In 1970, George Akerlof submitted a paper to three of the most prestigious economics journals in the world. Two of them rejected it on the grounds that the argument was too simple to be worth publishing. The third, the Quarterly Journal of Economics, ran it anyway. Thirty-one years later, Akerlof collected a Nobel Prize for it.
“The Market for Lemons: Quality Uncertainty and the Market Mechanism” is four pages of accessible prose and two pages of mathematics. Its central insight is not complicated: when one party to a transaction knows something material that the other party does not, the market does not simply become less efficient. It fails. The uninformed party, unable to distinguish quality, stops paying for quality. The informed party, unable to command a premium for it, stops offering it. The market converges on its worst possible equilibrium, and it stays there.
Akerlof was writing about used cars. The mechanism he identified, however, is not confined to used cars. It is a structural condition that emerges wherever information about quality is distributed asymmetrically across a transaction — and it produces recognizable consequences regardless of the industry it inhabits.
The luxury hospitality industry has spent thirty years constructing, one rational decision at a time, a demand problem illuminated by precisely this kind of information failure. It did not happen through incompetence. It did not happen through negligence. It happened because individually rational actors, operating under conditions of asymmetric information, produced a collectively damaging outcome that no individual actor can now reverse alone. The name of that outcome is OTA dependence. The mechanism behind it is a structural variant of the lemons problem — modified in its direction, different in its surface dynamics, but recognizable in its underlying logic. And the industry keeps paying for it, literally, every time it needs to buy more demand.
A careful reader will note that Akerlof’s original model positions the seller as the informed party and the buyer as the uninformed one — the used car dealer knows whether the car is a peach or a lemon; the buyer does not. The hotel-OTA system inverts that assignment in a specific way. Here the hotel is, in the relevant sense, the buyer — of demand — and the OTA is the informed seller who knows substantially more about the quality of the demand it intermediates than the hotel it delivers it to. The inversion modifies the mechanism’s surface form. It does not change the underlying logic: one party to a repeated transaction has systematically more information about what is being exchanged than the other, and uses that informational advantage to extract value at the uninformed party’s expense over time.
Understanding how this happened — and why it cannot be solved from within the game that produced it — requires a precise account of the information structure that governs luxury hotel demand. Not a marketing account. Not a technology account. An information account. Because the problem was never about marketing. It was always about who knew what, when, and what they did with it.
I. The Information Architecture of a Booking
Before there was an OTA, a luxury hotel booking was an information-dense transaction. A guest who contacted a property directly disclosed their identity, their preferences, the composition of their travel party, the occasion, the dates, the budget sensitivity, and dozens of behavioral signals embedded in how they asked questions, what they responded to, and what they ultimately agreed to. The hotel absorbed all of that information. It built a record. It built a relationship. It built, over time, a proprietary understanding of who its guests were, what they valued, and what it would take to bring them back.
This is not romanticism about a simpler era. This is a description of an information asset — one that compounded in value with every transaction and that belonged entirely to the hotel.
The OTA changed that information architecture in ways the industry failed to recognize as structural rather than commercial. When a hotel lists inventory on an OTA, the transaction that follows does not belong to the hotel in any meaningful informational sense. The guest browses an intermediary’s interface, filtered and ranked by criteria the hotel does not set and cannot see. The guest books through an intermediary’s system. The guest’s identity is disclosed to the hotel at the level required to execute the stay and no further. The preference data, the browsing behavior, the comparative consideration set, the price sensitivity signals, the abandonment patterns — all of that remains with the OTA.
The hotel receives a guest. The OTA receives a data point.
This asymmetry is not incidental. It is the OTA’s core business model. The OTA’s value to its equity holders derives precisely from its accumulation of demand-side information at scale — information extracted from millions of transactions that individual hotels conduct but cannot observe at the platform level. The OTA knows which guests cross-shopped your property against your three nearest competitors. It knows at what price point they converted and for which dates. It knows which amenities drove consideration and which drove abandonment. It knows more about your demand curve than your own occupancy reports reveal, because it has been observing the full competitive context of each booking decision for twenty years while you saw only the outcome.
