Every few years the conversation resurfaces.
A hotel conference panel. A trade publication opinion piece. A LinkedIn thread from someone who just paid their monthly Expedia commission statement. The question is always the same: why didn’t hotels just build their own platform?
It sounds reasonable. Hotels had the inventory. They had the guests. They had thirty years of watching Expedia and Booking.com extract billions in commissions from their revenue. If the problem was that obvious, and the cost was that painful, why didn’t the industry do something about it?
The answer is that a hotel-owned OTA was structurally non-viable under the actual market conditions and incentive structures that existed. Three simultaneous structural conditions made it so, and each one reinforced the others in ways that made partial solutions collapse under real-world conditions.
The Evidence Was Already There
The industry did try.
RoomKey launched in January 2012 as a joint venture among six of the largest hotel companies in the world: Choice Hotels, Hilton, Hyatt, IHG, Marriott, and Wyndham. This was not a fringe experiment or an underfunded pilot. It had founding-brand legitimacy from companies representing hundreds of thousands of hotel rooms globally. It had direct access to hotel inventory. It had experienced leadership. It had an explicit anti-OTA rationale: give travelers a trusted hotel-search experience while routing bookings back to hotel suppliers rather than to intermediaries.
RoomKey had a highly credible version of every ingredient the “we just needed better execution” argument demands. It had the brands. It had the inventory. It had the institutional backing. It had the financial rationale.
RoomKey stopped providing hotel-search functionality in June 2020.
The significance of that failure is not that one company executed poorly. It is that a highly credible version of the attempt, backed by major hotel companies with direct access to supply, still could not become consumer habit. That shifts the burden of proof onto the execution argument rather than the structural one.
Before RoomKey, the airline industry attempted something similar. Orbitz began as an airline-backed venture designed to create a lower-cost, supplier-influenced online travel distribution channel. It achieved significant scale, significant enough that Expedia acquired Orbitz Worldwide in 2015 for approximately $1.6 billion in enterprise value. But that outcome proves something different from what it might appear to prove. Orbitz did not fail because supplier-built platforms cannot reach relevance. It failed to remain what it was built to be. The moment a supplier-owned platform becomes valuable enough to matter, it becomes a target for acquisition by the intermediaries it was built to displace. Supplier-built platforms in this category have not demonstrated an ability to remain neutral and independent once they are worth owning.
These were not execution failures. They were structural outcomes.
The Collective-Action Problem
To understand why, start with an economic condition the hospitality industry has never fully applied to itself.
A hotel-owned OTA would require the largest, most powerful hotel companies in the world to fund, build, and maintain a platform that also made their direct competitors more visible to the same travelers they were competing for. Marriott would contribute capital and inventory to a platform that also prominently featured Hilton. Hilton would fund infrastructure that delivered bookings to Hyatt. Every independent property would benefit from the platform’s existence regardless of what it contributed to building it.
That is the structural trap. The industry could agree that OTA dependence was costly. It could not create a funding model where each participant’s private incentive matched the collective need. The largest brands had the most to contribute and the most to lose. Their own direct-channel economics, their loyalty program leverage, and their negotiated OTA positioning all gave them specific reasons to resist making the shared platform work for everyone equally. Smaller operators and independents would capture the benefits without carrying proportional risk.
There is also a dimension the industry rarely acknowledges publicly. Hotel brands are not monolithic actors who can simply decide to commit inventory and capital. A Marriott franchise owner is not Marriott. An IHG franchisee negotiates independently. The brands nominally at the table could not unilaterally deliver the inventory of their franchise networks. The gap between brand commitment and franchisee participation is not a management problem. It is a structural feature of how most major hotel companies actually operate. The brand signs the agreement. The franchise network does not.
The loyalty conflict compounds this further. The major hotel brands spent the same decade building their own direct-booking machinery: Marriott Bonvoy, Hilton Honors, World of Hyatt. Those programs were not just a parallel effort. They were a competing priority. They also solved a different problem. Loyalty captured known guests after brand preference existed. An OTA captured open-market demand before brand preference was established. Every dollar invested in brand.com and loyalty infrastructure was a dollar not invested in shared platform development. Every guest captured through Hilton Honors was a guest Hilton had specific reason not to route through a neutral platform that also featured Marriott. The brands were not simply slow to cooperate. They were actively building the alternative to cooperation at the same time they were nominally participating in the shared attempt.
The more useful the shared platform became, the more every large participant had reason to underfund it, negotiate for preferred positioning, protect its own channel economics, or simply wait for someone else to bear the cost of building it properly. The collective interest was real. The private incentive to act on it consistently was not.
