OTA Escape Velocity: how a hotel breaks free from 18% commission without losing occupancy
This kicks off Hospitality Loyalty Brief — a series on retention economics for hoteliers tired of paying OTAs the equivalent of a senior manager's salary.
The price of the "reliable channel"
Booking.com takes 15-22% of gross. Expedia — 18-25%. On paper, this is "marketing spend". On the P&L, it's the annual revenue of a 40-60 room hotel flowing to a platform that doesn't physically exist.
Let's run the numbers on a real mid-size property:
- 100 rooms, 65% average occupancy, ADR €85
- Annual revenue:
100 × 0.65 × 365 × €85 ≈ €2.02M - OTA share 55% (typical for resort):
€1.11M through aggregators - Effective commission 18%: ≈ €200K per year
€200K per year — that's two senior sales hires, a full digital team, or two turn-key Launch loyalty programmes with a six-year horizon.
Why "reduce OTA share" is the wrong target
The consultant's standard advice: "cut Booking 10 points, shift traffic to your direct site". This doesn't work in hospitality, because:
- OTAs deliver your shoulder traffic. In non-peak days — half of your bookings. Cut them off = −25% occupancy overnight.
- Direct acquisition costs more than it looks. Google Hotel Ads + retargeting + brand bidding typically lands CAC at 12-15%. So "direct booking" is sometimes only 3-5 points cheaper than OTA.
- A new guest's "direct booking" is a one-shot transaction. If they don't return in 14 months, you paid acquisition cost for one stay.
The right goal is not "reduce OTA" but increase LTV of every guest who's ever been to your hotel. Then OTAs become a one-off acquisition channel, not a lifelong dependency.
The math of escape: at what repeat-rate does loyalty pay off
Base model — a guest who came once either returns (repeat) or doesn't. The loyalty programme moves that probability.
Two scenarios for our mid-size hotel:
| Metric | No programme | With Launch programme |
|---|---|---|
| Guests / year | 23,720 | 23,720 |
| 12-mo repeat % | 14% | 23% (+9 pp) |
| Additional stays | 3,320 | 5,455 (+2,135) |
| Repeat ADR uplift | — | +8% (premium categories) |
| Extra revenue / year | — | +€196K |
| Programme cost | — | −€54K (one-off + 12 mo MRR) |
| Year-1 net | — | +€142K |
Sources: McKinsey 2024 "Hospitality loyalty pulse", Skift Research Q1 2026, HSMAI benchmark indices. Simplified — full P&L includes cannibalization, training, and redemption costs.
Break-even: a 9-point lift in repeat-rate
Rough rule: the programme pays for itself in 12 months if you can lift repeat-rate by 9 percentage points at current size and ADR. Most hotels achieve +6-12 pp in year 1 with a properly designed programme — but "properly" here means not "bought SaaS" but "wrote ICP, tier economics designed for your ADR, trained front-desk, launched multi-channel engagement".
Why 9 of 10 programmes die in year 2
Of the 50+ programmes I've seen launched in Russia, CIS, and MENA over the past 7 years, ~90% stop growing by end of year 2. Why:
- Tier design without a P&L model. Guest earns 8% cashback, but average redemption cost for the hotel is 11% margin. Programme is loss-making from first redemption.
- Reward catalogue out of the box. "Free night after 10 nights" — that's a Marriott-style default. For a resort with summer peak and winter desert, this reward creates exactly the low-season load when you need paid rooms, not freebies.
- No operational team. Software is live but no one segments the base, writes campaigns, moves tier qualifications.
- Not integrated with PMS. Guest checked in via Opera — but loyalty system says "didn't arrive". Front desk handles verbally. Trust evaporates fast.
What to do this week
- Calculate your annual OTA commission spend.
OTA revenue × average commission. That's the retention-infrastructure budget you already have. - Determine current 12-mo repeat-rate. If PMS doesn't compute it — pull a dump by email/phone. 14-16% is typical for resort, 8-10% city, 20-25% sanatorium.
- Build a simple model. "What happens to EBITDA if repeat-rate moves +5 / +10 / +15 points?" That's your loyalty investment threshold.
- Decide who will operate the programme post-launch. This is the critical question. 90% of programmes die because no one's left after go-live.
