Surveillance Pricing Definition and Data Research Note

Surveillance Pricing Definition and Data Research Note

Working definition

The FTC uses surveillance pricing for systems that draw on detailed information about people to categorize them and set or influence targeted prices. The agency's current surveillance-pricing hub says the information may include location, demographics, and browsing history. Its July 2024 study orders focused on intermediaries whose technology could use personal data to set individualized prices for the same goods or services.

The term should remain narrower than algorithmic pricing. A pricing algorithm may use inventory, time, capacity, public competitor prices, or other market data without using personal information. Surveillance pricing begins when observed, purchased, or inferred information about a person or group informs the offer.

What the FTC has documented

The FTC's January 2025 initial study findings reported that the intermediaries it examined could use direct consumer data, inferred data, and first-party or third-party sources. Examples included precise location, browser history, shopping history, channel, time, cart abandonment, and mouse movement. The agency described hypothetical uses because material collected through the 6(b) process must be aggregated or anonymized to protect confidential information.

That qualification is important. The findings establish industry capability and documented intermediary practices. They do not show that every client used every input or that any named retailer charged a named person more because of a specific trait.

The FTC's Issue Spotlight maps possible sources across direct collection, web activity, mobile and device information, location, transaction records, and data brokers. It also separates a pricing output from product steering, discount targeting, ranking, or other choice architecture. A higher-priced result at the top of a page can affect what a person pays even if the nominal price of an identical item is unchanged.

From observation to a pricing decision

The useful public model has six stages.

Possibility is not attribution. A claim about a particular seller needs evidence that connects the relevant input to the actual decision. Useful evidence may include an audit, system documentation, a disclosure, experiment assignment logs, discovery records, or a controlled comparison that holds product, time, location, inventory, fees, and eligibility constant.

Estimating willingness to pay

Willingness to pay is not directly visible. A seller can estimate it from choices at different prices, prior transactions, response to promotions, product searches, time spent, abandonment, or other signals. Jean-Pierre Dubé and Sanjog Misra's field research on personalized pricing used randomized prices to estimate demand and validate personalized offers. The work demonstrates a method and a set of welfare results in one large digital firm. It is not proof of a universal effect.

The OECD's Personalised Pricing in the Digital Era explains that digital markets can support fine-grained price differentiation through automated data tools. It also stresses that effects vary with competition, market coverage, and implementation. The public article should therefore avoid saying that personalized pricing always raises prices or always harms consumers.

Publication boundary

Use may, can, or documented capability when the source describes what an intermediary supports. Use did only when primary evidence connects a seller, data input, decision rule, and outcome. Protected-trait, health, disability, desperation, and vulnerability claims require particularly strong evidence because proxies and inferences can be wrong and the reputational stakes are high.