City Festivals Can Fill Hotels. First Prove Which Room Nights Are Truly Incremental

A citywide event can stimulate demand, but it does not prove incremental room nights at a specific hotel. Before committing inventory or rate, a property should define what would have happened without the partnership, tag the demand it can observe, and measure contribution after displacement and fulfillment costs.

The Shanghai Municipal Government reported on September 7, 2026 that the classic season of the 2026 Shanghai Tourism Festival began on September 1, with more than 170 cultural and tourism activities planned across the city. The programme also included cooperation with hotel groups on products combining accommodation with attractions or sporting events. This is useful city level context. It shows that event supply and hotel linked products are expanding. It does not show that any particular hotel will receive incremental bookings, achieve a higher net room rate or earn additional profit.

That distinction should shape the first management question. Do not begin with whether the hotel should join. Begin with how the hotel will prove which room nights came from the activity, which would have arrived anyway, what they contributed after costs, and what operating capacity they consumed.

Start with the counterfactual

Incremental demand is not the same as the number of bookings carrying an event name. The practical question is how many room nights, and how much contribution, the hotel would have received if the event partnership had not existed.

That counterfactual cannot be observed directly. It must be estimated from comparable periods and documented assumptions. The comparison should be as consistent as possible on the length of the stay window, weekday and weekend mix, room type, rate band, booking channel, normal seasonality, and known conferences or other demand shocks. Two comparable non event windows are more useful than one. If no close comparison exists, management can use the most relevant available period, but it should record the differences instead of hiding them.

The baseline will never create perfect certainty. Its purpose is to prevent management from treating every room night sold during a festival as demand created by that festival.

Build an eight field event incrementality bridge

For each date, room type and channel combination, maintain eight fields. The fields are the counterfactual baseline room nights, event tagged room nights, actual untagged room nights, realized room revenue, incremental commissions and discounts, incremental redemption and labour cost, actual refunds and compensation, and displacement adjusted incremental contribution.

The test period actual room nights equal event tagged room nights plus actual untagged room nights. Observed room night uplift equals actual room nights minus the adjusted counterfactual baseline. Displacement adjusted incremental contribution equals the total contribution during the test minus the contribution expected without the activity and minus activity specific one time costs.

The field definitions matter because an event tag proves traceability, not incrementality.

First, record the counterfactual baseline room nights. This is the room night level the property would reasonably expect without the activity, based on comparable historical windows, booking pace, channel mix and conversion patterns. Keep the assumptions visible. A comparison period affected by a major conference, a closure or a materially different pricing strategy is not a clean comparison.

Second, record event tagged room nights. These are bookings traceable through a dedicated channel code, booking page, package identifier, room type code, redemption record or partner guest list. A tagged booking may still have been made without the activity. A guest who would have booked the hotel directly may simply have selected the promoted package. Do not automatically count every tagged room night as a new room night.

Third, record actual untagged room nights. These are the room nights sold during the test without the activity identifier. If they exceed the adjusted baseline, the excess is an unattributed uplift candidate, not proven event spillover. Wider destination awareness, organic search, repeat guests, competitor pricing and ordinary seasonal demand may all be involved. The property should use booking pace, channel searches, price changes, competitor supply and other city activities to interpret this movement.

Fourth, record realized room revenue. Use the room revenue actually attributable to the booking, not the headline package price. If a package includes tickets, meals, upgrades or other benefits, assign their fair value before evaluating the room contribution. Keep this field as realized revenue. Do not also describe it as revenue after commissions and then deduct the same commissions again.

Fifth, record incremental commissions and discounts. Use the contract, campaign addendum and settlement statement to identify the additional distribution charges, hotel funded promotions, cashback, exclusive discounts and other costs created by the activity. Platform terms vary by property, channel and partnership structure. A single industry percentage is not a substitute for the hotel's own settlement data.

Sixth, record incremental redemption and labour cost. Include ticket verification, concierge assistance, breakfast upgrades, additional housekeeping, front desk handling, transport coordination, temporary staff and peak period food and beverage labour. These costs are often missed because they appear as overtime or operational disruption rather than as a package expense.

Seventh, record actual refunds and compensation. Include cancellations, refunds, date changes, failed redemptions and compensation connected with the activity. Weather, traffic restrictions, security arrangements, ticketing failures and event schedule changes may affect the result. Responsibility for those disruptions should be clear before launch.

Eighth, record displacement adjusted incremental contribution. This is the decision field. It should account for the difference between test period contribution and the no activity baseline, activity specific costs, rate dilution, and the contribution from higher value demand that the activity may have displaced. The result can be negative. More room nights do not necessarily mean more economic value.

Give each field a named owner

A bridge without named accountability becomes another spreadsheet that everyone recognises and no one maintains.

The revenue leader should own the counterfactual baseline, price floor, inventory cap and test window. The marketing leader should own the activity code, dedicated page and partner guest list. The property or reservation system owner should configure the required fields and audit tagging completeness. Finance should independently calculate incremental contribution from the contract, settlement statement and operating records. Front desk, housekeeping and food and beverage leaders should record their own added labour, materials and service exceptions. The general manager should decide at the scheduled review whether to expand, continue within limits, redesign, pause or exit.

The revenue team may define the baseline and control inventory, but it should not be the only function deciding whether its own activity attribution is valid. A cross functional review is a basic control, not an administrative extra.

Run a small and reversible pilot

A 14 day pilot is a useful starting point, not a universal industry rule. The appropriate window depends on booking lead time, demand rhythm and the event structure. A shorter cycle may require seven days. Another property may need 21 days or a window covering a complete event weekend and its cancellation period.