This pattern appears consistently when direct competitors attempt to build shared infrastructure that allocates the demand they are simultaneously competing for. The private defection incentive often becomes stronger than the collective build incentive when the contributors are also direct competitors for the demand the platform is supposed to distribute.
There is also a structural constraint that sits outside the industry entirely. Six major hotel chains building a joint booking platform would have faced significant regulatory scrutiny. Orbitz itself attracted antitrust review when the airlines built it. A hotel-owned equivalent designed to set distribution terms, control ranking logic, and coordinate across competing chains would likely have faced substantial antitrust scrutiny, forcing the venture to operate under legal constraints that raised coordination costs, slowed decision-making, and limited how aggressively the platform could compete. That regulatory burden would not replace the collective-action problem. It would make the collective-action problem harder to solve.
What Collapse Would Have Looked Like
The failure sequence would not have required incompetence.
Imagine the most optimistic version of the attempt. Marriott, Hilton, IHG, and Hyatt form a joint venture. They commit capital. They agree on a governance framework. They mandate inventory participation from their managed properties. Within the first operating cycle, the ranking disputes begin. Marriott pushes for contribution-weighted visibility. Hilton pushes for review-weighted ranking. IHG’s franchisees begin withholding best-available rates to protect their own direct channels. The governance body convenes. The lawyers arrive. Antitrust counsel advises constraint. The consumer-facing product stalls while the internal negotiation continues. The OTAs, watching this unfold, increase their marketing spend. The platform never reaches the traffic threshold required to change consumer search behavior.
A compromise is eventually reached, but the compromise degrades the product. Large contributors extract advantage inside the rules. Smaller participants reduce commitment as neutrality erodes. Consumers sense the bias and keep using the platforms that already feel complete. The hotel-owned platform stabilizes at a scale meaningful to its participants but invisible to the traveler making a booking decision on a phone.
The joint venture quietly winds down. Not because the executives failed. Because the structure of the problem made that sequence predictable before the first booking was ever processed.
The Governance Problem Has No Neutral Solution
Assume the funding problem was solved and the regulatory risk was managed. Assume every hotel agreed to contribute proportionally and the platform got built.
The next problem has no resolution that all parties would accept at the layer that matters most.
Who controls the ranking algorithm?
On Expedia, a traveler searches for a luxury hotel in Miami and a ranked list appears. The algorithm that produces that list is Expedia’s proprietary system. Expedia owns it and is accountable to travelers and shareholders, not to the hotels competing for position within it.
On a hotel-owned platform, the algorithm that produces that ranked list would have to be controlled by someone. And every hotel on the platform would have a direct financial interest in how it worked. The Ritz-Carlton wants to rank above the Four Seasons. The Four Seasons wants to rank above the St. Regis. Every independent property wants visibility above every competing independent property.
There is no ranking methodology all competing hotels would accept as neutral. Weight the algorithm toward rate and budget properties win. Weight it toward reviews and established brands win. Factor in availability and last-minute inventory floods the top results. Tie visibility to platform investment and the largest contributors capture the demand the platform was built to distribute fairly. Every parameter is a negotiation between parties whose financial interests are directly opposed at the moment of consumer decision.
This creates an adverse selection trap with no workable exit. If the platform governs ranking fairly and neutrally, the strongest brands lose their competitive advantage and have reduced incentive to remain on the platform or fund it adequately. If the platform allows ranking to be influenced by ownership stakes or contribution levels, travelers recognize a biased search environment and the platform loses the consumer trust that makes it useful.
OTAs sustain consumer trust in part by subsidizing the traveler experience. They invest billions in marketing, price transparency, loyalty rewards, and frictionless booking, all funded by commission revenue extracted from the supplier side. A hotel-owned platform governed by suppliers has a structural conflict between subsidizing consumers and protecting the supplier margins that justify the platform’s existence. An independent OTA can spend freely on the demand side because it profits from the supply side. A supplier-owned OTA cannot make that same investment without directly subsidizing its own competitors’ bookings.
Other industries have built functional shared platforms under imperfect governance. They work when the platform is not directly allocating demand among the same parties required to govern it. An MLS governs data visibility and listing access. It does not decide which hotel captures a traveler at the moment of booking. A hotel-owned OTA’s governance problem is not about shared data access. It is about demand allocation, specifically who gets the guest, among the exact parties required to govern the allocation neutrally. Hybrid structures, tiered ranking models, and algorithmic transparency mechanisms can work in other markets when the platform is not directly allocating demand among the same parties required to govern it. In hotel distribution it is. That is the difference that makes those models insufficient here.
No governance structure can make that allocation sufficiently tolerable for every major participant whose revenue depends on the outcome, because the participants with the most revenue at stake are always the ones in the most direct competition at the layer governance most needs to control.