This is the information asymmetry at the core of the problem. The OTA is the informed party. The hotel is the buyer of demand in a market where the intermediary has accumulated, over decades, a depth of intelligence about demand behavior that the hotel cannot reconstruct from its own transaction history. The hotel does not know what it is buying with the precision the intermediary brings to selling it. It knows what it paid.
II. The Prisoner’s Dilemma of OTA Adoption
Game theory is not a metaphor. It is a mathematical framework for analyzing strategic interactions in which the outcome for each participant depends on the choices of others. Applied correctly, it does not describe what happened in luxury hotel distribution in the abstract. It predicts it with precision.
The prisoner’s dilemma is the most studied problem in game theory because it captures, in its simplest form, the mechanism by which individually rational behavior produces collectively irrational outcomes. Two players, unable to coordinate, each choose the strategy that protects their individual downside, and both end up worse off than if they had cooperated. The formal condition for a prisoner’s dilemma requires a specific payoff ordering: the temptation to defect must exceed mutual cooperation, which must exceed mutual defection, which must exceed being the only party to cooperate. When that ordering holds, defection is the dominant strategy regardless of what the other party does.
The OTA adoption decision, as it presented itself to luxury hotel operators across the late 1990s and early 2000s, satisfies that ordering for the conditions that characterized that specific adoption period — when consumer search behavior was migrating to platform interfaces faster than any individual property could sustain direct discovery at equivalent scale, and when the first-mover advantage of early OTA visibility was commercially real. Consider the payoff structure as it appeared to any individual property under those conditions:
If the hotel listed and its competitors did not: incremental demand at commission cost, with first-mover advantage in OTA visibility. The best individual outcome — call it T.
If no hotel listed: no OTA platform achieves critical inventory mass, no consumer habit forms around OTA search, and the direct booking ecosystem survives intact. The best collective outcome — call it R.
If all hotels listed: OTA dependence becomes the industry norm, commission loads rise across the board, consumer behavior migrates to platform search, and the information asymmetry compounds permanently. A collectively worse outcome than the pre-OTA state — call it P.
If the hotel did not list and its competitors did: demand erosion as OTA-browsing travelers, now habituated to platform search, book comparable properties through the channel. The worst individual outcome — call it S.
The dominance ordering T > R > P > S holds under those adoption conditions, with the formal PD structure requiring that the incremental occupancy gain from unilateral listing exceeded the commission cost, that the collective listing cost exceeded zero, and that the demand erosion from unilateral non-listing exceeded that collective cost. Listing is therefore the dominant strategy — not because it produces the best collective outcome, but because it produces the best individual outcome regardless of what competitors do. This is the formal definition of a dominant strategy, and it is also the exact decision calculus that played out, property by property, across luxury hospitality over a decade.
No individual hotel made a mistake. Every individual hotel made the locally rational choice. And the aggregate of locally rational choices produced an industry-wide transfer of demand ownership, information rights, and pricing power to intermediaries who had no interest in the long-term health of the properties whose inventory they were selling.
The cage was not built by the OTAs. It was built by the hotels, one rational decision at a time, until the day it was complete and the door swung shut.
III. Why the Equilibrium Cannot Be Escaped from Within
The standard response to a prisoner’s dilemma framing is to invoke the repeated game. Robert Axelrod demonstrated in his landmark 1984 work, “The Evolution of Cooperation,” that in iterated prisoner’s dilemmas with sufficiently patient players, cooperative strategies — specifically tit-for-tat — tend to outperform pure defection over time. If hotels simply stopped listing on OTAs, or coordinated a collective withdrawal, the equilibrium would shift.
This argument fails in the luxury hospitality context for a reason that is structural rather than behavioral. The hotel-OTA relationship is not a symmetric iterated game between players of comparable capability. It is a platform-mediated interaction in which the information conditions have shifted so substantially over twenty years that the original payoff structure no longer applies to the current players.
The hotel that considers delisting from OTAs in 2026 is not the same player it was in 2002. Its direct demand generation capacity has atrophied. Its guest data is incomplete — shaped, in many cases, by two decades of OTA-mediated acquisition that transferred the identity layer of each transaction to the platform rather than the property. Its relationship with the high-value traveler segment has been intermediated for long enough that the direct channel has become structurally weaker, not merely underinvested.
