Research overview / Hospitality
One Hotel, Seven Versions
A family can find a hotel with connecting rooms and still need confirmation that those rooms will be allocated for its stay. The words on request change what it can safely assume.
One Hotel, Seven Versions examines how hotel information changes across websites, booking systems, structured data, intermediary platforms and AI answers—and what those differences mean for a traveller’s decision.
- Published
- 2026-07-27
- Updated
- 2026-11-05
Three observations that change the decision
The study compares 16 traveller-critical facts across seven representation roles for Hôtel Mont-Blanc Chamonix. It is an independent public-source case study.
Among the observations recorded in the paper:
| Traveller’s question | What the reviewed material showed | What still needs resolving |
|---|---|---|
| Can our family have connecting rooms? | The hotel’s pages described connecting rooms as available on request. The historical AI report mentioned them without foregrounding that condition. | Who confirms the allocation for this particular stay, and when? |
| Which checkout time should we rely on? | The third-party JSON-LD proposal contained a checkout value that differed from the reviewed official information. | A machine-readable value still needs checking against the appropriate source. |
| Can I move through the stay without steps? | Adapted-room information and broader accessibility labels did not, individually, resolve the complete journey for a particular guest. | The relevant routes and the traveller’s requirements need specific confirmation. |
These are historical observations, documented in Appendix A. They describe the material examined, rather than current guarantees about the property.
The paper calls the distance between what a property can deliver and what its representations allow someone to conclude the Property Truth Gap.
Seven representations, different responsibilities
The comparison covers the owned website, direct booking engine, structured data, Google and local-search surfaces, online travel agencies and reviews, destination and editorial sources, and AI synthesis.
Each contributes something different.
A room page can explain a configuration and its conditions. A booking engine can retrieve date-specific offers. An intermediary provides its own booking and support arrangements. A destination source can add relevant context. An AI answer can combine information from several of them.
The practical requirement is agreement on the facts that affect a decision. Identical wording across every channel would not establish that agreement.
The Omnichannel Decision Mesh describes the connections between those sources and the paths a traveller follows to investigate, compare and act.
Start with one consequential fact
Choose a fact whose wording could change someone’s decision: a room arrangement, an arrival condition, an access route or a cancellation term.
Then follow five steps:
-
Step 1: Align the fact and its owner.
Identify who can establish the information and resolve a disagreement. -
Step 2: Check the claim and its conditions.
Retain distinctions such as available, guaranteed and subject to confirmation. -
Step 3: Review its representations.
Examine the relevant pages, listings, structured data and summaries for altered meaning. -
Step 4: Connect changing information to its responsible system.
Availability, offers and confirmations need an appropriate current source. -
Step 5: Make the next action accountable.
Explain how to obtain confirmation, who handles exceptions and what happens if something goes wrong.
The paper develops this progression through the Source-to-Service Chain, connecting a description with verification, an offer, confirmation, action and recovery.
For help examining one guest promise, the Seven-Version Snapshot explains the scope of the assisted review. The separate Tamaga Hotel Demo illustrates mechanisms in a fictional setting.
Read the findings within their limits
The primary public-source review took place in July 2026. The reproduced Gemini report was generated on 20 December 2025. These materials do not form a simultaneous, current test of every channel.
The structured-data analysis concerns a third-party proposal, whose deployment on the hotel’s website has not been verified. Tamaga’s corrected JSON-LD example is illustrative.
The direct booking engine was not fully observable through the text-only review. That limitation does not establish that it lacks the information. No live, date-specific transaction established whether a direct offer was cheaper, more flexible or available.
The study was not commissioned by the hotel. It does not assess hospitality quality, commercial performance or legal compliance, and one property cannot establish how widespread these issues are across tourism.
The source note explains the sources, dates, evidence labels, limitations and correction route.
Read the full research
The maintained web paper contains the detailed argument, observations, figures, references and subsequent clarifications.
The August 2026 PDF preserves a fixed edition. Cite the version you actually consulted.
Sources
- One Hotel, Seven Versions. Tamaga, 2026.