The hotel is full. What does it leave?

1. Tourism growth is a starting point, not a hotel result

In its final 2025 release, published on 7 May 2026, the Bank of Greece reported travel receipts rising 9.4%, inbound travellers rising 6.4% and overnight stays rising 1.6% against 2024. These are national travel indicators, not a hotel margin statement.

For a commercial team, the useful next step is to reconcile the market story with the property’s own bookings. Did its rate improve for comparable rooms and dates? Did distribution become more expensive? Did a different mix of stays change servicing costs? Did the campaign acquire bookings that would otherwise have arrived through another route? A national total cannot answer these questions.

We use the final annual release here as a documented baseline. It is separate from Tourism Strategy’s analysis of the first half of 2026. Annual and first half series must not be spliced together as if they describe the same period. Travel spending also extends beyond accommodation, so spending per visitor night must not be treated as a hotel’s average daily room rate.

2. A country is a market dimension, not a profitability verdict

A hotel may reasonably ask whether to invest more in the United States, China or another source market. The weak answer is a league table of national spending averages. The stronger answer identifies which reachable segment fits the property, which dates need demand and what those completed bookings actually contribute. A business traveller, a family, a group and an independent holidaymaker can have very different economics within the same source market.

Country level travel datasets are useful for forming hypotheses and deciding where to investigate. They do not measure the acquisition cost, room mix, ancillary margin or cancellation profile of your own property. This analysis therefore does not assign an assumed value to US or Chinese guests. The worked examples use fictional segments A, B and C, with every input exposed.

3. Start with the right denominator

Occupancy is occupied room nights divided by available room nights. ADR is room revenue divided by occupied room nights. RevPAR is room revenue divided by available room nights. Each describes something important, but none deducts all the costs needed for this comparison. Two hotels can report the same occupancy and RevPAR while retaining different contribution because their distribution, acquisition and delivery costs differ.

Consider an illustrative 50 room hotel over 30 days. It has 1,500 available room nights. At 90% occupancy it sells 1,350 room nights. An ADR of €150 produces €202,500 of room revenue and RevPAR of €135. If contribution averages €80 per occupied room night after the variable and acquisition costs defined here, it produces €108,000 before fixed overhead and tax. The revenue number and the contribution number answer different questions.

4. Build a contribution bridge before comparing guests

Extra contribution means the revenue retained after the variable cost of the extra service. A restaurant bill is not all contribution. Nor should a transfer booked with an outside supplier be counted at its full selling price if most of it is paid away. Keep the boundary consistent across segments. If you already include breakfast cost in servicing, do not deduct the same breakfast cost again.

Separate costs that change with an occupied night from costs triggered once by a stay. Laundry, amenities and a checkout clean may follow different cost patterns. Use your own operating data. Fixed salaries, rent, financing and depreciation are outside the simplified calculators unless an explicitly incremental cost is added to the relevant input. A positive result is therefore contribution, not net profit.

This broader commercial perspective is consistent with Cornell’s 2017 research on total hotel revenue management, which discussed customer value, non room revenue and the importance of distribution and variable costs. It is a useful conceptual foundation, not a contemporary benchmark or evidence of a particular hotel’s performance.

5. Three segments, with the economics exposed

The following segments are invented to demonstrate the calculation. They do not represent a country, an industry average or a One9Six client result. All three use the same accounting boundaries. The model assumes one room per booking. Rename the segments and replace the figures in the calculator with comparable completed booking data.

Segment B contributes more per booking than A, but uses three times as many room nights. Its contribution per room night is lower. Segment C is higher on both measures in this example, yet that does not tell us how many additional C bookings are available, on which dates, at what marginal cost, or whether the property can fulfil the proposition consistently. The ranking identifies a commercial hypothesis. It does not produce a media allocation automatically.

6. Compare the mix at the same use of capacity

Consider a fictional ten room hotel over thirty nights, fully booked in both scenarios: 300 room nights. Initially, 150 nights come from two night stays, 90 from six night stays and 60 from higher rate three night stays. The second scenario uses 90, 90 and 120 nights respectively. Using the same rates and costs, the amount remaining rises from €26,284 to €30,578, a difference of €4,294. The hotel sells the same number of room nights.

