Why the FTC's Personalized Pricing Policy Must Contend With Goodhart's Law

The moment a retailer's algorithm selects which data points to weigh for a specific customer is the exact moment the Federal Trade Commission's new personalized pricing policy will either succeed or fail. On August 19, 2026, the FTC proposed an enforcement policy statement requiring businesses to clearly disclose when they use personal data to set individualized prices. The agency's logic is straightforward: if consumers know a price is tailored to their data, they can comparison shop or use privacy tools to avoid the markup. But this disclosure mandate collides directly with Goodhart's Law, the economic principle stating that when a measure becomes a target, it ceases to be a good measure. By forcing algorithms to declare the specific personal data they use, the policy mathematically incentivizes those systems to optimize for proxy variables that remain undisclosed, potentially leaving consumers with the same personalized prices but even less transparency about how they are calculated.[5][7]

The FTC's proposed enforcement policy statement, issued following a 2-0 vote, targets what the agency calls "surveillance pricing." The agency asserts that failing to disclose personalized pricing based on consumer data constitutes a deceptive practice under Section 5 of the FTC Act. The mandate requires businesses to clearly state three specific things: that a price is personalized, the basis for the personalization, and the types of data used. The proposal follows a July 2024 initiative where the FTC issued compulsory process orders to eight companies to study their data-driven pricing tools.[3][5][6]

The mechanism of personalized pricing relies on vast datasets to estimate a consumer's willingness to pay. Algorithms analyze variables ranging from location and device type to purchase history and real-time browsing behavior. A recent investigation highlighted by CBS News found that algorithmic price testing could cause grocery totals to fluctuate by as much as 23 percent, potentially costing families more than $1,200 a year. FTC Chairman Andrew Ferguson stated that when consumers see a listed price, they expect it to be the same price everyone else sees, "not the retailer's estimate of how much they are willing to pay based on their personal data." The FTC argues that hidden personalization deprives shoppers of the ability to make informed purchasing decisions.[4][5]

However, the application of Goodhart's Law—named after British economist Charles Goodhart—suggests a structural vulnerability in the FTC's approach. The law posits that any observed statistical regularity collapses once pressure is placed upon it for control purposes. In the context of algorithmic pricing, the "measure" is the specific personal data used, and the "target" is compliance with the FTC's three-part disclosure requirement.[7]

If a retailer must explicitly disclose that it uses "browsing history" or "zip code" to set a price, the algorithm's objective function shifts. To avoid triggering the disclosure requirement—or to minimize its impact on consumer trust—the system will naturally seek out proxy variables that correlate with willingness to pay but do not fall under the strict definition of the disclosed personal data.[2][7]

For example, instead of using a consumer's exact location, an algorithm might weigh the time of day a purchase is typically made, the specific combination of items in a cart, or the speed at which a user scrolls through a page. These proxies can achieve the same personalized pricing outcome without relying on the explicitly regulated data categories. The result is a system that complies with the letter of the FTC's policy while entirely circumventing its intent.[2][7]

The Computer & Communications Industry Association (CCIA) argues that the FTC's policy could also carry unintended consequences for consumers who benefit from algorithmic pricing. Personalized pricing often manifests as targeted discounts, loyalty rewards, or lower prices for price-sensitive shoppers. Mandating broad disclosures could disincentivize retailers from offering these discounts, effectively raising the baseline price for everyone.[1]

Furthermore, the legal framework of the FTC's proposal relies on the premise that consumers expect uniform pricing. Yet, dynamic pricing—where prices fluctuate based on market-wide supply and demand—is already widely accepted in industries like airlines and ride-sharing. The line between dynamic pricing and personalized pricing is increasingly blurred, making enforcement based on consumer expectations a moving target.[3][6]

The central assumption of the FTC's policy is that transparency automatically leads to consumer empowerment. While disclosure remains a foundational principle of consumer protection, algorithmic systems are uniquely capable of adapting to transparency mandates by obscuring their mechanisms further down the causal chain. If a model learns that explicit demographic data triggers a compliance flag, it will simply mathematically reconstruct that demographic profile from unregulated behavioral metadata.[2][7]

The success of the FTC's enforcement policy will not depend on whether businesses publish disclosures, but on whether regulators can mathematically prove that an algorithm is using proxy variables to achieve personalized pricing. Until that technical hurdle is cleared, the policy risks creating a compliance theater where prices remain personalized, but the data driving them becomes even harder to trace. The next verifiable checkpoint arrives when the 30-day public comment period closes and the FTC initiates its first enforcement action, testing whether Section 5 can actually constrain an algorithm that has already learned to optimize around it.[3][6][7]