Direct Booking Is Not Automatically Cheaper: How to Measure the Net Contribution of Every Booking by Channel

Direct Booking Is Not Automatically Cheaper: How to Measure the Net Contribution of Every Booking by Channel

Your official website charges no OTA commission. So why can a direct booking still end up costing more than an OTA booking?

The answer rarely sits in the room rate. It sits in the fact that booking revenue, acquisition spend, fulfilment cost, cancellation losses, and the cost of credit are rarely placed in the same table. When a hotel compares only commission rates, it sees the channel's price. Only when the net room revenue each channel delivers under the same stay-date structure is aligned against all attributable costs does management see which booking actually leaves cash behind.

The D-EDGE 2026 Europe & Asia hotel distribution report brings together data from more than 2,000 hotels across 2022 to 2025. It shows that direct channels perform strongly on average daily rate, booking lead time, and cancellation rates, and that OTA, GDS, wholesale, and direct each carry a different task across volume, value, and revenue stability. What the report proves is that channel differences exist in an international sample. It cannot prove which channel any single hotel in China should grow or cut. That judgement must return to the property's own PMS, CRS, payment, advertising, membership, cancellation, and receivables data.

Every booking carries four accounts

Start with the revenue that actually lands. Beyond net room rate, confirm how taxes, upsells, and ancillary revenue are attributed, and whether refunds, cancellation losses, and no-shows are deducted from revenue. If a property counts confirmed bookings but leaves a large volume of cancellations in the numerator, channel performance is flattered by mistake.

Then look at what was spent before the booking closed. Commission, affiliate fees, paid search, social media, membership benefits, promotional subsidies, payment fees, and technology fees all belong in the same period. Group advertising, website redesigns, or brand membership programmes may serve several channels at once; they should not be loaded onto a single channel when they cannot be split, but the allocation rule must be written down.

Third is fulfilment and cash cost. Front-desk handling, reconciliation, customer service, special requests, guest compensation, receivables terms, and bad-debt risk can all vary by channel. A short-credit corporate booking and a long-credit wholesale block cannot be compared on the same stay date alone; the capital tied up in credit must enter the comparison.

Finally, look at demand quality, but treat quality as judgement, not as profit in advance. Lead time, length of stay, cancellation rate, shoulder-season fill, share of new guests, and repeat business should all be listed separately. Repeat business that has not yet occurred cannot be written into this period's net contribution.

Put the four accounts into a channel net contribution bridge table

Each row of the table corresponds to one channel, one room type, and a comparable set of stay dates. Keep at least these fields: effective occupied room nights, net room revenue, ancillary revenue, refunds and cancellation losses, acquisition and promotion costs, payment and technology costs, fulfilment labour, guest complaints and compensation, credit cost, receivables balance, net contribution per room, lead time, cancellation rate, and demand task.

Net contribution per room can be calculated as:

Net contribution per room = net room revenue − attributable acquisition cost − payment and technology cost − fulfilment cost − credit cost

Net room revenue is the realised revenue from net rates plus collected ancillary revenue, minus refunds and cancellation losses. The formula does not use confirmed bookings as the denominator; it uses effective occupied room nights. Otherwise, channels with different cancellation rates will still produce a distorted comparison.

Three errors destroy the bridge table. The first is incomparable dates: mixing weekends with weekdays, or event days with ordinary days. The second is missing costs: filling in only commission while leaving out advertising, payment, labour, and credit. The third is writing unrealised repeat business, brand exposure, or subjective valuation into confirmed revenue. Once the table is complete, finance should confirm the revenue and cost definitions, revenue management should confirm comparable dates, sales and marketing should complete acquisition costs, and front office and customer service should complete fulfilment hours and compensation.

Run a 28-day minimum-cost test

The test does not need to cover a full year at once. Choose 28 consecutive days, two channels, one room type, and the same stay-date structure, and exclude event days, sold-out days, closed days, and major service anomalies that clearly distort price. The 28 days are for finding definition gaps and early signals; they do not, on their own, prove long-term causality.

Finance first confirms the definitions for revenue, tax, cancellation, and credit. Revenue management selects comparable dates. Sales and marketing backfill actual advertising, commission, and promotion spend. Front office and customer service record the hours and compensation attached to each booking. Before the test starts, write down four decisions: continue with current allocation, limit investment, correct the cost attribution, or pause scaling. It does not directly support closing an entire channel, because a channel may be carrying the task of international demand, long-stay demand, shoulder-season fill, or new-guest acquisition.

Verify only five outcomes: net contribution per room, net revenue after cancellation, receivables turnover, channel concentration, and the share of incremental demand. If net contribution deteriorates, or if cancellation rates and guest complaints rise together, stop scaling first and re-check the definitions. When cost attribution remains unclear, pause shifting budget across channels; do not use a lower surface commission to hide a larger fulfilment loss.

AI entry points belong in the acquisition account

The D-EDGE 2026 report also notes that artificial intelligence is already changing how hotels are searched, compared, and shortlisted, but has not yet changed the channel balance in booking data. For a hotel, AI is not a naturally cheaper new channel; it is an acquisition method that still needs verification.

The cheapest verification is not complicated. Distinguish natural search, paid clicks, AI search, and conversational entry in the CRM and order notes. Record whether each valid enquiry produced incremental demand. Then allocate the corresponding advertising, software, content, and technology spend to the booking. Judge the result with the same net contribution per room. If an AI entry point only adds exposure while crowding out direct bookings that would have happened anyway, or leaves no attributable revenue, exposure volume cannot be written down as acquisition success.

For the next channel budget meeting, bring one thing only: a channel net contribution bridge table completed under one consistent definition, then let a 28-day test verify the differences. Commission rates can stay on the agenda, but they cannot stand in for the amount of cash a booking finally leaves in the hotel.