A Book-Direct Programme Whose Business Case Hinges on Unproven Incrementality

Step 1 - Net revenue per room night by channel

• OTA: $164 x (1 - 0.17) = $136.12
• Brand.com: $164 x (1 - 0.075) = $151.70
• Gain per shifted room night: $15.58

Step 2 - Value of the 4-point shift 65,000 x 4% = 2,600 room nights shifted. Gross gain: 2,600 x $15.58 = $40,508. But the member rate is 5% off BAR, and the shifted bookers are exactly the ones who will take it. If half do: 1,300 x $164 x 5% = -$10,660. Net shift benefit: $29,848.

Step 3 - Value of the 600 incremental room nights Contribution per incremental direct room = $164 x 0.925 - $38 = $151.70 - $38 = $113.70. 600 x $113.70 = $68,220.

Step 4 - Total $29,848 + $68,220 = $98,068 against a $96,000 cost. Net +$2,068 - roughly a 2% return, which is inside the forecast error of every assumption in the model.

Step 5 - Stress the assumption that matters. The whole case rests on the 600 incremental rooms. Two objections:

• The hotel already runs 68% occupancy; new direct bookings may simply displace OTA or GDS bookings. If 40% of the 600 are cannibalised from GDS (net $145.96), the incremental value falls to about $42,300 and the programme returns -$23,800.
• The agency has not defined incrementality. Paid search on the brand name will capture demand that would have converted anyway.

Verdict: the case is not proven. It is break-even on optimistic assumptions and clearly negative on realistic ones. Do not commit $96,000 on this evidence; fund a properly controlled test.

Rate parity A member rate that is publicly visible and open to anyone is, in substance, a public discount, and most OTA agreements require that publicly available rates are distributed at parity - or at least that the hotel's own site is not undercutting. The exposure is twofold: the OTA can apply a rate-parity complaint, and the hotel's own distribution staff may be forced to load the lower rate everywhere, which converts the 'direct' strategy into an across-the-board 5% cut.

What is safe:

1. Genuinely closed user group rates - authenticated login, membership required before the booking date, not discoverable through metasearch or rate-shopping tools. Many agreements permit this if it is real.
2. Value-adds instead of rate cuts - breakfast, points, lounge access, free cancellation, upgrade. Parity clauses generally govern rate, not bundled value, and the guest's perceived value usually exceeds the cost.
3. Renegotiate - a defined narrow exception for loyalty rates, in writing, before launch.

Consent and retargeting A tracking pixel that fires before consent, or that builds ad audiences from stay data, is a data-protection problem in most jurisdictions the hotel sells into. Practical minimum: consent management platform with opt-in before non-essential cookies, no use of stay or booking data for ad targeting without a lawful basis, a documented data-processing agreement with the agency, and a retention limit on the audience lists. The legal team's question should be answered before the pixel goes live, not after.

Practical sequencing: parity-safe creative (brand search, value-add, loyalty signup) in the test; genuine member rates only behind login and only after written exceptions are in place. Ethics point worth raising to the owner: a member rate that anyone can access is not a segmentation strategy, it is a discount with a story attached.

Test design (90 days)

• Split by geography or daypart, not by date. Target two or three feeder markets or a defined day-of-week band; hold a matched control cell that receives current activity only. Splitting by time destroys comparability because demand moves.
• Separate the two mechanisms. Run brand search plus value-add in one cell and the 5%-off member rate in the other, so you learn which one produces the shift. Bundling them means you never know what you bought.
• Tag every booking with a campaign code and a first-touch source, and match against the loyalty file to see prior-stay share.
• Measure incrementality, not attribution. Compare total room nights and total net revenue in test vs. control, not bookings attributed to the campaign.

Metrics

• Net revenue per available room after channel cost - the number the programme is trying to move. Report it by channel, not blended.
• Cost per incremental direct booking (media + fees + discount give-up) / incremental rooms. Target: comfortably below the $15.58 channel saving, otherwise the programme is buying its own margin.
• Cannibalisation rate - share of 'direct' bookings that were previous OTA or GDS bookers.
• Direct share of member stays (currently 34%) and repeat-stay rate at 6 and 12 months, which is the real durable value of direct.
• Cancellation rate by channel and ADR by channel - a direct channel that buys volume with a 5% discount and a 27% cancellation rate has not created anything.

Stop rules Stop or restructure if: cost per incremental direct booking exceeds the $15.58 saving; cannibalisation across GDS/OTA exceeds 40% of 'incremental' rooms; any OTA issues a parity complaint or re-prices; or consent compliance cannot be evidenced before launch. Scale only if net revenue per available room in the test cell beats control by more than the $96,000 run rate, with a margin wide enough to survive the forecast error that this business case already demonstrated.