Klaus Ohlenschläger/picture-alliance/dpa/AP Images
Last month, I wrote for our special issue on pricing about an emerging development: companies charging different prices for the same product, based on a customer’s willingness to pay. New technologies for data collection and the ability to isolate the consumer in private settings are enabling what law professor Zephyr Teachout calls surveillance pricing.
In that story, Federal Trade Commission chair Lina Khan told me that, while economists have long studied price personalization, it was previously more of a “thought experiment,” something to muse about. But advances in data extraction and targeting made it “much more possible to be serving every individual person an individual price based on everything they know about you.”
Yesterday, Khan’s FTC followed through on this notion by launching an informational study of surveillance pricing, seeking information from eight middlemen that promise artificial intelligence–based pricing strategies to companies. The decision to commence a study was unanimous, with both Republicans on the commission authorizing it as well as the Democrats.
It’s obviously gratifying to see policymakers take action on your work. That’s especially because as a journalist, I don’t have the ability to compel document production. A federal agency is necessarily going to have more tools to understand what is an incipient marketplace for differential pricing.
“We have tried to gather information from other sources, from journalists and academics,” said an FTC official I spoke with on Monday. “We gathered everything we can without subpoena power. But a lot of this information is proprietary.”
The very nature of surveillance pricing makes it hard to document: The whole idea is that there’s no single price being publicly posted and served. But I was able to see a lot of excitement and hype among the kind of companies that want to make money by offering retailers surveillance pricing tools. Uber-consultant McKinsey even came up with the perfect euphemism for the practice: “digital pricing transformations.”
We’re clearly in a stage where the architecture to support surveillance pricing is being built, and it goes way beyond “dynamic pricing” efforts at Wendy’s or Walmart. We know that companies specialize in creating apps to deliver individualized offers. The McDonald’s app, for example, has 150 million active members. We know that smart TVs are adding capabilities for specific in-home advertising, where watchers can buy products right off their remote.
We’re clearly in a stage where the architecture to support surveillance pricing is being built, and it goes way beyond “dynamic pricing” efforts at Wendy’s or Walmart.
We know that apps and Wi-Fi-enabled devices can capture customer information about individual demographics, product preferences, ordering habits, locations, social interactions, and even the dates of a customer’s payday, when they have more money to spend. And we know that first-party data can combine with outside information to construct an identity graph, like email and browser activity, social media likes, subscriptions, viewing behaviors, app downloads, or travel and auto and retail and medical histories. All of this information can be exploited to pinpoint the willingness to pay, maximizing profit and straining customer wallets.
That much we know. And we’ve seen differential pricing in scattered practice in digital commerce. CEOs have even discussed it directly in earnings calls. But we don’t have the kind of fine-grained information necessary to build a regulatory or law enforcement framework. So the FTC is seeking that material.
The agency targeted eight intermediary companies—Mastercard, Revionics, Bloomreach, JPMorgan Chase, Task, PROS, Accenture, and yes, McKinsey. Some are big companies in their own right, while others are part of the burgeoning cottage industry of pricing consultants. They all have marketed themselves to a broad cross section of industries (retail, restaurant, grocery, travel, finance, and hospitality, to name a few), specifically cited AI and machine learning capabilities, and advertised surveillance pricing strategies like personalized ad inserts or location-based targeting.
The orders to the intermediaries are incredibly specific and far-reaching. The FTC wants all their sales data and business models, what customers they have transacted with, and each “targeted pricing solution” they have developed or licensed, including the technical details. They want to know the data sources that go into the targeting that helps create the pricing, how that data is being collected and/or retained, and what steps are taken to acquire user consent and safeguard privacy. They want to know how that data informs pricing, and whether the businesses that use their services obtain that data in accordance with all consumer privacy laws. And they want to know how it impacts consumers, now and in the future. The intermediaries have 45 days to deliver the information.
“We want to better understand consumer and sales information,” said the FTC official. “Some intermediaries have highlighted testimonials but it’s very vague. We want to know what companies are using pricing technology, and how, and to what extent.”
A 6(b) study like this attempts to expand agency knowledge and inform future policymaking. Doing it before surveillance pricing really gets off the ground is a smart strategy. “Enforcers made a mistake two decades ago with social media,” the FTC official said. “Now we’re taking it seriously, but the ability to uncrack the egg is going to be limited. We want to understand this early on.”
There’s an obvious consumer protection angle here, with the potential unfairness of price discrimination. But it could also tie in to data privacy, and the use of data collection itself as an unfair practice. And companies with the power and resources to engage in surveillance pricing may trigger competition concerns as well. The more data available for this kind of pricing, the better a company will get at it. And really only the biggest companies will have access to that much data.
In a way, this study is also playing catch-up. We allowed tech platforms and data brokers to go on a mass collection spree for a couple of decades, and we shouldn’t be surprised that all that information could be employed to exploit consumers.
Advocates are generally thrilled with this study. The White House has been through a tumultuous period. But at the agency level, officials tasked with protecting the public—in this case, on a bipartisan basis—are following that mission with a rare anticipatory strike at the kinds of business trends that usually end painfully for all of us. We don’t know what’s in store for Lina Khan and her colleagues in an administration with a new president that will start next year. There are already ongoing discussions about who might be replaced in a potential Harris administration. But Khan and her colleagues continue to prove their worth to the public.
