July Occupancy Rebounded. Why Is Hotel F&B Revenue Still Under Pressure?

July Occupancy Rebounded. Why Is Hotel F&B Revenue Still Under Pressure?

In July, the latest accommodation consumption index published by the China Hotel Association showed a split worth examining: the occupancy index climbed sharply month over month, the online booking index rose from a year earlier, while the hotel F&B revenue index remained well below its baseline, down nearly 20 percent year on year in the official measure. When owners and general managers see "occupancy rebounding," it is tempting to conclude that hotel performance has turned a corner and to shift attention back to rate and room volume, parking F&B as a later problem.

The real question is not whether to keep running a restaurant. It is which guest segment, which meal period, and which conversion path the decline is concentrated in. An industry index only raises the question; it cannot substitute for property-level attribution.

Put the two data sets side by side.

The China Hotel Association's accommodation consumption index uses May 2025 as its base of 100 and draws on the association's survey of about 30 hotel groups and a panel of member properties. In July the composite index was 97.3. The occupancy index recovered month over month but remained below the prior year; the average daily rate index also recovered month over month but declined year on year; the online booking index grew from a year earlier. The hotel F&B revenue index stood at 68.4, up month over month but down about 17.9 percent year on year. In the same period, the National Bureau of Statistics reported national restaurant revenue of RMB456.7 billion in July, up 1.4 percent from a year earlier.

These two sets of data differ in population, sample and definition. One is an index of association member properties; the other is a macroeconomic restaurant revenue figure. An index point is not a revenue amount, and an occupancy recovery is not proof that property F&B has turned. Together they point to a question worth verifying: whether the improvement in lodging-side indicators has actually converted into variable demand for hotel F&B. Where the break is, in demand, product, reach, delivery or contribution structure, can only be judged from a hotel's own records.

First layer: allocate F&B revenue into four guest-account buckets.

The most common mistake is to discuss restaurant revenue as one undifferentiated total and then blame "trading down" or "competition from restaurants outside the property." That closes the inquiry. A more useful approach is to allocate each revenue line to one primary purchasing-decision account, keeping the buckets mutually exclusive:

In-house individual guests, covering breakfast, light meals, in-room dining and convenience demand from late arrivals;

Meeting, group and banquet guests, covering meeting meals, group meals, weddings and event support;

Local destination diners, meaning nearby guests who come specifically for the food rather than an overnight stay;

Non-staying channel guests, from delivery, retail, space partnerships or channel-driven consumption.

Rate-inclusive breakfast, family packages, long-stay and resort packages are packaging or reach methods, not guest accounts. For each account, answer four questions: who comes, why, how they learned about it, and what verifiable value the spending leaves for rooms or meetings. This is how a decline in F&B revenue gets localized to a specific customer base.

Second layer: locate whether the problem is in product, reach or delivery.

Once the demand-side denominator is clear, examine where the supply-side break sits. A product break means the menu, price, portion or timing does not match the real scenario, or too many low-frequency items are kept for the sake of completeness. A reach break means the booking page, confirmation message, front-desk script, in-room information or sales process fails to explain the dining option clearly. A delivery break means breakfast peaks, late-night demand, meeting turnaround, output stability or staffing schedules are misaligned.

One discipline cannot be skipped: do not substitute promotion for diagnosis. If the product is wrong, discounting only amplifies low-contribution orders. If reach is broken, launching new items changes nothing because no one learns about them.

Third layer: read contribution by meal period, not total revenue.

Build a simple contribution table for each meal period, recording guest count and source, check average and revenue, direct food cost and incremental labor, energy and waste, any verifiable effect on rooms or meetings, and whether the period should be kept, fixed, outsourced or pared back.

Fix the simplest definition: period contribution equals net revenue, minus direct variable food cost, minus the labor, energy, platform and fulfillment costs added by that period. Fixed labor, shared energy and property depreciation are not allocated by ad hoc rules. Strategic value, such as breakfast's effect on rate or meeting meals' effect on deal closure, is tracked separately using indicators like take-up of rate-inclusive breakfast, meeting meal conversion, local repeat rate, F&B-related reviews or complaints, and incremental package revenue.

Four decisions: fix, pare back, replace, or stop.

Fix when demand exists, unit contribution can improve, and the main break can be resolved by the current team. Pare back when a genuine need or verifiable strategic value exists but the menu, hours or configuration exceed real demand. Replace when demand exists, the property lacks capability, and the partner's net contribution, food safety, revenue split and experience accountability can be controlled. Stop when contribution has been negative across several comparable periods with no verifiable strategic value for rooms or meetings.

The authorization boundary must be explicit: replacement or shutdown is approved by the owner or the appropriate decision level. A general manager does not change a major operating model on the basis of a fourteen-day test alone.

A fourteen-day, single-variable, low-cost test.

Before testing, review the last 28 days in PMS, POS and meeting orders. Select one high-frequency meal period, one target guest segment and one clear break; do not test several periods at once. Keep operations unchanged for days 1 to 7 and record a baseline under one consistent definition. From days 8 to 14, change only one low-cost variable, such as pre-arrival information, check-in confirmation, front-desk script, in-room information or meeting-sales wording, without new construction, equipment or large spend. Compare the same day-of-week structure, using a defined denominator, on reach, dining conversion, check average, incremental contribution, waste, complaints and service time.

Incremental contribution is incremental revenue minus variable food cost, incremental labor, and any new channel or material cost, with sunk fixed costs listed separately. Define continue, adjust and stop conditions before the test, and record disruptions such as large groups, holidays, temporary closure or price changes. The hard stop: if incremental contribution deteriorates, or complaints or food-safety risk rise, restore the original action immediately; if the single primary metric, such as dining conversion, shows no improvement for three consecutive comparable observation days, stop adding spend and return to the guest-account and break diagnostics.

Responsibility splits as follows: the general manager sets the test object and decision rules, the F&B lead records delivery, finance tallies contribution, and front office and sales record reach and source. The fourteen-day window is for screening one break into the next round of repair and one action to pause or cancel. It does not prove long-term causation and does not justify closing a whole meal period on its own.

Scope and counterexamples.

Resort, all-inclusive, banquet and events hotels should weight their own primary scenarios rather than copy urban limited-service hotels. Low-frequency or strongly seasonal business such as weddings, events and holiday packages needs a longer observation window or comparable-event analysis. Breakfast or meeting support with strategic value should not be cut simply because its standalone margin is low, but that strategic value must be verifiable. A limited-service hotel without an F&B base should not expand on the back of a single industry headline. Use the industry index to raise the question; base the decision on the property's own last-28-day history, the fourteen-day test, and subsequent comparable periods.

Today, stop the argument about "whether to do F&B." Have the general manager, F&B, finance, front office and sales choose one meal period and start the fourteen-day record. Check reach and conversion after a week, contribution and complaints after two, then decide fix, pare back, replace or stop.

MarvelBros C&T examines the operating breaks between hotel revenue, cost, service and organization. A drop in F&B revenue is rarely explained by a soft environment alone; it usually sits in one guest segment, one meal period, and one conversion path that has not been completed.

References: China Hotel Association, China Accommodation Consumption Index Report (July 2026); National Bureau of Statistics, Retail Sales of Consumer Goods, January-July 2026.