More precisely: the OTA has spent twenty years accumulating the demand-side information that would be required to rebuild direct relationships, and it has no incentive to return it. The hotel that attempts to defect does so with degraded information assets against an intermediary whose algorithmic demand model becomes more accurate with every additional booking conducted through its platform. The game has not merely continued. It has changed. And the hotel is now playing a structurally weaker version of itself against a structurally stronger version of its counterpart.
The iterated game logic also fails because the players are not symmetric and the strategies are not observable in the way Axelrod’s model requires. Tit-for-tat works when both players can observe each other’s prior moves and calibrate responses accordingly. Hotels cannot observe OTA pricing decisions, demand routing, or algorithmic ranking adjustments with the granularity required for credible retaliation strategies. They can observe their own occupancy and commission costs. That is not sufficient information to play a reciprocal strategy against a platform that controls the information environment.
The industry is not simply failing to cooperate. It is operating in an information condition that makes the cooperative escape structurally more costly than continued participation — which is a different and harder problem.
IV. Demand Quality and the Adverse Selection Dynamic
Akerlof’s most devastating insight was not that bad products drive out good ones. It was that the inability to distinguish between them is sufficient to produce the same outcome. The mechanism is adverse selection: when quality is unobservable, the uninformed party cannot price discriminate between high-quality and low-quality supply. They pay based on expected average quality. High-quality sellers, unable to command a premium, exit the market or withhold their best inventory. The average quality of what remains falls. The cycle continues until the market has converged to its worst available equilibrium.
The analogue in luxury hotel demand requires a precise and operational definition of what “demand quality” means, because imprecision here is where the argument becomes vulnerable.
Demand quality, as used in this analysis, is defined by three measurable variables. First: pre-transaction identity observability — the degree to which the hotel can establish, before the booking occurs, who the guest is, what their travel history looks like, and what their preference profile contains. Second: estimated repeat probability — the likelihood, assessable prior to the first stay, that the guest will return without requiring the same acquisition cost as the initial booking. Third: reacquisition cost delta — the difference in cost between re-engaging this guest directly versus re-acquiring an equivalent guest through the OTA channel on a subsequent occasion. A guest who scores high on all three variables is high-quality demand not because they will necessarily spend more on the first stay, but because they generate compounding value across a relationship lifetime. A guest who scores low on all three is not necessarily a bad guest. They may be extraordinarily wealthy. They may be a perfect fit for the property. But the hotel has no information by which to establish that before the transaction, and therefore no capacity to price, retain, or cultivate the relationship independently of the channel through which it was acquired.
In the language of information economics, this is an adverse selection problem — operating in modified form. The hotel cannot distinguish, prior to acquisition, between demand that will compound and demand that will not. It therefore cannot allocate acquisition investment toward the former or price to reflect the genuine long-term value differential between the two. The OTA, by contrast, has accumulated sufficient behavioral data to make exactly those distinctions across all three dimensions. It knows which guests are high-frequency luxury travelers with low reacquisition cost and high repeat probability, and which are once-a-decade occasion bookers who will return to the OTA for their next search regardless of how well the property served them.
Unlike Akerlof’s original mechanism, the luxury hotel market has not unraveled. Hotels continue to transact at scale. The adverse selection here produces not market collapse but market flattening: the hotel acquires an undifferentiated demand pool at undifferentiated acquisition cost, pays the same price for demand it will never see again as for demand that, properly cultivated, would return for a decade, and calls it a distribution strategy.
The market has not collapsed. It has converged toward a mediocrity that is profitable enough to persist and damaging enough to compound.
V. Signaling, Rarity, and the Algorithmic Cost of Discounting
Luxury is not a product category. It is a signaling system. Michael Spence, who shared the Nobel Prize with Akerlof in 2001, developed the theory of market signaling to explain how informed parties communicate quality credibly to uninformed parties when direct observation is impossible. In Spence’s framework, a credible signal must be sufficiently costly that low-quality parties cannot profitably mimic it, while remaining worthwhile for high-quality parties to sustain.
