OTA commissions vs your own booking engine: the actual math
First, the honest part
Online travel agencies earn their commission. Booking.com and Airbnb put your property in front of travellers who have never heard of you, in languages you don't speak, in markets you'll never advertise in. If someone tells you to "get off the OTAs," stop listening to them.
The problem isn't the OTAs. The problem is paying acquisition prices for guests you already acquired.
The returning family who stayed twice last summer. The guest who found you on Google Maps. The traveller your Instagram brought in. When they book through an OTA — and without a booking engine on your site, they have no other option — you pay 15–18% commission to rent back your own guest.
The arithmetic
Worked example — a 12-room guesthouse. Plug in your own numbers as you read; the shape of the math is what matters.
| Input | Example value |
|---|---|
| Rooms | 12 |
| Average nightly rate | €60 |
| Occupancy | 65% |
| OTA commission | 15% |
Room-nights sold per year: 12 rooms × 365 × 65% ≈ 2,850. Annual room revenue: 2,850 × €60 ≈ €171,000.
If every booking arrives through an OTA at 15%, the commission bill is about €25,600 a year. That's the number most owners have never actually multiplied out.
Now the realistic scenario — you don't move everyone, just the guests who already know you:
| Share of bookings moved direct | Commission saved per year |
|---|---|
| 20% | ≈ €5,100 |
| 30% | ≈ €7,700 |
| 40% | ≈ €10,300 |
A direct-booking site — availability, payments, confirmations, an admin calendar your front desk can run — is a one-time build. Against €5–10k of recurring annual savings, it typically pays for itself within the first year or two, and everything after that is margin. There's a second effect the table doesn't show: direct guests are your guests. You have their email, you can market to them next season, and no platform stands between you.
What "a booking engine" actually needs
Owners often think this means "a website with a calendar." The production checklist is a bit longer — and it's exactly the difference between a system guests trust with their card and one they abandon:
- Real-time availability — checked at the moment of confirmation, not when the page loaded, so two guests can't buy the same night.
- Payments your guests use — international cards via Stripe; for Georgian properties, TBC or Bank of Georgia checkout; deposits or pay-on-arrival where that's the norm.
- Instant confirmations — email with a calendar invite the second the booking lands, plus a notification to your team.
- An admin your staff can run — seasons and pricing, manual blocks for phone bookings, guest notes. If only a developer can operate it, it will die.
- OTA coexistence — you keep your OTA listings; the engine handles the direct channel. (Two-way channel-manager sync is a scope decision — worth it for larger properties, skippable for small ones that manage blocks manually.)
- Languages — the guest booking at 2am is rarely browsing in yours.
When you should NOT build one
The honest counter-cases, because they exist:
- You're rarely above ~40% occupancy. Fix distribution and reviews first; commissions aren't your biggest leak.
- Nobody books you by name. If 100% of demand is OTA discovery with no repeat guests and no direct search traffic, there's little booking share to move yet. Build the brand first, engine second.
- A €10/month widget genuinely covers you. A two-room B&B with simple pricing may be fine on a template widget. The custom route earns its keep when you need your own payment providers, policies, multi-language flows, or a real admin.
Where we come in
We build booking systems as fixed-scope projects — availability, payments, confirmations, and an admin calendar, shipped to production with the same race-safe machinery this site's own /book page runs on. Projects start at $8,000, and you'll get the scope and fee in writing before anything begins.
Run your numbers against the table above. If the savings line is bigger than the build line, book a 30-minute call — we'll tell you honestly which side of the math you're on, including if the answer is "don't build it yet."