This is not a prediction that the hotel can replace A with C. It is a way to estimate the value of investigating that possibility. In practice, the market may not supply enough C demand, acquisition cost may rise, or longer lead times may leave gaps. Shares are room night shares, not shares of guests or bookings. A segment’s booking share must be converted using its stay length before it can be used in this comparison.

7. Turn contribution into an acquisition ceiling

With an illustrative target of €70 per occupied room night, the acquisition ceilings for A, B and C are €31.80, €32.40 and €274.40 per completed booking. The current costs are €22, €35 and €45. B misses the target by €2.60 per booking under the stated inputs. C has substantial modelled headroom, but headroom is an upper comparison bound, not permission to bid up to that amount.

The average historical CPA can hide the cost of the next booking. A segment may have been cheap to acquire because it contained repeat guests, branded searches or an unusually strong week. The next increment in spend can reach a more expensive audience. Review marginal completed bookings and marginal contribution wherever a credible comparison is available, rather than multiplying the historical average by a much larger budget.

8. Direct is a channel, not an automatic margin advantage

Use the same illustrative €300 completed booking and €90 servicing cost. An OTA charge of 18% leaves €156. A direct booking with €25 acquisition cost and a 2% payment fee leaves €179. Direct is ahead by €23. It matches the OTA contribution when direct acquisition reaches €48, assuming all other values stay unchanged. If acquisition is €60, direct leaves €144 and is behind the OTA in this model.

Count agency, attributable production, paid media and relevant tools consistently in acquisition cost. Use effective channel charges that reflect the contract’s actual charging base. Do not assume every commission is calculated on the same tax or service base. Avoid counting the same commission twice. Then ask whether the direct booking was incremental or shifted from an existing route. Both may matter commercially, but they are different effects.

9. Put your own booking economics into the tools

The tools start with the fictional examples above. They run locally in your browser, require no account and submit no entries to One9Six. Export the outputs before leaving. Neither calculator predicts demand, establishes incrementality or optimises prices automatically. They make assumptions visible so a team can compare them and decide what to test.

10. Cancellations change the denominator

Suppose a campaign spends €3,000 on media and generates 100 initial bookings. Twenty cancel and none produce retained cancellation fees. Media cost is €30 per initial booking but €37.50 per completed booking. If attributable production and agency costs add €2,000, total acquisition cost becomes €62.50 per completed booking. Comparing a €30 media CPA with a completed booking contribution ceiling would mix definitions.

Timing matters as much as the arithmetic. A recently acquired cohort may not have reached its stay dates. Comparing it with a mature cohort can manufacture an improvement or deterioration. Establish a maturity rule, keep booking date and stay date separate, and record any retained fees or recovered room sales explicitly. Use an agreed cohort allocation for costs. Do not quietly move costs into another month to improve the result.

11. Build one commercial record from several systems

The PMS establishes whether a stay occurred and what was retained. The booking engine describes the reservation. Analytics observes part of the digital journey. The cost ledger records what was spent. Choose which system governs each field, document the join and reconcile discrepancies. Treat unattributed bookings as a visible category rather than forcing them into the campaign you want to look successful.

Download the segment data template

Google documents purchase and refund events linked through transaction_id. Its web stream guidance also explains purchase deduplication with a unique transaction ID. These are implementation mechanisms, not proof that a particular hotel’s tracking is correct. Validate test bookings, refunds and booking engine transitions against the internal record before using the numbers for budget decisions.

The planning tools need aggregated segment values, not guest names, email addresses or individual profiles. Keep any operational joins inside the hotel’s authorised systems and share an anonymised summary for the commercial discussion. A practical analysis can begin with a small set of consistent fields. It does not need an uncontrolled export of every customer record.

12. Give bidding a meaningful value signal

Google’s Maximize conversion value guidance allows the advertiser to define value, including revenue or profit margin, and requires conversion tracking with transaction specific values. The important implementation decision is therefore what the supplied value means. It should not be a convenient number chosen because it makes a target ROAS look attractive.