In luxury hospitality, the price signal operates through a mechanism that Spence’s framework illuminates but does not fully contain. Price does not merely signal the hotel’s quality to the traveler. It signals the composition of the guest set — and the composition of the guest set is itself a component of the luxury experience. A rate of $2,000 per night communicates something to the high-net-worth traveler that a rate of $400 does not, and that communication is not primarily about physical amenity. It is about who else the hotel has selected as worthy of that rate. Luxury is, at its informational core, a coordination game in which the signal price influences the composition of the guest set, and the composition of the guest set shapes whether the signal remains credible.
When a luxury hotel discounts through an OTA promotional window, the signal weakens in two compounding ways. First: the discounted rate attracts guests whose price sensitivity is, by revealed preference, materially higher than the guest set the rack rate was designed to select, changing the coordination game for guests who paid full rate. Second, and less visibly: the discounted booking is registered as a successful conversion event in the OTA’s demand model, adjusting the platform’s price sensitivity estimate for that property. Signaling theory does not prove that this adjustment is permanent and irreversible — that claim would require evidence beyond the theoretical framework. What it does establish is that a signal repeatedly available at a lower cost loses its separating power. A luxury hotel that has established a pattern of promotional availability through OTA channels has not temporarily lowered its signal. It has produced a data record — in the most comprehensive demand model in the category — of what its true demand floor looks like under pressure. Removing the promotion does not remove the data.
Every promotional booking communicates to the OTA’s system what the hotel’s demand floor is under pressure. Every communication improves the OTA’s model of that hotel’s demand curve. Every improvement in that model increases the OTA’s leverage in the next contract cycle.
You are, in effect, paying the intermediary to build a more accurate model of your own vulnerability.
VI. The Schelling Point of Mediocrity
Thomas Schelling, whose work on focal points in game theory earned him the Nobel Prize in 2005, identified a phenomenon essential to understanding why the luxury hotel industry has been strategically stagnant despite decades of visible damage from OTA dependence. In coordination games — situations where players benefit from making the same choice as others — outcomes tend to converge on salient focal points even without explicit communication. The focal point is selected not because it is optimal but because it is the choice that every player expects every other player to make.
OTA participation functions as the luxury hospitality industry’s focal point in this sense. It is the equilibrium that every hotel operator knows every other hotel operator has already selected, which makes departing from it a unilateral act rather than a coordinated one. Its salience is reinforced by two decades of institutional normalization — revenue management systems calibrated to OTA demand signals, sales cultures trained on OTA metrics, ownership reporting structures anchored to OTA-visible occupancy data, and brand standards that have accommodated rather than resisted OTA dependency.
The consequence of this focal point is not merely that hotels list on OTAs. It is that hotels have organized their entire demand generation apparatus around the OTA channel — not because that apparatus is optimal but because it is the one every other player in the market is also running. Departing from it unilaterally means abandoning the competitive field before a replacement field has been established. That is not irrationality. That is the logic of a coordination trap reinforced by network effects and twenty years of institutional embedding.
This is analytically distinct from the prisoner’s dilemma framing, and the distinction matters. The prisoner’s dilemma explains why OTA dependence was adopted: individual rationality in a strategic environment with first-mover dynamics produced a collectively damaging dominant strategy. The coordination logic explains why it persists after the damage is visible and widely understood: the cost of collective departure is prohibitive, and the cost of unilateral departure is immediate. No individual hotel has sufficient incentive to move first. No collective mechanism has emerged to move them together. The equilibrium persists not because it is good but because the game required to escape it has not yet been played.
VII. What OTAs Actually Do — and Why That Makes the Problem Harder
Before describing the structural escape, intellectual honesty requires a direct engagement with something the preceding argument has treated too lightly: OTAs are not purely extractive. They solve real economic problems, and those solutions have genuine value to both consumers and, in specific conditions, to hotels.
For consumers, OTAs dramatically reduce search and comparison costs. A traveler who might otherwise contact a dozen properties individually to compare availability, rates, and amenities can accomplish that comparison in minutes on a single platform. The review aggregation that OTAs provide offers a degree of trust mediation that independent hotel websites cannot replicate at equivalent scale. For smaller independent luxury properties with limited direct distribution reach, OTA participation genuinely expanded addressable demand during the platform adoption period — demand that, without the channel, would not have been accessible at all. These are real economic contributions. Dismissing them makes the argument ideological rather than analytical.
