Personalized price tags are not the most advanced tool available to companies that seek to extract more money from customers. Pricing methods now combine data collection, algorithms, and product design to shift more of the cost of a purchase onto the buyer.

Price Differences in a Single Store

In 2025, researchers in the United States ran a test with about 400 people in different cities. All participants opened the same grocery delivery app, selected the same store, and ordered identical baskets at the same time. Nearly 75% of items cost different amounts for different users. Eggs at the same store in the same minute ranged from $3.99 to$4.79. The price of a package of cookies varied by 23%. On average, identical items differed in price by about 13%. Investigators estimated that a family of four overpaid by about $1,200 per year. The delivery service called the estimate "wildly misleading" but did not deny that the price differences occurred. The company stated that it sometimes runs short-term random tests to help stores understand what customers want.

Historical Background

Price tags are a relatively recent development. They appeared about 150 years ago. Before that, prices were set through direct negotiation, and a seller could charge a wealthy buyer more. Such practices, however, risked losing customers to competitors and required negotiation in every transaction.

Mass price tags emerged partly from religious principles. In nineteenth-century America, Quakers held that everyone is equal before God and should therefore pay the same price at the register. Fixed prices also proved financially beneficial. They reduced the staff time spent on bargaining and disputes, and customers found them more convenient.

Fairness and Pricing Perceptions

In 1986, two future Nobel laureates conducted telephone surveys on pricing fairness. One scenario read: "A store sells shovels for $15. In the morning, it snows and the store raises the price to$20. Is this fair?" Eighty-two percent of respondents said it was not acceptable. From an economic perspective, the store acted rationally: demand rose, so the price rose. Respondents, however, did not accept this reasoning.

The researchers found that customers judged price increases differently depending on their cause. Most respondents accepted increases caused by supplier price rises. The researchers called this the "dual entitlement" principle. Customers expect a stable price, and companies expect a stable profit. A price increase driven by higher costs is generally accepted. A price increase made only because the company can raise it generates resentment.

In 1999, a Coca-Cola executive described to journalists a pilot project for a vending machine with a built-in thermometer that would raise the price of cola in hot weather. The company argued that a cold drink is worth more when it is hot. The press criticized the idea strongly. The Wall Street Journal published criticism on its front page. PepsiCo publicly stated that it would never adopt such a practice and described it as exploitation of consumers. Coca-Cola abandoned the plan. Business schools later used the case, with the lesson that companies must not make such pricing visible. Any such practice requires a justification, such as a cost increase.

Cost Pressures and Visible Justification

During the pandemic, rising prices across the economy made it difficult for consumers to distinguish genuine cost increases from opportunistic pricing. When the United States imposed tariffs under the Trump administration, a president of a regional Federal Reserve Bank asked business leaders whether they were raising prices because of the tariffs or because of news coverage of them. The answer was both. Some companies had already been affected by tariffs that were in effect, and others raised prices in anticipation of future tariffs.

Companies also adjust prices through package design. Reducing package contents, changing bottle bottoms, or offering oversized packs all act on the effective price per unit. Reducing the content of a one-liter milk container indefinitely is not viable, because at some point customers stop buying the product. The central question is at what point and how long such reductions can continue.

Six Methods of Revenue Extraction

The German consultancy Simon-Kucher & Partners made this question its focus. It employed dozens of PhD-level researchers. Until retailers began using barcodes and automated systems, measuring customers' willingness to pay was largely manual and difficult. Barcode data gave consultants, for the first time, the ability to measure willingness to pay across billions of purchases. The firm identified six methods for increasing revenue from consumers. The following sections describe five of them.

Method 1: Collusion and Algorithmic Pricing

Historical cartels. The first method dates to 1924. In Geneva, light bulb manufacturers agreed to limit product quality because their products were too durable. They set a standard lifespan of 1,000 hours. A manufacturer that produced a longer-lasting bulb faced a fine. The arrangement held until World War II, when the cartel dissolved because its members ended up on opposing sides of the conflict.

A light bulb in Livermore, California, has burned since 1901, a period of 125 years. It can be viewed through a personal webcam.

