<style type="text/css"> @page { size: 21cm 29.7cm; margin-left: 3.18cm; margin-right: 3.18cm; margin-top: 2.54cm; margin-bottom: 2.54cm } p { direction: ltr; widows: 0; orphans: 0; margin-bottom: 0.25cm; text-align: justify; line-height: 115%; background: transparent } a:link { color: #000080; text-decoration: underline } </style>

Introduction

The fixed price tag is so familiar that it is easy to forget how much economic trust it creates. It tells consumers not only what a product costs, but also that others are being offered the same terms. Personalised pricing unsettles this expectation. Yet the more important question is how such capabilities may reshape markets and society more broadly. As The Economist (2026) argues in a wider discussion of AI, the growing focus on cyber-security must be matched by “urgent thinking” about technology’s economic and social consequences. Personalised pricing is one such consequence.

Technology broadly refers to the practical application of knowledge through tools, systems, and methods that shape how economic activity is organised. It should not be treated as value-neutral because the same technology can have different effects depending on who controls it, what information it collects, and how it redistributes power between firms and consumers. More specifically, technology can be understood as mechanisms firms use to observe, classify, predict, and target consumers for pricing purposes. These mechanisms include, but are not limited to, data analytics, tracking systems, algorithms and artificial intelligence. Algorithmic pricing is not identical to personalised pricing, but it has become one of the most powerful means through which personalised pricing can now be carried out.

If personalised pricing becomes widely used, we should not expect a uniform effect. In transparent and contestable markets, it may mainly damage trust once price differences are revealed. In opaque and concentrated markets, it may allow firms to extract more surplus, weaken consumer choice, and reinforce data-based market power. Its strongest effect is institutional, where prices may cease to function as public, comparable signals and instead become private calculations based on behavioural surveillance.

Why Transparency Matters

To understand what is genuinely new about algorithmically enabled personalised pricing, it is first necessary to recognise that charging different consumers different prices is not itself a product of the digital age. Before algorithmic pricing, there was haggling, where the same product might be sold at different prices depending on the buyer’s identity, bargaining power, or urgency (Hower, 1938). Yet this also exposed a fairness problem: Prices reflected not just the value of the good, but also the information about the buyer’s identity, power, or need. In medieval Europe, this concern was reflected in the idea of the Just Price. Aquinas argued that selling a thing for more than it’s worth was “in itself unjust and unlawful,” while also recognising that “the just price of things is not fixed with mathematical precision” (Aquinas, ca. 1274/1947, II–II, q. 77, art. 1). The point was that exchanges should remain within a reasonable range of fairness, rather than allowing sellers to turn a buyer’s urgent need into a source of extra profit. Nineteenth-century retailers such as Le Bon Marché and John Wanamaker later turned fixed, visibly posted prices into a retail innovation, making prices more predictable and easier for consumers to compare (Miller, 1981; Hower, 1938). In doing so, fixed pricing helped institutionalise an expectation that prices should be visible and comparable, rather than privately adjusted to each buyer’s bargaining position.

Algorithmically enabled personalised pricing revives this older economic logic, but transforms it through unprecedented precision, speed, scale, and opacity. The central issue, therefore, is not price differentiation itself, but whether the differentiation is transparent, rule-based and contestable, or opaque, individualised and based on data consumers cannot observe or challenge.

Table 1

Consumer Attitudes Toward Different Forms of Price Differentiation

Item	Net acceptability
Loyalty-card discount	+32
Student discount	+27
Quantity discount	+17
Lower price for people who look poor	-22
Airline raises price when seats almost sold out	-56
Hotel charges Apple users more	-75

Note. Net acceptability equals the share rating a practice very/totally acceptable minus the share rating it very/totally unacceptable. Data from Poort and Zuiderveen Borgesius (2019, N = 1,202), used to compare relative attitudes rather than U.S.-specific views.

As Table 1 suggests, consumers do not reject differentiated prices outright. They are relatively accepting of visible discounts, such as loyalty-card, student, and quantity discounts, but strongly reject forms that appear opaque and exploitative.

Visible Differences, Hidden Pricing

Technology alone does not determine the effects of personalised pricing. More important are the market conditions that shape whether consumers can verify, contest, and resist price differences.