The problem is not that OTAs created value. The problem is the terms on which they created it, and what those terms have compounded into over time. A channel that initially expanded addressable demand while collecting a reasonable distribution fee has evolved into something structurally different: a platform whose information advantage over the hotels it serves grows with every transaction, whose commission leverage increases with every year of deepening dependency, and whose equity value reflects, in part, the future commissions of properties that have no realistic path to unilateral exit.
The OTA’s genuine value creation and its structural extraction are not mutually exclusive. They are the same mechanism at different time horizons. The distribution fee was real at year one. The information rent is what it has become by year twenty. A hotel executive who understands both simultaneously is better equipped than one who acknowledges only the extraction or only the value. The argument this article makes requires holding both.
VIII. The Boundary Conditions of This Argument
Before proceeding to the structural escape, intellectual rigor requires stating explicitly what this argument does and does not claim. Overreach on any of these points would weaken the case more than conceding them openly.
This article does not claim that OTAs create no value. Section VII addresses this directly. They reduce consumer search costs, aggregate trust signals, and expanded addressable demand for independent properties during the adoption period. That value was real.
This article does not claim that all luxury hotels are equally trapped. The argument applies most forcefully to independent luxury properties and branded properties in competitive markets whose acquisition history is substantially OTA-mediated. It applies less forcefully to properties with decades of relationship-driven direct acquisition, deep uncontaminated loyalty data, and brand gravity strong enough to sustain direct discovery without platform assistance.
This article does not claim that Spence’s signaling model proves discounting produces permanent, irreversible brand damage. It claims that repeated promotional availability through OTA channels creates a persistent data record in the platform’s demand model that adjusts how the hotel is priced and positioned algorithmically. Whether that constitutes “permanent” damage depends on conditions this article does not attempt to model with precision.
This article does not claim that no hotel can reduce OTA dependence through self-directed effort. Some can, and some have. It claims that for the majority of OTA-heavy properties, self-implementation of a genuinely clean demand signal — one assembled from sources independent of their own OTA-contaminated acquisition history — is rarely achievable at sufficient speed and scale through internal means alone.
This article does not claim that owned demand infrastructure is non-extractive merely by self-description. Section X specifies three structural conditions that must hold for an infrastructure provider to function differently from an OTA. Whether any specific provider satisfies those conditions is a question of contractual terms and data rights, not marketing language.
What this article does claim is narrower and more defensible: that the hotel-OTA relationship is characterized by an information asymmetry that compounds over time, that this asymmetry produces a modified adverse selection dynamic in demand acquisition, that the equilibrium it generates is stable and resistant to within-game optimization, and that the structural escape requires a change in the information conditions of demand acquisition rather than better play within the existing information environment.
IX. The Structural Escape — and Its Real Conditions
The formal solution to a prisoner’s dilemma is not better individual play within the existing game. It is a change to the payoff structure of the game itself. Every proposed solution to OTA dependence that operates within the existing information environment — direct booking incentives, loyalty programs, rate parity negotiations, metasearch investment, digital marketing optimization — is an attempt to play the existing game better. None of them change the game. None of them resolve the information asymmetry. They are responses to the symptoms of a structural condition that they leave intact.
What changes the game is a change in the information conditions under which the hotel acquires demand. Specifically: if the hotel can access demand whose quality is observable prior to the transaction — demand whose identity, preferences, travel behavior, and relationship potential are known before the booking occurs — the payoff structure of the acquisition decision changes. The hotel is no longer selecting blind from an undifferentiated demand pool. It is not paying a commission to an intermediary for demand intelligence the intermediary has accumulated at the hotel’s expense. It is transacting in a different information environment with a different set of available strategies and a different equilibrium.
This is the structural function of Owned Demand Infrastructure. Not a marketing channel. Not a CRM strategy. A mechanism that resolves the information asymmetry before the acquisition transaction occurs, and in doing so changes which strategies are available to the hotel and which outcomes are achievable.
Before describing how that mechanism works, it is necessary to address the sharpest objection a sophisticated reader will raise — because the argument’s critics were right to raise it, and it deserves a structural answer rather than rhetorical evasion.
X. The Middleman Paradox — and Why It Does Not Apply
If the OTA is destructive precisely because it sits between the hotel and its demand, accumulating information at the hotel’s expense, then a sophisticated reader will ask: how does a third-party demand infrastructure provider differ from an OTA in any structurally meaningful sense? If the solution is a different entity sitting between the hotel and its guests, has the game changed, or has the intermediary merely been rebranded?