Our recommendation is to keep financial meanings separate. Preserve revenue as revenue in the commercial record. If a contribution signal is used for bidding, define it, version it and test its delivery as a distinct value. Be careful about deducting the very media cost the bidding system is deciding, which creates a circular input. A pre acquisition contribution signal and a separate acquisition constraint are easier to interpret than an unexplained net number.

Do not send one conversion twice under competing primary actions or replace transaction revenue with margin without documenting the downstream consequences. Confirm the booking engine integration, consent handling, conversion lag, refunds and account eligibility before changing bidding. This feature proposes a commercial framework. It does not prescribe a universal bid strategy for every hotel or account.

13. Attribution is not incremental demand

An attributed booking is assigned to a route under a measurement rule. An incremental booking is one that would not have happened without the intervention. A campaign can claim credit for demand that already intended to book. Another may influence a stay that later appears through an unattributed route. The first reporting problem is consistency. The separate causal question is what changes because the advertising exists.

Google describes Conversion Lift as a controlled comparison of exposed and control groups to assess incremental conversions. Availability and usable precision depend on the account and study conditions. For a hotel with limited volume, a test may remain inconclusive. A simple before and after comparison also cannot isolate advertising from prices, events, weather, flight access or changing availability.

Use a control where a valid design is feasible. Otherwise present the result as a monitored commercial test with unresolved alternative explanations. Decide the spending limit, observation period, comparison basis and stopping condition before launch. A result that fails to clear those conditions is information to improve the plan, not a reason to rename the KPI after the fact.

14. A booking also occupies dates you cannot sell twice

An average contribution per night can conceal a calendar constraint. A six night stay may include a high demand Saturday that would otherwise sell at a better contribution, together with weaker weekdays. Accepting or rejecting it requires looking at the whole stay and the credible alternatives, not only the average. The relevant comparison is the contribution displaced on the constrained dates, adjusted for the uncertainty that the alternative demand will materialise.

• Compare room type and arrival pattern, not just a monthly market average.
• Look for minimum stay rules that protect genuinely constrained dates, then test their effect on conversion and gaps.
• Treat opening an additional service, room block or season as a possible step in cost, not a free use of existing capacity.
• Keep a diversified demand base where it protects the property against access, timing or channel risks.

15. Make the offer fit the segment before scaling the audience

A more valuable segment is only a commercial opportunity if the property can attract and serve it. The website must answer the questions that delay a booking. Creative should show the specific stay, not repeat generic luxury language. The landing page should make the room, location, access, cancellation terms and next action understandable in the audience’s language. Distribution and front desk teams should recognise the same proposition.

16. A practical sequence for the next commercial decision

1. Reconcile the baseline

   Select a comparable historical cohort. Match retained room revenue, room nights, channel fees and acquisition costs. Record missing fields and decide which comparisons are not yet defensible.

2. Choose one commercial hypothesis

   For example, a specific stay proposition may improve contribution from a segment on dates with available capacity. Define the desired outcome without assuming the campaign will cause it.

3. Set the budget and the stop

   Name the maximum test spend, the contribution target, a review date and conditions that would stop or change the activity. Agree who has authority to pause it.

4. Align message, page and delivery

   Create an offer the hotel can deliver, a landing journey that explains it and measurement that follows the agreed booking definition. Keep the test distinct enough to interpret.

5. Wait for a usable cohort

   Review leading indicators without pretending they are completed value. A short test period may be useful for technical validation yet too short for a commercial conclusion.

6. Expand, revise or stop

   Compare the completed results with the original decision rule. Record uncertainty, operational costs and alternative explanations. Choose the next limited step rather than defending the initial idea.

17. What would change the recommendation?

The apparent best segment can change when acquisition rises, a room category sells out, additional staffing becomes necessary, extra spending fails to materialise or a strong historical cohort proves unrepresentative. A weak average can also hide a valuable subsegment or a useful contribution on dates that would otherwise remain empty. Recalculate with ranges for the variables that matter. Do not add precision to inputs you have not measured.

The visitor mix calculator is deliberately a transparent scenario model. It does not solve unconstrained demand, forecast cancellations, model multi room bookings or optimise the hotel’s date level inventory. Use it to identify where better evidence would change a decision. A revenue manager, the operating team and marketing should then review the same assumptions before spending changes.