most. They can watch the external marketplaces and cancel (undelivered/unfulfilled) sales they believe are linked to arbitrage.
All they need to do is make it inconvenient enough to discourage 80-90% of customers from participating in arbitrage.
But they are doing this all for what? Won't the market average out to the same unit price at the end of the day even if they can successfully create discriminatory spreads?
Think more in terms of behavioral psychology rather than idealized market dynamics which require rational actors and easily accessible information. Each corporation wants to optimize their customers' behavior for efficient extraction of wealth.
They want each customer effectively siloed in an ephemeral, eternal now: whatever the phone screen presents in this moment, and little else. The consumer may have a few scattered memories for context when presented with a potential purchase, but ideally isn't tracking prices or doing much research. The goal is to create those circumstances and (within them) reduce friction spending money as close as possible to zero.
Do that to as many customers as you can. Subvert their software and turn their own computers against them to achieve it. Instill learned helplessness and stimulus-response leading to purchase. Unit price and revenue will sort themselves out once you have a bunch of addled addicts staring at your shiny products in a digital environment you design and control.
That's the game. And that's why these companies will oppose arbitrage with all they can bring to bear, and fight with the brutal jealousy of gangs defending turf.
But imagine trying to sort out X number of people who each want a different basket of items from, say, the Walmart app. Each of those items fluctuating daily in price for each customer independently makes arbitrage almost prohibitively difficult to coordinate.
The best case scenario is something like Steam sales, where a wishlist function notifies you when items you've "watched" are on sale. There are third parties like, for example, Deku Deals that track this pricing data across time for console games.
But Amazon is already trying to banish external AI agents from any access to its data. And what does a price history graph even mean if prices are specific to each customer and stochastically varied each day to induce impulse purchases?