Airline fare publication. In the late 1980s, airlines began publishing fares through a shared database. The stated purpose was to help travel agents. The actual function was to coordinate prices. Carriers announced fare increases and their effective dates and expected competitors to match them. If nine carriers announced a fare increase on a route, effective in one week, and a tenth carrier declined to follow, the nine would quietly withdraw their increases. They would then target the non-compliant carrier on its most profitable routes until it agreed to join the group. The U.S. Department of Justice estimated that passengers overpaid by about $2 billion as a result. The case was resolved through a settlement in 1994.

Algorithmic coordination. Internet-based tools expanded these practices. In 2024, an investigation began into RealPage, a company that held data on 16 million apartments. The service recommended rental prices to property owners based on the rents that neighbors actually paid, vacancy rates in the area, and current discounts offered by competitors. The algorithm was designed to raise prices. When the market rose, recommendations increased. When the market fell, the algorithm did not reduce prices. In a complex with declining rental demand, a "revenue protection" mode activated. The program preferred to keep some units vacant rather than lower prices.

Landlords were encouraged to enable an auto-accept option instead of reviewing each recommendation manually. In this mode, the system changed prices by up to 3% per day and up to 8% per week. RealPage's software controlled pricing for 45% of apartment complexes in Denver. The company calculated that its algorithms generated $136 more per month from each household than would otherwise have been the case.

In November 2025, the U.S. Department of Justice reached a settlement with RealPage. The company did not admit wrongdoing and did not pay a fine. It agreed to stop using non-public data from competitors. Several U.S. cities subsequently banned such software. Among them are San Francisco and Philadelphia.

The distinction is significant. If colleagues agree to maintain certain prices and not to lower them, the agreement constitutes illegal and condemned collusion. If the same coordination occurs through an algorithm, it is often treated as legal.

Method 2: Restricted Access

Restricting customers' access to products and repairs is the second method. In the United States, a recurring example is the ice cream machine at McDonald's. On average, one in every seven McDonald's restaurants cannot sell McFlurry desserts because of mechanical failures. The machines are manufactured by Taylor. Under the service contract, restaurant employees may not open or repair the equipment. They must call a technician and wait. A service call costs $300, and the repair itself takes about 15 minutes. Service generates roughly a quarter of Taylor's annual revenue.

The practice extends beyond restaurants. Mechanics in the U.S. Army cannot open and repair a generator, because doing so would void the warranty. They must call an external contractor.

In 2015, John Deere officially stated that a farmer is not the owner of the software installed in a tractor. Farmers were required to sign agreements prohibiting independent repairs. Dealers' technicians charged high fees for repairs. In response, many American farmers obtained pirated firmware from gray market forums to operate their tractors.

In 2026, John Deere agreed to pay $99 million to settle a class action lawsuit. For 10 years, the company must provide farmers and independent repair shops with the same diagnostic software, programs, and instructions it provides to its authorized dealers.

Another example involves smart cribs that cost about $2,000. They include multiple sensors, track sleep, play white noise, and rock the infant when crying is detected. After a period, some functions become subscription-based at$20 per month. Owners who bought the crib discovered that its core functions depended on a monthly payment for rocking.

Method 3: Concealing the True Price

The third method hides the actual cost from the buyer. Gambling businesses apply well-known techniques: no windows, no visible clocks, and entrances designed to be inconspicuous. One reported experiment used pleasant scents near slot machines on the premise that customers stay longer, relax, and increase their bets more readily. The author of the source material was not certain whether scents are used in practice, but the other techniques are consistently applied.

Online, similar techniques exist. An interface may make cancellation complex while making subscription simple. A user can subscribe to a streaming service with one click and then spend ten minutes trying to cancel.

In 2025, Amazon was fined $2.5 billion for using deceptive design to enroll customers in its Prime subscription, which many could not cancel easily. A similar case is ongoing against Uber. Regulators found that canceling Uber One required passing through 23 screens and completing 32 actions. The case remains in progress.

Drip pricing. A common tool that conceals the true price is the add-on charge. Consumers are shown a low initial price, and fees are then added step by step: a service fee, a booking fee, a processing fee, and others. Airlines use this extensively, charging for checked baggage, seat selection, and the option to sit next to a companion. In the United Kingdom, authorities are examining whether charging families for seating next to a child is lawful, or whether separating parents from children in the cabin is acceptable.