For instance, Instacart provides a relatively basic case of this problem, showing how even small differences in prices for identical goods can undermine price transparency and consumer trust. In a 2025 experiment, Groundwork Collaborative, Consumer Reports, and More Perfect Union asked 437 shoppers across four cities to add the same groceries from the same stores at the same time. Nearly three-quarters of tested items appeared at multiple prices, with some shoppers seeing prices up to 23% higher for the same product and total baskets differing by about 7% on average (Groundwork Collaborative et al., 2025). This does not prove fully personalised pricing since Instacart said the programme was randomised price testing, not dynamic or surveillance pricing (Associated Press, 2025).

Nevertheless, this distinction does not remove the economic effect. Groundwork explicitly argued that when prices are no longer transparent, shoppers cannot comparison-shop, and when prices are no longer predictable, they cannot properly budget (Groundwork Collaborative et al., 2025). The reputational effect was also visible in Instacart’s own response, where the company admitted that the tests left “some people questioning the prices they see on Instacart” and that customers “should never have to second-guess the prices they’re seeing” (Associated Press, 2025). The controversy also became a regulatory issue when New York Attorney General Letitia James demanded more information about Instacart’s algorithmic pricing practices, warning that charging different prices for the same products left shoppers “feeling cheated” and raised serious concerns under New York’s algorithmic pricing disclosure rules (Office of the New York Attorney General, 2026). Therefore, the Instacart case shows that, while in markets for standardised and easily comparable goods, algorithmic price variation may allow firms to test willingness to pay, but once exposed, its wider effect is to make pricing appear less transparent, less trustworthy, and increasingly exposed to regulatory scrutiny.

Yet, the airline industry complicates the Instacart case because price differences are much harder to isolate and challenge. Unlike groceries, air tickets are not standardised goods sold in a transparent market. The final price may vary according to route, departure time, seat availability, fare class, baggage rules, refund conditions, and connecting flights. These complications make it difficult for consumers to know whether a higher fare reflects ordinary demand-based pricing, product differentiation, limited competition, or the use of personal data.

This opacity did not emerge from digital pricing alone: Decades of deregulation, hub-and-spoke networks, loyalty programmes, and consolidation had already made air travel a market in which fares vary widely, where switching is often costly. Digital technology therefore does not create pricing power from nothing; it operates on top of an already complex pricing environment. The recent JetBlue lawsuit shows why this shift is controversial. The complaint alleged that JetBlue failed to disclose that website trackers were used to set prices, conduct behavioural analytics, and share consumer data with third-party analytics and pricing technology providers such as FullStory and PROS Holdings(Bullock et al., 2026) . JetBlue denied that fares were based on cached data or personal information, but the dispute itself reveals the problem: In the airline industry, consumers struggle to distinguish ordinary dynamic pricing from pricing based on behavioural surveillance (Bullock et al., 2026). The airline case therefore pushes the argument beyond Instacart: Where identical groceries made price differences visible and contestable, airfares are easier to justify, harder to compare, and more difficult to resist because personalised pricing can be hidden within a pricing system that already appears complex.

The Data Feedback Loop

The effects of algorithmic pricing do not stop with individual transactions. By rewarding firms that already possess large datasets, advanced pricing systems, and established consumer relationships, it can turn data into a durable source of market power. Firms with large datasets can estimate willingness to pay more accurately than smaller entrants (Steinberg, 2020; Grochowski et al., 2022). Even a modest increase in extracted surplus can then fund further investment in data infrastructure, analytics, and customer retention, creating a feedback loop in which data improves pricing, pricing raises margins, and higher margins reinforce the firm’s data advantage. Its effects are therefore unlikely to be evenly distributed across firms. Grochowski et al. (2022) note that dynamic pricing can widen the gap between algorithmic and non-algorithmic sellers and create winner-takes-all tendencies. In oligopolistic markets, the risk is sharper because if firms rely on similar pricing software or respond continuously to each other’s algorithmic prices, competition may come to resemble tacit coordination rather than independent price-setting. The consumer’s problem is therefore not only that prices become harder to understand, but that, over time, the market offering those prices may become less contestable.

Rethinking Willingness to Pay

The strongest defence of personalised pricing is allocative efficiency: If prices ration scarce goods by directing them towards consumers with the highest willingness to pay, then technology that identifies willingness to pay more precisely might improve market allocation. But this defence depends on what willingness to pay represents. Graeber’s work on value exposes the danger of reducing value to price and preference. The medieval just-price problem makes the point starkly: A starving prisoner may surrender his entire fortune for an egg, but that willingness does not make the bargain fair (Graeber, 2005). What appears economically as intense demand may instead be desperation, dependence, or a lack of meaningful alternatives.

