Earlier articles in this series made two arguments. The Evidence Your Commercial Team Is Not Looking At argued that the demand signal inside a hotel's own calls, emails, and chats is the evidence set most commercial teams never analyze. Reduce OTA Commission by Capturing the Direct Bookings You're Already Losing argued that hotels lose direct demand they already generated and paid for.
This article is about how to measure it.
Most hotels can report how many reservations they received. Far fewer can report how many qualified guests asked about a stay, what those guests needed, how the hotel responded, and why the inquiry did or did not become a booking.
Notice what that means for the metric every commercial team wants: a direct-demand conversion rate. Most hotels cannot report one. Not because the arithmetic is hard, but because they cannot report the denominator. Nothing they own counts the qualified inquiries that arrived, and nothing follows enough of those inquiries to a verified outcome. Before a hotel can judge its conversion, it has to be able to see its demand, which is why the framework below treats coverage, the share of inquiries the hotel can actually follow to a result, as a first-class measure alongside conversion itself.
That missing record is lost demand.
Lost demand is not everyone who visited the website without booking. It is not an estimate of travelers who considered the destination. It is not every unanswered phone ring or spam submission.
It is qualified commercial interest that entered a channel the hotel could observe but did not become a reservation, together with the evidence needed to understand the outcome.
A hotel measures lost demand by starting with qualified inquiries, assigning a consistent outcome, preserving the reason when it is known, and keeping "unknown" visible when it is not.
What counts as lost demand for a hotel?
A useful definition begins with the inquiry rather than the booking.
A qualified direct inquiry is an observable expression of interest connected to a plausible stay, event, group, or other commercial need that the hotel could evaluate.
It might arrive through:
- A phone call.
- An email.
- Web chat.
- A contact or inquiry form.
- A messaging channel.
- A wedding, group, or corporate RFP.
- A direct conversation with reservations or sales.
Not every interaction qualifies.
A current guest requesting towels is a service interaction, not new booking demand. A robocall is not an inquiry. A vague message with no identifiable commercial need may remain unqualified until the hotel has enough context to evaluate it.
A practical qualification rule should require some combination of:
- A property or destination.
- Dates or a usable time frame.
- A room, event, or service requirement.
- A traveler or buyer segment.
- A real availability question, purchase condition, or request for an offer.
The purpose is not to make the denominator as large as possible. It is to count the demand the hotel had a credible opportunity to convert.
Doesn't sales already track this?
Partly, and that is the strongest argument for the rest of the framework.
Group and event sales teams have kept a version of this record for decades. The lost business report is an established artifact of hotel sales operations: when a group RFP is turned down or turned away, the sales and catering system records the opportunity, its value, and a turndown reason. A DOSM reviewing lost business by reason and by account is applying exactly the discipline this article describes.
The gap is everything that never enters that system.
The leisure caller asking about connecting rooms for a school-holiday week, the 11pm chat about accessible rooms, the email that hesitated over a deposit policy: none of these is an RFP, so none of them earns a record. Group sales measures its lost demand because its tooling forces the question. The rest of the hotel's direct demand, which at most properties is the larger share, gets no equivalent.
So the proposal here is not a new idea imported from outside the industry. It is an extension of a practice hotel sales teams already trust, applied to every qualified direct inquiry rather than only the ones large enough to become a lead.
Why do most hotel reports miss this denominator?
Website analytics, booking engines, Hotel Center, Google Ads, the PMS, and the RMS all measure valuable parts of the commercial path.
A website funnel can show sessions, booking-engine entries, and completed reservations. Google Hotel Center, the layer Google uses to hold and validate hotel rates and availability, and which we examined in Google UCP for Lodging: From AI Recommendation to AI Reservation, can report booking-link impressions, clicks, and click-through rate. Google Ads hotel conversion reporting can connect configured campaign activity with booking outcomes.
Those systems make defined paths measurable. They do not automatically create one record of every qualified inquiry across phone, email, chat, forms, RFPs, and channel changes.
A guest might discover the property through search, browse the website, call about an accessibility requirement, receive an answer by email, and book through the engine two days later. The reservation is visible. The inquiry that shaped it may remain separate.
A planner might request a proposal, ask about one contract condition, and stop responding. The lead may exist. The reason the opportunity stalled may remain buried in a thread.
As explained in RMS vs PMS vs Commercial Intelligence, established hotel systems answer different questions. Measuring lost demand does not require diminishing those systems. It requires a common inquiry record that can connect their outcomes to the evidence expressed before a reservation existed.
What should a lost-demand record contain?
