Guest data: the strategic asset hotels are losing.

Hotels obsess over commission and underweight data. The numbers explain why: commission is a clean P&L line, deducted from each booking, easy to total at month-end. Data loss is diffuse — a slow erosion of guest knowledge that never appears as a single charge on a single statement.

But the data loss compounds, and over a decade it can cost more than the commission ever did.

What "guest data" actually means in 2026

The phrase is used loosely. To make the strategic argument concrete, it helps to distinguish four layers:

• Identity: Real name, real email, real phone. Not an alias-relay address that bounces marketing back to the intermediary.
• Behaviour: What they searched, what they considered, when they booked, how they paid. The patterns that predict the next stay.
• Preference: Room type, floor, dietary needs, special occasions, communication preferences. The operational details that drive repeat satisfaction.
• Cross-property context: Where else they've stayed in your market, what categories of property they prefer, their travel frequency. The strategic context that informs segmentation.

A hotel that owns all four layers can market intelligently, personalise operationally, and forecast accurately. A hotel that owns only one or two of them is running on guesses.

What OTAs capture vs what reaches the hotel

The mismatch is structural. When a guest books a hotel through an OTA, the OTA — by design — sits between guest and hotel for the entire lifecycle of the relationship.

The result: the OTA acquires a marketing-addressable customer. The hotel acquires an operating data point. The asymmetry holds for the lifetime of the guest's travel behaviour.

The compounding effect

The asymmetry compounds in three ways that are easy to miss in any single year:

1. The re-marketing gap widens

Every OTA booking strengthens the OTA's behavioural model of the guest. The OTA learns when they travel, what they consider, what they convert on, and what re-engagement messages work. The hotel learns none of this. By year five, the OTA can predict the guest's next booking window; the hotel can only hope.

2. Repeat economics deteriorate

The economics of a returning guest are dramatically better than a first-time booking — no commission, lower acquisition cost, higher satisfaction. But repeat bookings require remarketing channels. If marketing channels were never established because the data was never owned, the repeat rate plateaus far below where it could be.

3. Operational personalisation degrades

The preference data captured during one stay can transform the next one — provided it's actually attached to a recognised guest record. When the guest re-books through an OTA, often using a different masked address, the property has no reliable way to connect the new booking to the prior preference profile. The work invested in personalisation evaporates at every channel hop.

How AI bookings change the flow

AI distribution offers a structurally different deal. When a traveller asks a neutral AI assistant to book a room, and the AI uses a verified hotel-direct connection (typically via MCP) to execute the booking, the data flow looks different:

• The booking record is hotel-direct. Identity, contact details, and stay context arrive at the property as a complete record — not a stripped subset.
• Marketing consent can be hotel-direct. The booking flow can ask for permission to communicate with the guest about future stays — without OTA contract clauses forbidding it.
• The guest is recognisable across visits. The same authenticated identity that booked once can book again, with the property able to maintain a continuous record.

This is not automatic. AI booking flows can be built in ways that re-intermediate the data — for example, if the AI assistant itself becomes the relationship layer and the hotel just receives a transaction. The structural opportunity exists, but capturing it requires the hotel to actually integrate, not just be listed.

"Commission is the rent you pay for the booking. Data loss is the equity you give up forever." — Sigtrip Strategic Analysis, 2026

Reclaiming the asset

The shift back to data ownership is incremental, not all-at-once. Three priorities for the next 12 months:

1. Audit what each channel actually delivers

Most hotels know their channel mix by revenue. Few know it by data quality. For each channel — OTA, metasearch, direct, agency, group, and now AI — score what the guest record looks like when it arrives. Real email or masked? Marketing consent transferable or not? Preference data attached or not?

The resulting matrix usually surprises operators. Channels that look comparable on revenue look very different on data ownership.

2. Treat data quality as a channel-mix variable

Once the matrix exists, the strategic question changes. Instead of "which channels drive volume?", it becomes "which channels drive volume and deliver an owned guest record?" The latter is what compounds. Direct bookings and hotel-direct AI bookings sit at the top of that list.

3. Invest in identity continuity

The infrastructure that recognises a returning guest across stays, channels, and contact details is what turns isolated bookings into a relationship. This is a CRM problem the industry has under-invested in for two decades. As AI distribution drives more guests directly, the cost of not having a working identity layer rises.

The strategic frame

The decade ahead will be defined by which hotels treat guest data as an asset to actively accumulate, and which keep treating it as a by-product of bookings. The first group will compound their direct-channel advantage every year. The second group will keep paying intermediaries to stand between them and people who have already stayed at their property.

OTAs understood the value of the data asset in 2005. The hotels that catch up are the ones that act like it's still 2005 — but with a 2026 toolkit.

The bookings come and go. The data, once given up, doesn't come back.