The Network Effects Were Already Gone
Even if the funding problem, the regulatory risk, and the governance problem were all resolved, even if the industry produced a capitalized and adequately governed platform, the window had already closed by the time the industry was ready to act.
OTAs are winner-take-most markets. The platform with the most hotel listings attracts the most travelers. The platform with the most travelers attracts the most hotel listings. Each addition to supply increases platform value to demand. Each addition to demand increases platform value to supply. The two sides compound each other in a loop that becomes progressively harder to enter from the outside.
By the time RoomKey launched in 2012, Booking.com had been building its network for fifteen years. Expedia had been building for sixteen. They had not merely accumulated listings. They had accumulated trust, search engine authority, mobile app installs, loyalty program members, conversion data, and the behavioral default of millions of travelers who had learned to start their hotel search on those platforms.
A new platform entering that environment would not need to gradually build toward relevance. It would need to simultaneously replicate sufficient inventory breadth, consumer trust, search presence, and booking conversion to become the place travelers defaulted to for hotel search, before the OTAs simply outspent it into irrelevance. Those conditions cannot be purchased sequentially. They compound together or they do not compound at all.
The counterargument is that venture-backed platforms have built dominant network effects even against entrenched incumbents. Airbnb is the most relevant example, establishing itself starting in 2008 in the middle of the window this article argues had already closed. But Airbnb was creating a new supply category that did not exist in the incumbent platforms. It was not displacing an established consumer habit with a direct substitute. It was building demand for something travelers could not get from Expedia or Booking.com. A hotel-owned OTA was not creating a new category. It was asking travelers to replace an existing platform habit with a functionally identical one, from a coalition of the suppliers the traveler already knew were competing for their booking. Those are categorically different competitive problems.
A subscale platform does not solve the problem either. Suppliers still need OTA volume and continue supporting the incumbents that deliver it. Consumers prefer completeness and default to the platforms with the broadest inventory. Marketing efficiency collapses below the scale threshold required to shift consumer search behavior. Without full relevance the platform cannot maintain supply discipline or demand preference, and both revert to the incumbent platforms that never stopped operating. Partial success is not a stable equilibrium. It is a slower path to the same outcome.
The obvious workarounds do not escape the structure. Mandatory participation breaks down against franchisee autonomy and brand loyalty priorities, because the brand signs the agreement and the franchise network does not. Contribution-weighted ranking solves the funding problem at the cost of neutrality and trust, turning the platform into a branded directory travelers learn to distrust. Regional or niche pilots solve scale temporarily but cannot attract the participants required for national relevance, because those participants have no private incentive to join a local experiment. Each workaround appears to narrow the problem but reproduces the same contradiction in smaller form: the platform needs cooperation from parties whose private incentives improve when they defect.
The OTAs’ structural advantage by 2012 was not simply capital. It was fifteen years of accumulated consumer behavior that had to be unlearned before the new platform could be learned. That is not a problem that marketing spend solves under competitive pressure from incumbents with both the resources and the incentive to ensure the attempt fails.
What the Question Was Actually Asking
The industry keeps asking why hotels never built their own OTA as if the answer is embarrassing, as if the right leader, the right moment, or the right pooling of resources would have produced a different outcome.
The collective-action problem meant the platform could not be reliably funded without the parties with the most to contribute having the most private reason to underfund it. Franchise structure meant the brands nominally at the table could not deliver what they appeared to control. Loyalty investment meant the most capable participants were actively building against cooperation at the same time they were nominally participating in it. The regulatory environment added coordination costs that raised the threshold for viable collective action further. The governance problem meant the platform could not be neutrally operated among parties whose financial interests were in direct conflict at the layer governance most needed to control, and the adverse selection trap ensured that any attempt to resolve that conflict would alienate either the suppliers or the consumers the platform depended on. The network effects problem meant the platform could not reach the scale required to displace established consumer habit, and subscale success was not a stable resting point but a slower reversion to the same outcome.
These conditions reinforced one another. Weak governance made funding harder to sustain. Inadequate funding prevented the network scale required to make governance conflicts worth tolerating. Lack of scale intensified the ranking conflicts that governance could not resolve. Partial solutions to any one condition collapsed under pressure from the others.
The question was never really about execution. It was always about structure.
The hotels were not outmaneuvered. They were not outspent. They were not the victims of poor leadership or insufficient ambition.
They were operating inside a market structure where the necessary conditions for success contradicted one another. The platform needed collective funding, but the incentives favored defection. It needed neutral governance, but the product itself allocated demand among competitors. It needed network scale, but the incumbents already owned the consumer habit.
That is why the hotel-owned OTA was not a missed opportunity. It was a structurally non-viable answer to the wrong question.
