Food Delivery Apps Use Secret Pricing Algorithms to Charge Customers Different Rates

When two neighbors in the same apartment building open the same food delivery app and search for the same items from the same store, they might reasonably expect to see the same prices. They would be wrong.

A growing body of evidence — from federal investigations, academic research, and consumer advocacy groups — reveals that the food delivery and grocery delivery industry has quietly built a pricing infrastructure that tailors what you pay based on who you are, where you live, how you browse, and what the algorithm calculates you're willing to spend. The practice, which regulators have dubbed "surveillance pricing," represents one of the most significant and least understood shifts in how Americans pay for food.

The Instacart Experiment

The clearest window into how algorithmic pricing works in practice came from a monthslong investigation by Consumer Reports and the Groundwork Collaborative, published in December 2025. Their findings were striking: nearly 75% of products tested on Instacart were offered at different prices to different shoppers at the same time, from the same store [1].

The investigation relied on more than 400 volunteers participating in live online shopping sessions. They found price variations ranging from 7 cents to $2.56 per item, with some products carrying as many as five different price points simultaneously. Identical grocery baskets differed by up to $9.59 between customers [1]. Extrapolated across a year of shopping, the price inflation could cost a family of four an additional $1,200 annually [2].

The engine behind these experiments was Eversight, an AI-powered pricing platform that Instacart acquired in 2022 [6]. The technology allowed retailers to run what the industry calls "in-market experiments" — testing untried price points on live customers to gauge their sensitivity to higher and lower prices across different product categories [6]. Major grocery chains including Albertsons, Costco, Kroger, Safeway, Sprouts Farmers Market, and Target were found to be participating in these algorithmic pricing tests through the Instacart platform [1].

The backlash was swift. Within weeks of the investigation's publication, Instacart announced it was ending all item price tests, and retailers would no longer be able to use Eversight technology to run pricing experiments on its platform [3]. The Federal Trade Commission followed with a $60 million settlement, requiring Instacart to provide refunds to hundreds of thousands of affected customers — not only for the AI pricing practices but also for deceptive marketing of "free delivery" that concealed mandatory service fees of up to 15%, and for opaque subscription enrollment practices [4].

Source: FRED / Bureau of Labor Statistics

Data as of Mar 15, 2026CSV

The Scale of the Problem

The Instacart case, while the most thoroughly documented, is far from an isolated incident. In January 2025, the FTC released a landmark study on what it formally termed "surveillance pricing," revealing that intermediary companies worked with at least 250 clients — from grocery stores to apparel retailers — to set individualized prices using a staggering range of personal data [5].

The data inputs catalogued by the FTC go far beyond what most consumers would expect. Companies are harvesting precise location data, demographic information, search and browsing history, purchase records, device type, browser settings, language preferences, and even micro-behavioral signals like mouse movements and scroll depth on a webpage [5]. All of this feeds algorithms designed to calculate, in real time, the maximum price each individual consumer is likely to accept.

The food delivery sector is particularly susceptible to these practices because of its structure. A $30 restaurant order that arrives at your door can carry a delivery fee, a service fee, a "small order" fee, a "busy area" fee, regulatory response fees, and menu markups that restaurants apply to offset the 15% to 30% commission they pay to platforms — none of which are presented as a single transparent total [7]. With DoorDash commanding 56% of the U.S. market, Uber Eats at 23%, and Grubhub at 16%, these platforms collectively process tens of billions of dollars in orders annually in an industry projected to exceed $74 billion by 2033 [8].

Research from George Washington University added another dimension to the concern. A large-scale analysis of Chicago-area ride-hailing data — drawn from tens of millions of rides and cross-referenced with U.S. Census Bureau demographic data — found that ethnicity, age, housing prices, and education levels all influenced dynamic fare pricing [9]. Passengers picked up or dropped off in lower-income communities or neighborhoods with higher minority populations were charged more per mile. While the study focused on ride-sharing, the underlying algorithms and the companies deploying them overlap significantly with food delivery operations — Uber Eats and Uber ride-hailing share the same parent company and technological infrastructure.

"When machine learning is applied to social data, the algorithms learn the statistical regularities of the historical injustices and social biases embedded in these data sets," wrote researchers Aylin Caliskan and Akshat Pandey [9].

How the Pricing Machine Works

The mechanics of surveillance pricing in food delivery operate on multiple layers. The most visible is surge or dynamic pricing — the practice of raising delivery fees during peak demand periods, bad weather, or when driver supply is low. Platforms have long justified this as a market-balancing mechanism that incentivizes more drivers to accept orders.

But beneath the surge layer lies a more opaque system. Platforms collect and analyze user behavior continuously: how often you order, what time of day, whether you comparison-shop, how quickly you complete a purchase, whether you abandon carts, and your history of accepting or rejecting suggested items. This behavioral profile feeds machine learning models that optimize not for the lowest possible price, but for the highest price each user segment will tolerate without switching platforms or abandoning an order.

The fee structures themselves are deliberately complex. An analysis by Cybernews found that on Postmates, hidden fees — including service fees, delivery fees, and expected tips — constituted approximately 46.4% of the average order value, the highest among major platforms [7]. DoorDash layers at least five distinct fee types on every order, with some percentage-based, some flat-rate, and at least one invisible to the consumer [7]. Restaurants compound this by marking up menu prices on delivery platforms to offset commissions — a $12.80 pad thai in the restaurant might appear as $16 on the app, with no indication that the markup exists [7].