This is the sharpest objection the article faces. It requires a structural answer — one that specifies not just the direction of information flow in principle, but the precise conditions by which that flow is governed.
The distinction between an OTA and an owned demand infrastructure provider is not a difference of degree. It is a difference in three specific structural conditions that together determine whether an intermediary extracts or transfers. These conditions are not met by assertion. They are met by contractual terms, data rights, and operational architecture — or they are not met at all, in which case the provider is functioning as an OTA regardless of how it describes itself.
Condition One: Pre-Transaction Information Transfer
In the OTA model, the hotel receives guest identity only after the booking is confirmed, at the level required to execute the stay. No behavioral profile, no comparative shopping history, no pre-qualification data transfers to the hotel at any point. The information generated by the guest’s pre-booking behavior belongs entirely to the platform. In an owned demand infrastructure model, the hotel receives the guest’s identity, preference profile, and qualification status before the booking occurs. The informational value of the relationship is transferred to the hotel at the point of introduction, not withheld as a precondition for continued access.
Condition Two: Post-Transaction Data Ownership
In the OTA model, the booking data is retained by the platform and used to refine its model of the hotel’s demand curve. The hotel’s transaction funds the OTA’s intelligence. In an owned demand infrastructure model, the guest record generated by the stay belongs to the hotel. The hotel can re-engage that guest directly, without returning to the infrastructure provider for permission or paying a repeat commission for access to a relationship it has already established.
Condition Three: Compounding Asset Location
In the OTA model, scale benefits the platform: more hotels generate more data, improving the platform’s demand model and increasing its leverage across all hotels. The platform becomes more powerful with every transaction. In an owned demand infrastructure model, scale benefits the hotel: each direct relationship established compounds into a retention asset that reduces future acquisition cost without increasing the infrastructure provider’s leverage over any individual property.
A provider that fails any one of these conditions is functioning as an OTA under a different name. A provider that satisfies all three is functioning as infrastructure: its continued relevance to the hotel depends on the quality of the demand it introduces, not on the depth of the dependency it has engineered. Whether any specific provider satisfies these conditions is determined by the contractual terms governing data rights, not by the language used to describe the service.
XI. Why the Hotel Cannot Self-Implement the Escape
The middleman paradox addressed, a second objection remains. If the structural escape is a pre-qualified demand audience assembled outside the hotel’s own transaction history, why cannot a sufficiently sophisticated hotel build that audience independently?
The answer is structural for the majority of OTA-dependent properties, and it derives from the same information problem that produced the original condition.
The value of a demand audience is not a function of database size. It is a function of the quality of the demand signal — specifically, the degree to which the audience has been assembled from sources independent of the OTA-mediated transaction history that constitutes most luxury hotels’ existing guest data. A hotel that builds a first-party demand audience from its own guest records is not solving the information problem for those records. The guests in that database were, in many cases, acquired through OTA channels in the first place. Their profiles carry the information deficits of the channel through which they were initially acquired. Inspecting that data more carefully produces a more detailed record of a signal that was limited at its origin.
This argument applies most directly to independent luxury properties and branded properties in competitive markets where OTA-mediated acquisition has been the primary growth channel for the better part of two decades. It applies less forcefully to the smaller set of properties with decades of genuinely direct, relationship-driven guest acquisition and loyalty data assembled outside OTA channels. Those properties face a different version of the problem.
For the majority, a genuinely qualified demand signal requires assembly from sources that predate and are independent of the hotel’s own acquisition history — a cross-property, cross-market audience whose travel behavior, spend capacity, and preference profiles have been observed and verified outside any single property’s commercial context. For most OTA-dependent properties, achieving that at sufficient speed and scale through internal means alone is rarely feasible: it requires cross-market scale, an accumulation period that extends beyond typical planning horizons, and independence from the very acquisition channels whose limitations the audience is designed to replace.
The hotel that attempts internal self-repair is not wrong to try. It is working with a signal whose limitations are embedded in its origin. The escape requires a different source of information, not a more thorough analysis of the existing one.
