Should a Tour Operator Automate Booking Changes or Simplify Its Change Policy First?

The inbox that fills up the week before every departure

Your reservations team spends the busiest part of the week not selling. They are moving a family of four from Thursday to Saturday, working out whether a guest who cancelled at hour 23 gets half back or a quarter, and explaining to someone who booked through Viator why the policy on your own website says something different.

The obvious fix is software. Every booking platform in the category will show you a self-service change flow, automated refund calculations, and a dashboard that tracks cancellations by channel. Those tools work. We are not arguing against them.

We are arguing about the order. Most tour operators who buy change automation are buying a machine to enforce a policy nobody would design on purpose today. The policy is the cost driver. The software just runs it faster.

Page one couples the two and gets the sequence backwards

Search for help with tour cancellations and you get two kinds of result.

The first is booking software content. It treats automation as the way to apply the policy you already have: configure the refund windows, let guests self-serve, and stop touching every request by hand. The implicit assumption is that the policy is a fixed input.

The second is policy templates. They hand you a tiered refund ladder and tell you it balances guest flexibility against protection for the business. The implicit assumption is that more tiers mean more fairness.

Put those two assumptions together and you get the standard outcome. An operator adopts a template, then buys software to administer it, then discovers the software project is mostly about encoding exceptions.

Count the outcomes in a typical policy

Take the free template that currently ranks for day tour cancellation policies, published by Automate.travel for 2026. It is a sensible document, and that is the point. A careful, well-intentioned policy still produces a lot of branches.

The standard guest cancellation ladder alone has four outcomes: a full refund at 48 hours or more, 50 percent between 24 and 48 hours, 25 percent between 12 and 24 hours, and nothing inside 12 hours.

Then the modifications. Date changes are free more than 24 hours out, and inside 24 hours they are subject to availability and possibly a rebooking fee. Name changes are free at any time. A no-show gets nothing, except that a one-time rebooking credit may be offered at the operator's discretion.

Then the operator side. Cancellation for low numbers or safety means a full refund or a rebooking. Force majeure means a full refund or a credit valid for 12 months. The weather clause sets separate thresholds for wind, rainfall, wave height, lightning distance, and temperature.

Then groups of ten or more, which get their own ladder: a 25 percent non-refundable deposit, and refund steps at seven days, three days, and 48 hours, plus an allowance to shrink the group by 10 percent without penalty.

That is well past a dozen distinct outcomes before a single booking arrives from a reseller. Several of them contain the words "may" or "at the operator's discretion," which is policy language for "a person decides."

The channels multiply it

Now add distribution. Viator's default Standard policy gives a full refund for cancellations at least 24 hours before the experience, and the platform also supports "All Sales Final" and custom policies. GetYourGuide defaults to free cancellation up to 24 hours out, and sells travellers a cancellation upgrade that extends the window to 60 minutes before start. It also charges suppliers a fee for cancellations it considers unjustified.

So a guest's refund does not depend on your policy. It depends on your policy as it was configured on the channel they used, on the day they booked. The consumer advocacy group Truth in Advertising has criticised Viator on exactly this seam, where some listings flagged as strict require seven days' notice despite free cancellation messaging elsewhere on the site. When the rules differ by channel, the guest experiences that as inconsistency, and your phone rings.

The mechanism is arithmetic. The cost of handling changes scales roughly with the number of distinct rules multiplied by the number of channels those rules have to be reconciled across. You can reduce either number. Only one of them is free.

Cutting channels costs revenue. Cutting tiers costs a policy meeting.

Why every tier is a cost, not a feature

A refund tier looks like a line of text. In operations it is four things.

It is a branch someone has to build. Every booking system can handle a refund window, but each additional window is another configuration to maintain per product, per season, and per channel.

It is a branch someone has to test. When you change a departure time or add a morning slot, every tier boundary shifts with it. The 12 hour cutoff on a 7 a.m. tour falls at 7 p.m. the previous evening, which is when nobody is in the office.

It is a branch someone has to explain. A four step ladder guarantees that a meaningful share of cancellations will land just the wrong side of a boundary. Those are the guests who call, and they call to negotiate.

