The Industry That Got Played Twice

While you were processing the last mistake, they started building the next one.

The governance failure is documented. Accept it as premise. This article begins where it ends.

It ended by asking what, if anything, exists outside the compromised stack. Nobody answered. Nobody asked. And while the industry stood there processing what had been done to it, the next version of the same structural transfer was assembled in plain sight.

That is what this article is about.

The First Time Was Ignorance

A platform showed up. It filled rooms. Nobody modeled what twenty years of compounding data transfer would do to the asset’s pricing power or direct booking equity. The mechanism was new. The consequences were not yet legible.

That was ignorance. Ignorance is forgivable.

This time the mechanism is visible. The outcome is documented. The industry is not operating from ignorance.

This time it is negligence.

What Is Being Built Right Now

Search is familiar territory. A traveler forms a query. Results appear. Your property can compete in that environment. You can optimize for it. You can buy your way into it. You can measure it.

This is not that game.

AI recommendation does not wait for a query. Before the traveler types anything, before they open a browser, a preference model has already shaped which properties are worth surfacing and which are not. That model was not built on your content or your website. It was built on behavioral signal accumulated over twenty years by the platforms that sat between you and your guests: repeat booking sequences, look-to-book ratios across price bands, cancellation behavior, transaction history across properties and markets.

You know who accumulated that signal. You paid them to accumulate it with every commission check you wrote.

The OTA did not merely distribute your inventory. It weaponized your own guests’ behavior, building a demand intelligence model using data you generated and financed, and that model now exceeds anything you hold internally. That accumulated behavioral record is now a structural input to the AI preference systems being built on top of the travel stack. Properties that do not exist in that signal base face structural exclusion from the consideration set before a traveler forms any conscious preference.

Not ranked low.

Not present.

You paid to build the model. The model is now shaping whether you exist.

The Commission Was the Cheap Part. Again.

The industry spent twenty years focused on the commission percentage. 18%. 22%. 25%. That was the problem as ownership understood it. Fix the number on the invoice.

The commission was never the problem. The commission was the fee charged for the actual transfer: the guest relationship, the behavioral record, the right to define your property’s value before you said a word. Ownership focused on the invoice. The platform kept the asset.

Now it is happening again at a new layer. The industry is running AI personalization pilots. It is sitting in panels where speakers explain that AI will redefine travel and then return to properties generating negligible first-party signal outside the OTA stack.

The pilots are the cheap part. Again.

What is actually being ceded this time is the pre-discovery layer. The moment before the traveler searches. Before they compare. Before they have any awareness of your property at all. That moment is being shaped by preference models right now. The shaping reflects the behavioral data that exists. The behavioral data that exists lives at Expedia and Booking.com.

Not at your hotel.

The industry is optimizing conversion while the consideration set is being assembled without it.

The Window Is Not a Metaphor

These models do not start over. They do not give late entrants equal footing. They update on new data with weight given to established patterns, which means the advantage held by the platform with the deepest behavioral record compounds with every passing quarter. A property generating owned demand signal outside the OTA stack, first-party transaction history and direct audience engagement that does not flow through a platform with opposing commercial interests, is building structural presence in that model. A property generating no meaningful signal outside the OTA stack is being encoded as peripheral. Peripheral becomes absent faster than the industry has ever moved to correct a structural problem.

Every quarter spent running pilots instead of building that signal is a quarter the OTA’s behavioral advantage compounds further into the models shaping the next generation of travel discovery. The same mechanism that turned manageable OTA booking share in 2001 into a structural dependency by 2010 is operating at the AI preference layer right now. The timeline is shorter because these models update faster than market share moved.

The window is open. It is measurable. It is closing.

What the Industry Is Doing Instead

Committees. Conferences. Carefully worded AI readiness documents that will be obsolete before they are approved.

This is the same posture the industry held in 2001 when OTA booking share was still small enough that the structural transfer seemed manageable. The technology was maturing. There was time to watch and wait. The technology matured. The window closed. The exit disappeared.

The industry is now watching an identical mechanism operate in real time with full visibility into what the outcome looks like, and it is responding with workshops.

This is not caution. Caution requires uncertainty. There is no uncertainty here. The mechanism is documented. The outcome is on record. The industry is not being careful. It is hiding behind process to avoid the decision that process cannot make.

Every week spent waiting is a week the OTA’s behavioral data compounds further into the models that will define where affluent travelers go next. Waiting is not a neutral position. It is the decision to let someone else’s data define your property’s place in the next discovery layer while you finalize the agenda for the next AI strategy session.

The Only Response That Is Structurally Correct

You cannot optimize your way into a preference model being shaped by OTA behavioral data. Paid social does not generate first-party signal at the pre-discovery layer. Your CRM holds guests you already have. This problem is about the traveler who does not know your property exists yet and is about to have an AI shape a preference before they ever search.

The only mechanism that operates at that layer is owned demand infrastructure. An audience that exists before the traveler searches. Behavioral signal generated outside the OTA stack at the level of introduction and preference formation, not at the level of conversion and transaction. First-party engagement data that reflects your actual demand pool, accumulated over time, in a system the OTA does not own and cannot absorb.

This is not a campaign. It is not a technology pilot. It is the same category of asset the OTA built between 1995 and 2010 while the hotel industry was focused on the commission percentage. The OTA understood the audience was the asset. The hotel industry is repeating the failure while the next version of it is being decided.

If you do not own the signal, you will not exist in the system that defines where affluent travelers go next.

The Verdict

The first time, the mechanism was invisible and the cost was not legible until it was too late to reverse. That is the only available defense. It was spent twenty years ago.

This time the mechanism is visible. The cost is documented. The timeline is measurable. This is not an analogy. It is the same structure operating at a new layer.

The first time was ignorance.

This time it is a choice.


Close