Source: Consumer Reports Bill Tracker

Data as of Mar 15, 2026CSV

The Regulatory Reckoning

The regulatory response has been unprecedented in speed and scope. New York became the first state to enact an Algorithmic Pricing Disclosure Act, signed by Governor Kathy Hochul in May 2025 and taking effect in November of that year [10]. The law requires any business using consumer-specific data to set personalized prices to display a clear disclosure: "THIS PRICE WAS SET BY AN ALGORITHM USING YOUR PERSONAL DATA" [10]. Companies that fail to comply face civil penalties of up to $1,000 per violation.

By early 2026, DoorDash and Uber were among the first apps to display these mandated disclosures, triggering a new wave of scrutiny as consumers began seeing the warnings in practice [11]. New York Attorney General Letitia James issued guidance specifying that algorithmic inputs triggering the law's disclosure requirement include a consumer's location, income level, and previous shopping habits [10].

The legislative momentum has been remarkable. In the first seven months of 2025 alone, state legislators across 24 states introduced 51 bills targeting algorithmic pricing — a fivefold increase from just 10 bills in all of 2024 [12]. By the end of the year, more than 100 price transparency bills had been introduced across 33 states and Washington, D.C. [12]. For 2026, new proposals have emerged in Arizona, Florida, Hawaii, Illinois, Kentucky, Nebraska, Oklahoma, Pennsylvania, Tennessee, Vermont, Virginia, and Washington [12].

At the federal level, the Stop AI Price Gouging and Wage Fixing Act was introduced in Congress in 2025, targeting the use of algorithms to inflate prices [13]. California Attorney General Rob Bonta launched an investigative sweep in January 2026 focused specifically on surveillance pricing, sending demand letters to businesses with significant online presence in the retail, grocery, and hotel sectors [14]. New York's attorney general sent a separate enforcement letter directly to Instacart regarding its pricing practices [15].

The legislative approaches converge around three regulatory targets: prohibiting algorithmic price-fixing (where competing companies use shared algorithms or data to coordinate prices), banning surveillance pricing (using personal data to charge individuals more), and restricting dynamic pricing in essential sectors like groceries and transportation [12].

The Industry Defense

The food delivery industry has pushed back against characterizations of its pricing as discriminatory. Platforms argue that dynamic pricing serves legitimate functions: matching supply with demand, ensuring driver availability during peak periods, and sustaining a business model that provides convenience to consumers while offering flexible employment to millions of delivery workers.

DoorDash, Uber Eats, and Grubhub have all invested in subscription programs — DashPass, Uber One, and Grubhub+ — that offer reduced or eliminated fees for a monthly charge, which the companies frame as consumer-friendly alternatives. Grubhub went further in its competitive battle with DoorDash, eliminating all delivery and service fees on restaurant orders over $50, saving customers an average of $13 per transaction [16].

The platforms also point to New York City's complex fee cap negotiations as evidence that pricing regulation can have unintended consequences. In April 2025, DoorDash, Grubhub, and Uber Eats settled their legal challenge to the city's fee caps, resulting in a revised structure allowing total restaurant fees of up to 43% — nearly double the original 23% cap — broken into a 15% cap for core delivery, 5% for marketing, 3% for credit card processing, and an additional 20% for enhanced services [17].

What Consumers Can Do

Consumer advocates recommend several strategies for navigating the algorithmic pricing landscape. Ordering directly from restaurants — either through their own websites or by phone — typically avoids the platform markup and algorithmic fee layering entirely. Comparing prices across multiple apps before ordering can reveal significant disparities; research has found pricing gaps of up to 28% between platforms for the same order depending on timing, location, and order size [7].

Using browser privacy modes or VPNs can limit the behavioral data available to pricing algorithms, though platforms that require login can still access purchase history and account-level data. Subscription programs may offset fees for frequent users, but their value depends heavily on order frequency and size.

Perhaps most importantly, consumers in New York — and soon potentially in a dozen more states — can now see when an algorithm is setting their price. That transparency, advocates argue, is the first step toward accountability.

The Bigger Picture

The surveillance pricing debate in food delivery is a bellwether for a much larger transformation in how consumer goods and services are priced across the economy. The same AI-powered pricing tools being deployed in grocery and food delivery are spreading to airlines, hotels, retail, entertainment, and healthcare. The FTC's study found that the intermediary companies enabling personalized pricing work across all of these sectors simultaneously [5].

The food delivery industry's $74 billion trajectory means these pricing practices touch an enormous and growing share of American food spending. The Bureau of Labor Statistics' Consumer Price Index for food away from home has climbed steadily — from an index value of 364.5 in March 2024 to 391.7 in February 2026, a 7.5% increase over less than two years [18]. While algorithmic pricing is not the sole driver of that increase, it operates on top of already rising food costs, compounding the burden on consumers least equipped to absorb it.

The question now is whether the regulatory response — from New York's disclosure law to the FTC's enforcement actions to the wave of state bills — can keep pace with the technology. The algorithms are getting more sophisticated. The data collection is getting more granular. And every time you open a delivery app, the price you see is increasingly a price calculated just for you.