Define three separate windows before launch. The booking observation window should cover the main booking lead time. The stay and fulfilment window should cover the actual event and service delivery. The settlement review window should cover cancellations, refunds and platform settlement. A short pilot cannot prove repeat purchase if the normal repurchase cycle is longer.

Keep the first test controlled.

1. Choose one stay window. Avoid established peak dates and known large conferences where possible. If the activity must run during a traditional peak, use a restricted inventory test and include displaced high value demand in the calculation.

2. Use one channel. An existing direct or semi direct channel usually creates a cleaner test than launching through several new channels at once.

3. Apply one comparison method. Use comparable historical windows with the same weekday mix, rate band and channel. Record unavoidable differences.

4. Set the inventory cap and price floor before bookings open. Do not release every eligible room type at launch.

5. Set decision thresholds before the first booking. The hotel should define the minimum acceptable contribution and the tolerable range for cancellations, complaints, price floor breaches, data completeness and operational disruption from its own baseline.

Watch three supplementary signals

The financial bridge should be accompanied by three operating signals.

The first is the pre arrival cancellation rate, compared with the same type of baseline. A package that books strongly but cancels unpredictably can damage forecasting and late stage inventory recovery.

The second is the post arrival complaint rate. This shows whether the hotel can deliver the package without excessive waiting, redemption disputes or service recovery.

The third is a short term retention signal, such as consented membership registration or a reliably identifiable future booking. Repeat booking is a useful lagging measure, but it should not be required to improve within a 14 day pilot.

These signals should not be applied mechanically. Expansion should depend on the combination of net contribution, cancellation and complaint movement, operational capacity and the quality of the data. A single weak signal requires investigation. A material deterioration beyond the pre agreed tolerance range requires redesign, pause or exit.

A hypothetical example

Consider a hypothetical midscale hotel testing an attraction and two night stay package through one local lifestyle platform. It reserves 20 rooms per night and runs the offer for an initial 14 day pilot.

Comparable windows suggest a counterfactual baseline of 168 room nights. During the pilot, the hotel records 192 event tagged room nights and 38 actual untagged room nights. The figures show where bookings entered the system. They do not prove that the activity created 192 new room nights, or that all 38 untagged room nights were event spillover. Total observed room nights are 230, so the observed uplift over the baseline is 62 before attribution, rate, displacement and cost are considered.

Suppose the cancellation rate is 9.4 percent during the pilot compared with 5.8 percent in the comparison window. That is a useful warning signal, but it is not enough to determine the financial result. The hotel still needs actual room revenue, additional commissions and discounts, redemption and labour cost, refunds and compensation, and the contribution expected under the no activity baseline. Without those amounts, management cannot honestly declare the net contribution positive or negative.

This is precisely why a hypothetical example should demonstrate the calculation structure rather than manufacture a profit result. The numbers above are illustrative only and are not an industry benchmark or a forecast.

Know the conditions that reverse the conclusion

A controlled offer in a low demand period can create genuine value if it fills rooms that would otherwise remain unsold, preserves rate and inventory control, limits costly inclusions and fits within existing service capacity.

The opposite can happen during a high demand period. Discounted event bookings may displace higher rated direct or corporate demand. Occupancy can remain high while total room contribution falls.

Low season inventory is not automatically safe. Expensive redemptions, outsourced cleaning, dedicated transport or intensive guest assistance can consume the value of rooms that appeared inexpensive to fill.

An outcome at one property should not be promoted as an industry rule. Guest mix, channel structure, labour model, pricing position and service capacity differ by hotel.

Set pause and exit triggers before launch

The partnership should pause or exit when the activity source cannot be reliably identified, the partner cannot provide usable guest lists or settlement details, the price floor or inventory cap is breached, or the hotel cannot reconcile the financial fields. The same applies when the settled displacement adjusted contribution does not exceed the no activity baseline, when cancellation, complaint or fulfilment exceptions exceed the property's pre set tolerance range, or when activity related workload remains above available staffing capacity.

These triggers should be based on the hotel's own historical baseline. Example percentages are not universal standards. The statement that one hotel's result cannot be generalized to the industry is a communication boundary, not by itself a project exit trigger.

The manager's checklist

Before approving or expanding an event linked product, the management team should complete the following actions.

1. Name the partner, channel, product format and proposed stay window.
2. Define the counterfactual baseline and document the comparison periods.
3. Create a dedicated channel code, property system tag, package identifier or room type code.
4. Keep event tagged room nights separate from actual untagged room nights.
5. Establish all eight fields of the event incrementality bridge before bookings open.
6. Assign revenue, marketing, systems, finance, operations and general management responsibilities by name.
7. Set the booking observation window, fulfilment window and settlement review window.
8. Set the inventory cap, price floor, minimum contribution and operational tolerance range.
9. Run one controlled pilot through one channel and record the three supplementary signals.
10. Review the result after settlement data is complete, not only when occupancy is visible.
11. Record an explicit decision to expand, continue within limits, redesign, pause or exit.
12. Preserve the assumptions and evidence so the next partnership is judged from data rather than publicity or memory.

City festivals can create demand, visibility and useful hotel partnerships. But the operating decision should rest on room nights the hotel can identify, contribution it can reconcile and delivery it can sustain. A large event calendar is a reason to test a proposition, not proof that a specific hotel has earned the right to expand it.

MarvelBros C&T

MarvelBros C&T