Prices once reflected the product and the market. Increasingly, they can also reflect what a company knows — or thinks it knows — about you.
Surveillance pricing uses personal data, behavioral tracking, and algorithms to estimate what a consumer may be willing to pay. If the system can infer that you are in a hurry, financially stressed, or unlikely to walk away, your vulnerability can become part of the calculation.
Companies watch us. We deserve tools to watch them back and make exploitation costly.
Websites, apps, and data brokers can collect or infer information from searches, purchases, location, browsing history, and even mouse movements. Combined, those signals can build a detailed profile of what you want and how price-sensitive you may be.
The Federal Trade Commission began investigating this market in 2024, issuing orders to eight companies. Its initial findings showed that intermediaries serving at least 250 clients could use granular data to tailor prices, discounts, and product offers. The FTC cautioned that its published examples were “hypothetical.”
Surveillance pricing differs from ordinary dynamic pricing, in which prices move with inventory, time, or demand. Surveillance pricing is personal: the price, discount, financing term, or product placement can vary according to a consumer profile or group. The change may be subtle — a withheld discount, a costlier product ranked first, or a different financing offer.
The idea is older than the AI boom
In 2012, a Wall Street Journal investigation found that Staples.com displayed different prices after estimating customers’ locations. Shoppers near an OfficeMax or Office Depot were more likely to receive discounts; those farther away could see higher prices for the same item.
In 2015, ProPublica found that the Princeton Review offered the same online tutoring package for $6,600 to $8,400 depending on ZIP code. Residents of heavily Asian ZIP codes were 1.8 times as likely to receive the higher price, even in lower-income areas.
The company said the pricing reflected regional business conditions rather than race. Regardless, the outcome shows how supposedly neutral data can produce serious demographic disparities.
RELATED: Seattle votes to ban grocery ‘surveillance pricing.’ Your data shouldn’t set your food bill.
Orbitz also experimented with showing Mac users pricier hotel options first after finding they were more likely to book luxury rooms. It did not charge Mac and PC users different prices for the same room, but the lesson was clear: Even device type can become a signal about what a customer may spend.
More recently, a 2025 Consumer Reports investigation found that Instacart’s AI-enabled pricing experiments offered different prices for identical groceries at the same stores and times. About three-quarters of tested products varied, with some differences reaching 23%. The investigation did not show that personal profiles determined who received which price, and Instacart later ended the program. But the experiment showed just how easy it was to target basic needs.
Your urgency can become valuable data
The troubling question is what happens when individualized pricing meets urgency.
A search for an emergency plumber at 2 a.m., last-minute airfare after a family crisis, or a baby thermometer at midnight may signal that waiting is not an option. Searches for eviction help, payday loans, or emergency medical care can reveal pressure without a company ever seeing a pay stub or medical record.
Economist Benjamin Shiller showed why such information could be valuable. In a 2020 economic model using browsing data to estimate willingness to pay for Netflix, personalized pricing increased simulated profits by 12.99%, versus roughly 0.25% using demographics alone. It was a model, not proof of a retailer secretly charging desperate consumers more, but it showed the commercial value of behavioral data.
Data brokers have long packaged financial vulnerability. A 2013 Senate investigation found categories such as “Rural and Barely Making It,” “Retiring on Empty” and “Credit Crunched.” The report warned that such lists could appeal to sellers of high-cost loans looking for people who need quick cash.
RELATED: Welcome to the new traffic stop, where it’s your word against AI’s
The balance of power shifts
Two people can see different prices, discounts, products, or financing offers on the same website and each assume everyone else received the same thing. Algorithms do not need a field labeled race, age, sex, or income to make sensitive inferences; ZIP code, device type, location, and shopping behavior can act as proxies.
Targeted pricing is not inherently abusive. It could reward loyal customers or offer discounts to families under pressure. The problem is secrecy: Sellers can know far more about buyers than buyers know about the offer in front of them.
That imbalance can turn personal data into a tool for extracting the highest price a consumer will accept. It also weakens comparison shopping when prices quietly change according to profiles the shopper cannot see.
Transparency is the minimum
Companies should disclose when personal data or AI materially shapes a price or offer and identify the categories of information involved. Regulators have a role, but entrepreneurs can also build independent tools that compare prices across accounts, devices, and locations.
Companies watch us. We need better tools to watch them back.
When a company knows you are desperate, that desperation should never become a hidden advantage it can use against you.