The record should be simple enough to apply consistently and detailed enough to support an investigation.
| Field | Question it answers | Example values |
|---|---|---|
| Inquiry identity | Can this interaction be followed without creating duplicate demand? | Contact, lead, conversation, or anonymous inquiry ID |
| Property | Which hotel or destination was considered? | Property, cluster, or market |
| Channel | Where did the inquiry arrive? | Phone, email, chat, form, message, RFP |
| Need | What was the traveler trying to solve? | Room configuration, event, policy, amenity, experience |
| Commercial context | What made the request specific? | Dates, occupancy, room type, segment, event size |
| Response | How did the hotel handle it? | Answered, quoted, escalated, followed up, no response |
| Outcome | What happened? | Booked, pending, unavailable, lost, unqualified, unknown |
| Reason | Why did that outcome occur, when known? | Price, policy, product fit, information, response, operations |
| Timing | When did the interaction and outcome occur? | Inquiry time, response time, stay dates, resolution date |
The hotel does not need every field before the record can exist. It needs a consistent way to improve completeness over time.
An inquiry with an unknown outcome should remain in the denominator if it was qualified. Removing unknowns makes the report look cleaner by hiding the exact measurement gap the hotel is trying to understand.
Which outcomes should hotels track?
Use a small set of mutually understandable outcomes.
| Outcome | What it means |
|---|---|
| Booked | The inquiry became a verified reservation, contract, or other defined commercial conversion |
| Pending | The opportunity remains active; the guest or buyer has not accepted or declined, and the hotel still expects another action |
| Unavailable | The hotel could not offer the required dates, room configuration, space, inventory, or service |
| Lost | The inquiry was qualified and did not convert, and the opportunity is no longer active |
| Unqualified | The interaction did not represent a plausible commercial opportunity under the hotel's agreed qualification rules |
| Unknown | The inquiry was qualified, but the final outcome could not be verified |
These categories should not be collapsed prematurely.
"Unavailable" is different from "lost after an offer." "Unqualified" is different from "unknown." A pending wedding lead should not be counted as a failed conversion because the reporting period ended before the sales cycle did.
Which reasons for lost demand are useful?
A reason taxonomy should help teams make decisions, not create false precision.
Recommended top-level reasons are:
| Reason | What it means |
|---|---|
| Availability | Required dates, inventory, room configuration, or space could not be offered |
| Price or value | The guest rejected the price or did not see sufficient value in the offer |
| Policy | A deposit, cancellation, payment, minimum-stay, age, pet, or other condition blocked the decision |
| Product fit | The property could not meet the requested location, amenity, experience, or configuration |
| Information | The answer was missing, inconsistent, unclear, or difficult to verify |
| Response | The reply was late, incomplete, misrouted, or never sent |
| Operational | A current property reality made the stay or event a poor fit |
| Alternative selected | The guest reported choosing another property, channel, date, or solution |
| Guest withdrew | The traveler's plans changed for a reason unrelated to the hotel where known |
| Unknown | The hotel could not verify why the inquiry did not convert |
Do not infer a reason from silence.
A guest who stops replying after receiving a rate did not necessarily reject the price. They may have changed dates, chosen another property, postponed the trip, or booked through a different channel. Unless the evidence supports a reason, keep it unknown.
How does AI visibility fit into lost-demand measurement?
AI visibility and lost-demand measurement observe different parts of the commercial path.
AI visibility is outside-in. It shows whether an AI platform can find the property, associate it with the traveler's need, describe it accurately, cite useful sources, and provide a path toward the hotel or an intermediary.
Lost-demand measurement is inside-out. It begins once qualified interest reaches a channel the hotel can observe.
This means "not cited by an AI platform" should not automatically become a lost-demand reason. If the traveler never discovered or contacted the property, the hotel did not observe an inquiry to classify. Treating that invisible demand as measured loss would make the denominator speculative.
The two evidence sets become useful when they are compared:
| AI visibility finding | Owned-demand finding | Commercial interpretation |
|---|---|---|
| The hotel is rarely recommended for a relevant need | Direct inquiries repeatedly express that need | The public information environment may not reflect demand the hotel already attracts |
| The hotel is recommended for a need | Few corresponding inquiries reach the property | The association may be weak commercially, misaligned, or occurring through a path the hotel cannot observe |
| AI answers rely on intermediaries rather than the hotel's site | Relevant direct inquiries still arrive | The hotel may have demand but limited control over the public evidence and routing around it |
| The hotel is accurately represented and frequently recommended | Qualified inquiries arrive but conversion remains weak | The problem is more likely downstream, such as availability, policy, product fit, information, or response |
| An AI agent cannot verify property facts, availability, or a direct path | Guests ask the same unresolved questions directly | The hotel has both a machine-readable information gap and an observed conversion-friction signal |
These comparisons do not prove causation. They tell the commercial team where to investigate.
