Two operators sell the same product. Same coastline, same boats, same caliber of guides, same $95 half-day kayak tour.
Operator A runs a strict policy: no refunds inside 72 hours, no exceptions. The logic feels airtight. Slots are hard to refill. Cancellations cost us money. Rules protect the business.
Operator B runs a graduated policy: full refund outside 48 hours, 50% credit between 48 and 24 hours, no refund inside 24 hours except for unsafe weather, in which case full rebook credit valid for 18 months.
A season later, Operator B books about 10% more direct traffic, carries a lower effective cancellation cost (because cancelled slots refill better thanks to earlier notice), and produces measurably more repeat bookings the following year.
Same product. Same market. Different policy. The difference isn't about generosity. It's about design.
We don't think about cancellation policy much, most of us. It feels like a legal document we wrote once, tucked into the terms page, and stopped reading. But policy is doing real work in our business every day, and the version most of us are running isn't the one that would actually serve us best.
The Three Levers Policy Is Pulling
Every cancellation policy is doing three things at once, whether the operator who wrote it thinks of them that way or not.
It affects conversion. Stricter policies cost bookings at checkout because customers hesitate. The hesitation is invisible (no operator gets a notification that says "customer abandoned because of your refund terms") but it's real and it's measurable in conversion data across thousands of operators.
It affects margin. Looser policies cost money when seats don't get resold. The cost is most visible in the operator's monthly P&L (refunds processed, credits issued, deposits returned) and it's the cost most of us focus on.
It affects trust. How customers feel about us after the booking shapes their reviews, referrals, and repeat behaviour. A policy that felt fine to write feels different to a customer who actually needed to invoke it.
Most operators optimise for the margin lever and accept the cost to conversion and trust as inevitable. They aren't inevitable. The right policy moves all three in your favour at once.
The OTA Standard Is Now the Floor
Here's an uncomfortable fact about our pricing power on policy: we don't really set the customer's expectations anymore. The OTAs do.
GetYourGuide, Viator, and Airbnb Experiences have spent years training a generation of travellers to expect free cancellation up to 24 hours before the experience. For our customers, this stopped being a feature years ago. It's a baseline assumption.
When our direct booking site offers stricter terms than the OTA version of the same product, two things happen at the same time. Conversion drops at checkout, invisibly, because there's no event that says "lost because of refund terms." And the buyers who do complete a booking often do it through the OTA, where the terms are more forgiving, paying us less margin on bookings we could have owned directly.
This is the hidden tax of restrictive direct policy. It looks like margin protection. It functions like a channel-mix penalty.
The practical implication: treat OTA cancellation terms as the minimum acceptable direct policy, not the maximum. Being more generous than the OTA is a competitive weapon. Being more restrictive is a silent revenue leak we're paying every month and not noticing.
One Policy Doesn't Fit All Our Products
A blanket cancellation policy is almost always wrong, because it treats unlike products as if they were the same. The walking tour we run twice a day with 25 seats has different reselling economics from the private fishing charter we run twice a week with one boat. They shouldn't share terms.
The framework that works starts with one question: how recoverable is the slot?
Shared, high-frequency, easy to refill. A walking tour with multiple daily departures. Cancelled seats refill within hours. Policy can be permissive without much margin cost: 24-hour free cancellation, no-show forfeiture, a low-friction rebook path.
Shared, low-frequency, hard to refill. A weekly food tour with 12 seats. Cancelled seats inside 72 hours often don't refill. Policy needs more protection: 72-hour free cancellation cutoff, partial refund inside that window, credit-based rebooking after.
Private or semi-private with dedicated staffing. A private hike with one guide assigned per booking. Once the guide is committed, the slot is functionally gone. Tiered refunds (full outside 7 days, 50% between 7 and 3 days, credit only inside 72 hours) with clear language explaining why.
High fixed-cost departures. A small boat with a four-person crew and a minimum threshold. Two-layer policy: customer-facing cancellation window, plus an operator-side minimum threshold clause that lets us cancel with full refund if the departure doesn't hit viability.
The point isn't these four categories specifically. It's that policy should be designed per product, not per company.
The Single Most Important Number in Your Policy
How far out should free cancellation end? This is the highest-leverage number in the whole policy, and most of us set it by gut.
Two forces pull in opposite directions. A longer window gives us more time to resell cancelled slots (good for margin). A shorter window signals flexibility and lifts conversion (good for top line).
The answer isn't universal. It comes from three questions about your specific business.
What percentage of bookings arrive inside the window you're considering? If 60% of your bookings come in during the 72 hours before departure, a 72-hour cancellation cutoff is irrelevant to most buyers but maximally restrictive for the minority who booked early. Worst of both worlds.
