Why a Firm Might Abandon the Clever Price
The firm-level subject is the decision to adopt tracking and personalized pricing when consumers may react to the pricing consequence. The specific question is why a technically available way to estimate willingness to pay can become unattractive when buyers expect to receive a bad deal; understanding that distinction separates a technology problem from a commitment problem.
From the buyer’s side, the decision was whether to enter a data-dependent exchange. From the firm’s side, the decision is whether collecting and using the data will create enough value after accounting for the participation it may lose. The two decisions are connected, but they are not the same. A consumer can dislike a personalized price without the firm immediately abandoning the system. A large enough pattern of consumer reluctance, however, can change the firm’s incentive to adopt it.
The Symptom Is Not a Technical Failure
Suppose a firm can use personal information to make more informed offers. The technology may work exactly as designed: it can collect signals, estimate a buyer’s willingness to pay, and vary the price. Yet the firm may still not use it. The reason can be that the people who dislike being priced this way are also the people whose participation makes the market valuable.
This is the shift in the unit of analysis. The consumer asks, “Will this offer leave me better off?” The firm asks, “Will the additional pricing information raise returns more than the response it provokes?” The first is an individual participation decision. The second is an adoption decision. Treating the second as a simple extension of the first hides the feedback between them.
A 2026 economic model examines that feedback in an online market with two differentiated goods, privacy-sensitive consumers, and alternative monopoly and duopoly structures. Its reported result is qualitative but important: when the proportion of privacy-sensitive consumers is high, no firm uses personalized pricing in equilibrium. When that proportion is low enough, adoption can be supported. The model makes privacy sensitivity part of the demand environment facing the firm rather than an issue outside the firm’s calculation.
The word high matters here. The available evidence is the publisher-supplied abstract, not the full equations or a parameter table. It does not provide a numerical cutoff. The result should therefore be read as a conditional model finding, not as an estimate of how many people in a real population are privacy-sensitive.
The model’s logic can be understood without a numerical threshold. If only a small share of buyers object to tracking, a firm may gain more from better price estimates than it loses from refusals. If a large share object, the firm faces a participation constraint: the pricing method must leave enough buyers willing to transact for adoption to pay. A system that extracts more from each participating buyer can still reduce the number of participating buyers. At some point, the extra extraction is not worth the lost participation.
A simple hypothetical makes the feedback visible. Imagine 100 potential buyers. A personalized system would let the firm earn more from each person who accepts it, but 60 buyers refuse because they expect their information to be used against them.