It is a branch someone has to override. The discretionary clauses exist because rigid tiers produce outcomes that feel unfair in the moment. Every discretionary call is a manual task that no automation will take off your hands, because the whole point of it is judgement.

This is the same pattern we see in contract logistics, where a warehouse buys billing automation to administer an accessorial rate card that has grown one exception at a time. We wrote about that in whether a 3PL should automate billing or standardise its accessorial rules first. The software industry is different. The shape of the problem is identical.

What vendor guidance leaves out

Tour booking platforms, including Bokun and Atlasperk, give reasonable advice about what to automate first. They sequence it by volume: confirmations and vouchers, then reminders, then changes and cancellations. That is a sensible order for the work the software does.

What none of that guidance asks is whether the rule set is the thing generating the volume. It would be an odd question for a vendor to raise. Their product is priced on bookings and features, not on how many tiers you deleted.

We sit on the other side of that table. Our job is to ask the question before you sign.

The tiers that usually survive a hard look

We will not hand you a replacement policy here. The right one depends on your cost structure, your supplier terms, and how much of your volume comes through resellers with their own rules. But the questions that decide it are consistent.

Does the middle tier recover money or just generate calls? A partial refund band recovers some revenue on late cancellations. It also creates the boundary disputes. Many operators have never compared the revenue a middle tier keeps against the staff time it consumes.

Does your direct policy match your dominant channel? If most of your volume comes through an OTA with a 24 hour free cancellation default, a different ladder on your own site mainly produces confused guests and a second set of rules to maintain.

Which discretionary clauses are really rules in disguise? If staff grant the no-show rebooking credit nine times out of ten, it is a rule. Write it as one and remove the judgement call, or drop it.

Which of your supplier terms force a tier? Some tiers exist because a boat charter, a park permit, or a restaurant deposit runs on its own cancellation clock. Those are real. Many tiers exist because a template had them.

Every tier that fails those questions is one fewer boundary to dispute and one fewer branch to configure. The change volume the automation was being bought to absorb starts falling before the software is even chosen.

Where regulation limits how simple you can go

Simplification has edges, and they are legal, not preference.

If you sell packages to travellers in the EU, the Package Travel Directive (EU) 2015/2302 gives travellers the right to terminate before departure on payment of an appropriate termination fee, and to cancel without a fee if the organiser significantly changes the package or raises the price by more than 8 percent, with refunds due within 14 days. A policy cannot contract out of that.

In the US, the FTC's Rule on Unfair or Deceptive Fees, in effect since May 2025, targets drip pricing for short-term lodging and live event tickets. It does not cap change fees, but it is a clear signal about how regulators view charges a consumer only discovers late in the process. A rebooking fee buried inside a discretionary clause is the kind of thing worth reviewing with counsel.

Neither of these argues for more tiers. Both argue for fewer rules that are stated plainly.

Then automate what is left

Once the policy is short, automation becomes a much smaller and cheaper project. Self-service date changes, automatic refunds on a two step ladder, and channel settings that match each other are close to out-of-the-box configuration on most modern booking platforms.

The expensive part of change automation was never the software. It was translating a policy with a dozen outcomes and several judgement calls into logic a system could run. Remove the complexity first, and the implementation shrinks to match. You also avoid paying a vendor to hard-code rules you would have deleted next year anyway.

The same principle runs through when not to use AI in business: automating a process tends to freeze it, so the process should be one you want to keep. And if you are on the hotel side of travel, inconsistent PMS usage across properties is the same story told with rate codes instead of refund tiers.

In short, your refund ladder has too many rungs. Take a couple out before you pay someone to climb it.

Where a conversation helps

An article can show you why tier count times channel count drives your change volume. It cannot tell you which of your tiers are protecting margin, which are protecting a supplier obligation, and which are left over from a template adopted in 2019.

That takes a look at your cancellation data by channel, your supplier contracts, and the actual time your team spends on changes. If your reservations staff spend their busiest days negotiating refund boundaries, and you are being quoted automation to make that faster, talk to us before you sign. We will tell you how much of the problem the software would solve, and how much a shorter policy would solve for free.