What's the historical refill rate on cancelled slots at different windows? Most of us have never measured this. When we do, the results often surprise us; refill rates inside 48 hours are often higher than expected for popular products and lower than expected for niche ones.
How does our window compare to OTA terms for the same product? If our direct window is longer than the OTA's, we're actively training customers to book through the OTA.
A worked example to make this concrete. A food tour selling 16 seats at $120 a head. If 48-hour cancellations refill at 75% and 12-hour cancellations refill at 20%, the math for a 24-hour window becomes obvious. The window captures almost all of the refillable cancellations while preserving the conversion benefit of flexible terms. A 72-hour window would protect a little more margin on the refill side but leave a lot of bookings on the table at checkout.
No-Show Is a Different Problem
Cancellation and no-show often get written into the same clause. They shouldn't. They're different problems with different solutions.
A cancellation is a guest who tells us they can't make it. A no-show is a guest who disappears. The first gives us a chance to resell. The second doesn't.
Cancellation policy should balance conversion and margin. No-show policy should focus on reducing the no-show rate itself, because no-shows are almost entirely unrecoverable revenue once they happen.
The most effective no-show reduction isn't punitive. It's communicative. A 48-hour confirmation request with a one-click "I'll be there" button. A 24-hour reminder with arrival instructions and a contact number. A 2-hour day-of nudge for activities where guests commonly get lost or run late.
Operators who implement this kind of three-touch pre-arrival sequence often see no-show rates drop from 8 to 12% down to 2 to 4%. The recovered revenue dwarfs anything a punitive no-show clause would claw back, and the automated sequence costs roughly nothing per booking once it's running.
No-show forfeiture should still exist in the written policy, both as a margin backstop and as a signal that attendance is expected. But the real work is done upstream in the reminder sequence, not in the clause.
Weather and Other Things We Can't Control
The most emotionally loaded category. A family that drove three hours for a whale-watching trip and gets cancelled for high seas is already disappointed before the refund question comes up. How we handle the next 48 hours determines whether that family leaves a two-star review or a five-star one.
Three principles that hold up across most operators.
Make the call early and communicate early. A cancellation notified at 5pm the night before lands very differently than one issued at 8am on the day of. Guests adjust. We rebook more of them.
Default to credit, not refund, for operator-initiated cancellations. A rebook credit valid for 12 to 18 months keeps the revenue in the business and gives the guest an implicit reason to return. Most guests accept this willingly if the alternative (full refund) is also offered transparently. Forcing credit without the refund option creates chargebacks; offering both builds trust.
Separate "we cancelled" from "you cancelled" in policy language. Guests intuitively understand that a guest-initiated cancellation has different terms from an operator-initiated one. Policies that blur this line feel unfair even when the economics are reasonable.
The Cancellation Moment Is a Retention Moment
Here's the counterintuitive finding in this whole topic: the guests who cancel are often the guests most likely to rebook, if they're handled well.
A guest who cancels has already decided they want the experience. They picked it, they paid for it, and then something got in the way. If the cancellation experience is frictionless and respectful, their willingness to come back is usually higher than a guest who browsed and left without booking at all.
The cancellation flow isn't a loss-mitigation process. It's a retention touchpoint. Designed accordingly:
- Offer rebooking as the first option, not the refund. A one-click "pick a new date" flow converts a meaningful share of cancellations into rescheduled bookings.
- If a refund is the outcome, process it quickly. Same-day or next-day refunds generate goodwill that slow refunds (5 to 10 business days) quietly destroy.
- Follow up two to four weeks later with a soft offer, not a sales push. "We saw your plans changed. If you're still interested, here's a reminder of what you missed, and a small credit if you'd like to try again."
The operators who do this well turn what looks like a leak in the funnel into a measurable source of repeat bookings. The metric to track is cancellation-to-rebook conversion, separate from new-booking conversion. Most of us don't measure it at all. The ones who do usually find it's higher than expected.
Designing Policy Like We'd Design Product
The through-line of all of this: cancellation policy isn't a document to copy from a template. It's a set of connected decisions that shape how customers find us, how much of our revenue survives cancellations, and how likely our guests are to come back.
Most of us set policy once, write it into our terms, and never revisit it. The operators who treat it as a living part of the commercial stack, reviewed seasonally against actual data on conversion, cancellation rates, refill rates, and repeat bookings, consistently outperform on all three of the levers this piece opened with.
Dazhboards handles the operational mechanics behind a well-designed policy (the rebook flows, the credit tracking, the refund processing, the cancellation-to-rebook follow-up) so the policy you've designed actually executes the way you intended, rather than getting bottlenecked at the operations layer. The hardest part is the design; the execution is the part where infrastructure matters.