XII. The Tax on Informational Dependence
It is worth being precise about what the OTA commission actually is, because the industry’s language for it has functioned as consequential misdirection — not deliberately, but with real structural effects.
A commission framed as a distribution fee implies the hotel is purchasing reach — access to a demand pool it could not otherwise address. This framing was partially accurate for early OTA transactions, when the platforms were genuinely expanding the addressable market for properties with limited direct distribution reach. It has become progressively less accurate with every subsequent transaction, for a reason that is structural rather than contractual.
Each OTA transaction generates data that improves the OTA’s model of the hotel’s demand curve. A more accurate model means a more precise understanding of the hotel’s demand floor — the minimum rate at which the property will convert guests during distressed periods. A more precise demand floor means more accurate leverage in future pricing negotiations and algorithmic ranking decisions. The hotel’s commission payments, in aggregate, are not merely purchasing bookings. They are funding the continued development of the most accurate external model of the hotel’s demand vulnerability that exists.
By the time a hotel has operated through OTA channels for a decade, the commission is not primarily a distribution fee. It is a tax on informational dependence — a recurring payment for continued access to demand the hotel’s own transaction history funded and the OTA’s platform captured. The hotel’s guests generated the OTA’s data. The hotel now rents a portion of that data back, in the form of bookings, at a rate set by the party with the more complete information about the hotel’s demand curve.
This is the information rent that flows to the platform — levied on the asymmetry between what the OTA knows about the hotel’s demand and what the hotel knows about its own. The hotel is not paying for distribution. It is paying to remain in the information condition that makes it necessary to keep paying.
The OTA’s equity value reflects, in part, a claim on the future commissions of hotels that will continue to fund the platform’s intelligence advantage. Every new hotel that joins the platform adds data. Every additional data point makes the platform’s model more accurate. Every improvement in accuracy extends the information rent. The OTA’s growth is not merely commercial expansion. It is compounding informational leverage.
Conclusion: The Game Has to Change
Akerlof’s lemons paper ends not with despair but with a specific observation: markets develop institutions in response to quality uncertainty — warranties, licensing, certification, brand standards — each of which represents an attempt to resolve the information asymmetry that would otherwise cause the market to fail. The institutions are imperfect. They are partial solutions. But they change the information conditions sufficiently to allow higher-quality transactions to occur.
The luxury hotel industry is at the same institutional inflection point. The mechanisms that governed demand acquisition for three decades have produced a stable, damaging equilibrium. The prisoner’s dilemma of OTA adoption locked the category into a collective outcome that no individual actor can reverse through optimization alone. The coordination logic of OTA participation has made collective departure practically impossible without an alternative mechanism. The information asymmetry has compounded to the point where the hotel’s own transaction data is insufficient to reconstruct the demand intelligence it transferred to the platform over twenty years of mediated acquisition.
The industry recognizes the damage. What it has not yet recognized with sufficient precision is that the damage is not primarily commercial. It is structural. It lives in the information architecture of how demand is acquired, not in the marketing tactics deployed once demand arrives. Every solution that operates within the existing information environment — however well-funded, however technically sophisticated — leaves the architecture intact and the asymmetry compounding.
The game changes when the information conditions change. When the hotel accesses demand whose quality is observable before the transaction, rather than after. When the data generated by the booking accumulates to the hotel, rather than to the intermediary. When the compounding asset belongs to the property, rather than to the platform that delivered the guest.
That is not a marketing proposition. It is the institutional response to an information problem — the same kind of response that Akerlof identified as the genuine solution to the lemons problem in markets where it has taken hold.
George Akerlof identified the mechanism in 1970. The luxury hotel industry built a recognizable illustration of its consequences by 2010. The question in 2026 is not whether the diagnosis is directionally correct. The question is whether the industry has the structural imagination to understand that you cannot solve an information problem with a marketing budget.
The market for lemons does not fix itself. It waits for someone to change the information conditions. And it collects the tax on every transaction until they do.
Americas Great Resorts operates a proprietary audience of pre-qualified affluent travelers assembled independently of OTA-mediated transaction history. Owned Demand Infrastructure is not a marketing service. It is the institutional mechanism by which the information conditions of luxury hotel demand acquisition change — and with them, the direction in which the compounding asset of each transaction flows.
