The OTA commission is the largest line item in most hotels' distribution P&L. It's also the one operators have been quietly told, for 20 years, is non-negotiable. The bargain was simple: OTAs deliver reach you cannot build yourself, and in exchange they take a percentage of every booking they touch.

That bargain is starting to break down — not because OTAs are weakening, but because the reach side of the trade is no longer something only OTAs can provide.

The deal hotels actually signed

The OTA value proposition has three components, and it's worth being explicit about each:

  • Demand aggregation: Travellers comparison-shop on the OTA, so the OTA gets to be the discovery layer.
  • Marketing infrastructure: The OTA spends on Google ads, SEO, brand campaigns, retargeting — far more than any individual hotel could afford.
  • Conversion machinery: Polished booking flow, multi-currency, multi-language, customer support, refund handling.

In exchange, the OTA takes 15–25% of the room rate, depending on the model, the market, and any "preferred partner" upcharges. For an independent or small group, that's been the price of being visible.

The compounding cost

A 100-room boutique property running an ADR of €200 and 70% occupancy through OTAs takes in roughly €5.1M in OTA-mediated revenue per year. At a blended 18% commission, that's €920k handed over annually — not for one stay, but every year, forever, for as long as the dependency persists.

Over a decade, that's nearly €10M, much of it consumed by margin that never returns to the property.

What changed

The OTA bargain held because no alternative actually delivered the reach. Direct-booking campaigns clawed back single percentage points but never restructured the channel mix. Metasearch (Trivago, Google Hotel Ads) shifted some discovery upstream but mostly re-fed the OTAs. Loyalty programmes worked for large chains, not for independents.

Three things shifted, almost simultaneously:

1. Discovery moved into AI assistants

Travellers asking "find me a hotel in Lisbon" inside ChatGPT, Claude, or Google AI are now a meaningful slice of top-of-funnel travel queries. The answer the AI gives is the discovery. The OTA isn't necessarily in the loop.

2. Booking moved into the same conversation

Agentic AI — assistants that can execute, not just chat — closes the loop. The discovery and the booking happen in the same interface, often without the user visiting any third-party site. The traditional OTA funnel is bypassed entirely.

3. The protocol layer matured

MCP and equivalent standards let a hotel publish a verified inventory feed that AI platforms can consume directly. No intermediary. No revenue share. The connection is hotel-to-AI-to-guest.

The new maths

Take the 100-room property from earlier. Their P&L assumed OTA commission as a fixed cost. With AI distribution as a working channel, the model changes:

  • Year 1: 10% of OTA-eligible bookings shift to direct AI channels. €510k revenue, ~0% commission. Net retained margin: ~€90k.
  • Year 2: 25% shift as AI booking habits scale. ~€230k retained.
  • Year 3: 40% shift in mature markets. ~€370k retained.

These numbers are illustrative — the actual shift rate depends on the hotel's visibility, the AI platform's defaults, and the broader pace of AI booking adoption. But the directional point is unambiguous: the channel mix is no longer fixed. The OTA share is now a strategic variable.

"Commission was a fixed cost when there was no alternative. The moment there is one, it becomes a choice — and choices get audited." — Sigtrip Strategic Analysis, 2026

What "breaking the cycle" actually looks like

This is not a "kill the OTA" argument. Most properties will continue running OTAs as a meaningful channel — they still deliver real volume, especially for distressed inventory and unfamiliar markets. The point is to stop treating OTA dependency as immutable and start actively managing it down.

Three practical moves:

  1. Establish hotel-direct AI presence. Verified MCP endpoint, structured data, machine-readable rates. Without this, AI assistants will default to whichever OTA structured their data first — and the OTAs are sprinting.
  2. Audit channel mix every quarter, not every year. The AI channel will grow fast or slowly depending on dozens of factors outside any single hotel's control. Quarterly visibility is the minimum cadence to spot the shift and reallocate spend.
  3. Move marketing budget upstream. If OTAs no longer monopolise discovery, the marketing investments that used to be "spend on Google to drive direct bookings on our website" can shift to "make sure neutral AI assistants quote us accurately". Same outcome, different channel.

The strategic frame

The commission cycle hasn't been broken because hotels are willing to pay it. It's been broken because no one offered hotels a credible reach alternative. AI distribution is the first credible alternative in two decades.

That doesn't mean the cycle ends overnight. It means the trajectory is now negotiable for the first time since the early 2000s. Hotels that recognise this — and start treating distribution mix as something to actively shape rather than passively accept — will compound the margin advantage over the next decade.

The commission is still being paid. The question is for how much longer, and on how much of your inventory.