As discussed in AI Visibility Shows Who Mentions Your Hotel. It Cannot Show Who Booked., public platforms can show how a property is understood and routed. The hotel's inquiry record shows whether the same needs appear in real demand and what happened when they did.
The valuable measure is not an AI citation added to the lost-reason taxonomy. It is the gap between how external systems understand the property and what qualified travelers actually ask the hotel to provide.
How should a hotel calculate lost-demand performance?
No single percentage explains the full picture, and the first measure to fix is usually not conversion. It is coverage. A hotel that cannot follow its inquiries to a verified outcome cannot trust any conversion rate it computes, because the losses it cannot see are missing from the math.
Anana recommends beginning with four separate measures.
1. Qualified inquiry volume
How many observable commercial inquiries met the hotel's qualification rule?
This establishes the denominator and should be segmented by property, channel, segment, need, and stay period.
2. Direct-demand conversion
Verified booked inquiries ÷ total qualified direct inquiries
This is broader than a website conversion rate because it includes qualified demand from other direct channels. It should be reported alongside, not instead of, website and booking-engine conversion.
3. Outcome coverage
Qualified inquiries with a verified outcome ÷ total qualified inquiries
A low result does not necessarily mean conversion is poor. It means the hotel cannot reliably follow enough inquiries to their outcome. This is the number that tells a commercial team whether its conversion rate deserves any confidence at all, which is why improving it usually comes before optimizing anything downstream.
4. Known-reason coverage
Lost or unavailable inquiries with a supported reason ÷ lost or unavailable inquiries
This shows how much of the non-converting demand can be investigated without pretending every outcome is explainable.
These terms are a recommended analytical framework, not universal hospitality standards. The definitions should be documented and applied consistently before different properties or periods are compared.
Why should the measures be segmented?
A portfolio-wide average can hide the operating problem.
Phone inquiries should not automatically be compared with wedding RFPs. A request for ten rooms has a different sales cycle from a one-night leisure stay. A guest asking about an unavailable accessible room should not be grouped with a guest who received no follow-up.
Segment the measures by:
- Property and market.
- Channel.
- Traveler or buyer segment.
- Need or request type.
- Stay date or event period.
- Response time.
- Outcome.
- Reason lost.
Then look for the gaps.
A property may have high inquiry volume and low outcome coverage, which means the first problem is follow-through and measurement.
Another may have strong outcome coverage but repeated losses around one policy, which creates a policy or communication question.
Another may convert simple leisure inquiries well but lose a high share of small-group requests after the first proposal, which points toward a sales-workflow investigation.
The average becomes useful only after the underlying demand remains distinguishable.
How should hotels handle cross-channel journeys?
Do not assume the channel where the guest booked is the channel that created or converted the demand.
A caller may book through the website. An email inquiry may become a reservation entered manually. A chat user may return through a branded search campaign. A planner may communicate across a form, email, and phone before signing.
Where identity and consent allow, connect interactions to one inquiry or opportunity record. Use reservation IDs, lead IDs, verified contact details, dates, property, and other appropriate matching evidence.
Where the connection cannot be verified, do not force attribution. Preserve the inquiry and its unknown outcome.
The goal is not perfect attribution. It is a more complete and honest account of the demand the hotel actually observed.
What decisions improve when lost demand is measured?
The framework becomes valuable when it changes a decision.
Revenue can distinguish price resistance from unavailable inventory, product constraints, and weak response handling.
Marketing can see which traveler questions appear repeatedly in direct inquiries but remain unclear on the website or in campaign messaging.
Reservations can identify questions that require better answers, escalation paths, or follow-up.
Sales can see where group, corporate, and event opportunities stall.
Operations can see which property realities affect purchase decisions before the stay.
Commercial leadership can compare who asked with who booked without hiding unknowns or reducing every lost opportunity to price.
This is the practical purpose of measuring lost demand. It turns disconnected interactions into evidence a team can investigate.
Where does Anana fit?
Anana captures and structures direct interactions hotels already receive across calls, emails, chats, leads, and related workflows. That gives commercial teams a way to investigate recurring needs, responses, outcomes, and known reasons demand disappears.
The goal is not to replace website analytics, the booking engine, the PMS, the RMS, or the CRM. Those systems continue to own the records they were built to manage.
The goal is to preserve the commercial evidence that appears before, between, and around those records.
A hotel should know more than how many rooms it sold. It should know how much qualified demand reached the property, what those travelers needed, how the team responded, and where the opportunity disappeared.
Start with who asked. Keep the unknowns. Then connect the outcomes.