How to win back direct bookings from OTAs
Not by outbidding them for strangers. By using the guest history you already hold, and answering the phone when somebody is standing on an OTA page reading you a price.
Answered in short
4 things that decide this
- 01Win back the guests you already served first, because they cost nothing to reach and your PMS already knows who they are. A guest who stayed twice and books through an OTA the third time is the most expensive booking you take, and the easiest one to move.
- 02OTAs captured 61% of independent bookings in 2024, at 15 to 30% or more in commission (Cloudbeds 2025), while costs rose faster than revenue across the sector that year (CBRE 2025). That gap is why the direct question is a margin question rather than a marketing one.
- 03Answer the phone during the wave, because a caller quoting an OTA price is a direct booking standing in front of you. Your line can offer your direct rate and what it includes, by rules you set, at the moment the guest is still deciding.
- 04Then reconcile, so you can tell whether any of it worked. Channel commission matched against what accounting actually received is what turns direct share from a feeling into a number.
You cannot outspend an OTA for a stranger, and you don't have to
Most direct-booking advice is really advice about paid search: bid on your own brand, build a better booking engine, run a rate-parity campaign. Some of that is worth doing. None of it changes the basic asymmetry, which is that an OTA can spend more to acquire a guest than you can, because it amortises that spend across every property in the city and you have one.
So the winnable ground isn't the stranger comparing forty hotels. It's the two groups the OTAs don't own: guests who have already stayed with you, and callers who are on your phone right now with an OTA tab open. Both of those people already know your property. Neither of them needs to be acquired again.
That reframing matters because it changes what you build. A better booking engine helps a guest who already decided to come direct. It does nothing for the guest who never thought to check, which is most of the 61%.
Your past guests are a direct channel you already paid for
Your PMS holds who stayed, when, in what room type, at what rate and how often. That's a better targeting dataset than anything you could buy, and most properties do nothing with it beyond a quarterly newsletter that goes to everybody at once.
What works instead is narrow and timed. A guest who came for the same festival two years running gets an offer before that festival, not in January. A business guest who stayed four Tuesdays gets a different message from a couple who came for an anniversary. And the offer is a reason to book direct rather than a discount that undercuts your own channels, which is usually something the OTA can't sell: the room they liked, a late checkout, a room-type upgrade at your discretion.
None of this needs a new system. It needs the stay history read properly, consent captured before anybody is messaged, and the send tied to the guest's own pattern instead of a calendar somebody maintains by hand.
- 01Segment by stay history, not by a single mailing list: season, room type, party shape, repeat pattern.
- 02Offer what an OTA can't: the room they liked, a late checkout, an upgrade at your discretion.
- 03Capture consent before a guest is messaged, and log every touch so you can read what was sent and when.
The OTA-to-direct moment happens on your phone, at 4pm
A caller who says "I'm looking at you on Booking.com and it's showing this price" has already chosen your hotel. They're asking whether to give you the commission or not, and the answer depends entirely on whether somebody picks up and what they're allowed to say.
In practice that call lands in the middle of the check-in wave, gets a hold button, and converts on the OTA instead. So the mechanism is unglamorous: answer the call, know the guest is quoting a channel price, and respond by a rule you wrote in advance. What your direct rate includes, what you'll add, and what you won't match are decisions you make once and the system applies consistently, rather than decisions your newest desk agent makes under pressure.
This is also where most of the value sits, because the guest is at the point of purchase rather than at the top of a funnel. It's the cheapest conversion in the building and it usually goes to voicemail.
- Past guestAlready yours. Reached from stay history, timed to their own pattern.
- The caller with an OTA tab openAt the point of purchase. Won or lost by whether you answer.
- The in-stay momentRebook for next year while they're still in the room.
- ReconciliationCommission against payout, so direct share is a number, not a feeling.
Notice what isn't here: the stranger comparing forty properties on an OTA page. That one you rent, and it's fine to.
Rate parity, walks and chargebacks are never a rule the software decides
Every serious direct strategy runs into your channel contracts, and that's where automation has to stop. What you may advertise, what you may offer on the phone, and what counts as parity are legal and commercial questions with real consequences, and they belong to whoever signed the agreement.
So we build the system to prepare rather than decide. It shows you the commission, the payout and the gap. It can offer a rule you set on a call. It never sends a rate change to a channel on its own, never decides who gets walked when you're oversold, and never processes a chargeback response.
- 01Rate parity and what you may offer on a call are your decisions, set as rules in advance.
- 02Overbooking walks stay with your duty manager, always.
- 03Chargebacks are prepared with the evidence attached and answered by a person.
- 04Card details never enter a call transcript or a message log.
Some of the systems we have shipped
Related questions
01Should we just leave the OTAs?+
Almost never, and the properties that try usually come back. OTAs are genuinely good at reaching people who have never heard of you, and that's worth paying for. What isn't worth paying for is commission on a guest who already stayed with you twice, or on a caller who was on your phone asking about the price. Win those back and keep renting the rest.
02Doesn't rate parity stop us offering a better direct price?+
Your contracts decide that, and they differ by channel and by market, so this is a question for whoever signed them rather than for a software vendor. What's usually available without touching rate is value: the room they liked last time, a late checkout, an upgrade at your discretion, or something the OTA has no way to sell. We build the rules you set, and never send a rate change to a channel on its own.
03How do we know if it worked?+
By reconciling channel commission against what accounting actually received, month by month, and reading direct share out of your PMS source report rather than out of a booking engine dashboard. Direct share quoted from memory is almost always a few points optimistic. Deciding what you'll measure before you start is the part most direct campaigns skip.
04Do we need a new booking engine for this?+
Usually not. A booking engine converts a guest who already decided to come direct, and that's a real job, but it's downstream of the problem. The guests you're losing never got to your site at all, so the work sits earlier: in the stay history you already hold and the phone that rings during your check-in wave.
05Is emailing past guests allowed?+
With consent, and consent has to be captured properly rather than assumed from a stay. That's the first thing we build, and every automated touch gets logged so you can read what was sent to a guest and when. A win-back programme that annoys your best repeat guests costs more than the commission it saved.
Related
- Operations automation →Win-back, reviews, requests and the loops behind them.
- Hotels →The whole chain for a property, built around your PMS.
- AI guest desk →The call where the OTA-to-direct moment is won or lost.
- How should a hotel handle the 4pm check-in wave? →Why that call goes to a hold button in the first place.
- Cloudbeds, Mews and Opera APIs →What it takes to read stay history and write a reservation back.
- AI, automation and custom software for hospitality →Which page owns each part of the guest chain.
