26 Oct 2025 Search

One other comment on living with a bigger ad duopoly: what about personalized/surveillance pricing? What if, even though the surveillance advertising duopoly has to keep driving up customer acquisition costs, brands can pay for it by using ad personalization to charge each customer what the surveillance data says they’re willing to pay? Personalizing every step of the customer journey is usually about at what price a product or service is offered. Most products and services aren’t personalized. There are some exceptions, such as custom guitars from Creston Electric Instruments, but for most brands, personalization doesn’t mean they’re going to reschedule an airliner or launch their CAD software. In practice, personalization is a polite expression for surveillance pricing.

Quite a bit of discussion of the surveillance pricing issue lately, even though the FTC has taken down the Request for Public Comment Regarding Surveillance Pricing Practices (PDF). There’s even one argument in favor of surveillance pricing: In defense of ‘surveillance pricing’: Why personalized prices could be an unexpected force for equity by Prof. Aradhna Krishna. In reality, though, surveillance pricing is ML figuring out who has the option to get a better alternative elsewhere, so tends to identify, for example, shoppers with disabilities who are often limited in where they can shop due to mobility limitations, difficulty finding accessible transportation, or other factors related to their disability. When the same or similar products are available to rich and poor shoppers, the poor pay more for the same thing. The longer you can wait to buy, and the more options you have, the lower the price you get.

Prof. Nikita Garg writes, in AI is using your data to set personalised prices online. It could seriously backfire, The challenge for business is to deploy AI pricing ethically and transparently, in ways customers can trust. That’s harder than it looks. Research shows that Algorithms Can Drive Up Prices by learning to collude to raise prices, even if no collusion was programmed in. People don’t have a lot of time to learn about this stuff while shopping, so a safe assumption is that the personalized price is somehow unfair—whether it’s because of exploiting information that the seller knows, or collusion, or both.

But there’s one more problem, for the brand side. When making the decision to price discriminate or not, remember that anyone who is good at price discrimination is also going to be good at setting their own price. If a brand uses one of the big surveillance companies to match an ad to a high-spending consumer, the surveillance company will bid up their own ad rate to capture more and more of the difference between the price offered and the minimum price the brand would accept from a bargain-hunting customer A brand built on surveillance advertising and surveillance-enabled personalized pricing is like a popular San Francisco restaurant—when the landlord sees the restaurateur making money, they raise the rent. Jakob Nielsen’s Search Engines as Leeches on the Web applies to any auction-based advertising platform that can dynamically adjust advertising rates to capture as much as possible of the expected profits from a sale. Look carefully and you’ll spot brands trying to dig out of the CAC crunch, sometimes by going old school: Why startup brands are increasingly flocking to this Midwestern department store. Spotify is a ripoff for recording artists, so it’s back to vinyl. Surveillance ad platforms are a ripoff, so it’s back to…the mall?)

I still don’t think that a surveillance advertising ban is realistically possible in the current political climate, but when it happens it will be a one-time cash cheat code for the whole economy. By the time a jurisdiction realizes that a ban is necessary, the hard part will be timing it so that all the interested political parties or factions can get a piece of the credit for the resulting good news in the statistics.

More: advertising personalization: good for you?