Algorithmically enabled personalised pricing makes this problem newly scalable. It can turn asymmetric information into a pricing tool: Firms infer urgency, loyalty, weak outside options, or vulnerability, while consumers often cannot know why they face a particular price. Its effect is therefore not merely that some consumers pay more than others. It changes the meaning of price itself from a public signal within exchange into a private estimate of how much each consumer can be pressured to pay. In this sense, the consumer is no longer treated as a contracting party responding to a public offer, but as a data profile whose predicted behaviour can be used to manipulate pricing (Steinberg, 2020). This is the moral issue at the centre of personalised pricing: Fairness matters because consumers’ belief that prices are legitimate is one of the conditions under which markets function. When pricing reflects genuine differences in demand, the efficiency defence has force; when it monetises vulnerability, it undermines trust in the fairness of market exchange.

Conclusion

The final judgment should therefore be calibrated rather than absolutist. Personalised pricing may widen access when it is transparent and socially defensible, such as the case with student discounts. However, the danger lies in opaque, data-driven pricing in markets where consumers face weak choice, high switching costs and concentrated corporate power. A blanket ban would remove beneficial price differentiation and would be difficult to enforce. Yet regulation must avoid creating compliance burdens that only the largest firms can absorb. If smaller rivals are deterred by audit and reporting costs, regulation may end up deepening the concentration that makes exploitative personalised pricing dangerous. The most plausible response is targeted regulation: not against price differentiation outright, but against opacity, discriminatory data use and algorithmic extraction where consumers cannot meaningfully compare, contest or exit. If widely used, personalised pricing may produce efficiency gains, but its stronger effect would be institutional, testing whether markets can still be understood as systems of exchange instead of systems of surveillance.

References

Aquinas, T. (1947). Summa theologica (Fathers of the English Dominican Province,

Trans.). Benziger Bros. (Original work published ca. 1274). https://www.newadvent.org/summa/3077.html

Associated Press. (2025, December 22). Instacart ends a program where users could

see different prices for the same item at the same store. AP News. https://apnews.com/article/instacart-pricing-customers-retailers-c9a0a52e959ce46d2152aa664308d228

Bullock, H. S., Lizmi, D. J., Shen, L., Mancall-Bitel, S., Provance, M. D., Demko, A.

J., & Lowry, V. B. (2026, May 11). New class action targets "surveillance pricing": What businesses using personalized pricing need to know. Mayer Brown. https://www.mayerbrown.com/en/insights/publications/2026/05/new-class-action-targets-surveillance-pricing-what-businesses-using-personalized-pricing-need-to-know

Graeber, D. (2005). Value: Anthropological theories of value. In J. G. Carrier (Ed.), A

handbook of economic anthropology (pp. 439–454). Edward Elgar Publishing.

Grochowski, M., Jabłonowska, A., Lagioia, F., & Sartor, G. (2022). Algorithmic price

discrimination and consumer protection: A digital arms race? Technology and Regulation, 2022, 36–47. https://doi.org/10.26116/techreg.2022.004

Groundwork Collaborative, Consumer Reports, & More Perfect Union. (2025,

December 9). Same cart, different price: Instacart's price experiments cost families at checkout. Groundwork Collaborative. https://groundworkcollaborative.org/work/instacart/

Hower, R. M. (1938). Urban retailing 100 years ago. Business History Review, 12(6),

91–101. https://doi.org/10.1017/S0007680500022005

Miller, M. B. (1981). The Bon Marché: Bourgeois culture and the department store,

1869–1920. Princeton University Press.

Office of the New York State Attorney General. (2026, January 8). Attorney General

James demands answers from Instacart about algorithmic pricing. https://ag.ny.gov/press-release/2026/attorney-general-james-demands-answers-instacart-about-algorithmic-pricing

Poort, J., & Zuiderveen Borgesius, F. J. (2019). Does everyone have a price?

Understanding people's attitude towards online and offline price discrimination. Internet Policy Review, 8(1). https://doi.org/10.14763/2019.1.1383

Steinberg, E. (2020). Big data and personalized pricing. Business Ethics Quarterly,

30(1), 97–117. https://doi.org/10.1017/beq.2019.19

The Economist. (2026, April 16). America wakes up to AI's dangerous power.

https://www.economist.com/leaders/2026/04/16/america-wakes-up-to-ais-dangerous-power