Cruise lines generate about 40% of their onboard revenue from seasickness medication, beverages, and food. The same logic has been proposed for rental housing. Landlords could charge separately for a concierge, parking, storage, meeting rooms, washing services, trash removal, and package reception. Landlords have complained that base rent alone does not generate enough income and that they need new monetization methods.

Method 4: Dynamic Pricing

The fourth method is flexible pricing, often called search pricing. A cloud appears and the taxi fare immediately rises to a level comparable to calling a rescue boat. Digitization of nearly every service encourages companies to adopt this approach.

In a resort town, the price of sunscreen can vary with temperature and even with the day of the week, rising closer to weekends. Energy drinks may be sold at the maximum price in the morning and evening and at a promotional price during the day. Raincoats and umbrellas can double in price as soon as a forecast predicts rain. Taxi fares involve so many variables that no one can state the true cost of a commute from home to work in advance.

Electronic shelf labels make such changes possible across stores. Walmart converted all 2,300 of its supercenters to electronic price labels, at a cost of $1 billion. Rema 1000 employees acknowledged that the company can change up to 2,000 prices per day using electronic labels.

New services have emerged around these capabilities. Some firms track physical inventory at competitors' stores and signal when a company should raise prices because competitors are running out of products. For event tickets, a sellout within an hour is now considered a failure by some sellers, because resellers may purchase large volumes and capture the margin. Sellers therefore release tickets in batches, with an algorithm raising prices as sales progress.

Method 5: Personalized Pricing

The fifth method, described as the most sophisticated, is personalized pricing. In the United States, it is called surveillance pricing. Under this method, a buyer pays more than others based on information about the buyer.

Early instances were identified in 2015. A company that offered exam preparation services, including books and tests, set prices based on the customer's ZIP code. Customers in wealthier areas paid more.

Tinder, in an earlier practice, charged $10 for its paid subscription to users under 30 and$20 to users over 30. Regulators subsequently banned this practice.

Cars can also collect data without the driver's permission. The data may include how sharply the driver brakes and accelerates and whether the driver exceeds the speed limit. General Motors collected data on its drivers for years and sold it to insurance companies. Drivers discovered that their insurance premiums had increased without an apparent cause. In January 2026, the U.S. Federal Trade Commission barred General Motors from selling such data for five years. The order does not prohibit collection.

A former director of global policy for Facebook in the Middle East and North Africa reported that the company tracked when teenage girls deleted their selfies. The company reportedly considered that moment ideal for advertising cosmetics.

Personalized pricing is only beginning to reach grocery retail. Grocery stores have less interest in charging a customer $2 more for potatoes than in collecting information. Kroger, the largest traditional supermarket chain in the United States, treats data sales as a separate business line. In 2025, this line generated$1.5 billion. The chain holds tens of petabytes of customer data. The data will reportedly be used to train artificial intelligence systems.

Consumers are increasingly using AI chatbots instead of traditional search and asking them for advice on decisions. The comparison is to asking a friend who works for a large corporation and earns a percentage of sales. That friend knows the consumer's interests, remembers past conversations, and has access to a large volume of purchase history.

Regulation in China and the United States. China has been the most aggressive regulator of personalized pricing, because it found that loyal long-term customers would be charged the most. Algorithms assume such customers will not leave. In 2021, a customer with diamond status at a Hilton hotel in China paid 3,000 yuan for a room. She later learned that the same room had cost others 1,300 yuan on the same day. A court awarded her triple compensation for price deception and false advertising. The court could not prove that an algorithm was responsible. From 2026, a new Chinese rule prohibits internet platforms from setting different prices for the same product based on willingness to pay, ability to pay, preferences, or habits.

In New York City, a law requires businesses to disclose personalized prices with the statement: "This price was set by an algorithm using your personal data."

The principal difficulty is that personalized pricing is very hard to prove. Whether disclosure improves the situation is unclear.

Conceptual Considerations

The price tag has long been common knowledge. It allowed buyers to compare value. Without a visible price, buyers are easier to mislead, as in a traditional open-air market.

A second question concerns the value of money. If one friend pays more for the same item than another, does that mean that one friend's money is worth less? Can a store determine the value of money?

Price tags were introduced on the principle that everyone is equal before God, and therefore equal at the register. The principle was never an official law. It was a social convention. One party to that convention no longer finds it acceptable.