RevPAR is forecast to grow 1.2% globally this year, and roughly 1.4% across Europe. Most of that comes from rate rather than occupancy: CoStar's Q3 2026 forecast, published 31 August, puts ADR growth at 0.8% and occupancy growth at 0.5%, falling to 0.3% next year.
For a hotel that is already close to full in season, that is the whole ceiling. You cannot sell the same room twice, your rate is bounded by the property down the road, and the occupancy points that used to be available have mostly been taken.
The number that is not bounded is what each guest spends once they have booked. That is where the growth is now, and it is the one line most independent hotels have never measured.
What ancillary revenue actually covers
Everything a guest pays you for that is not the room.
Breakfast, room upgrades, early check-in, late check-out, parking, spa and wellness, restaurant and bar, airport transfers, experiences and tours, pet fees, bike hire, minibar. At a property with a restaurant it also includes covers you would otherwise lose to the place across the street.
Two things separate it from room revenue and both matter.
It carries higher margin. Selling an upgrade from a standard double to a room you already own costs you nothing but the rate difference. Parking is close to pure margin once the spaces exist.
And it is not capacity-constrained in the same way. A hotel at 95% occupancy in August has no rooms left to sell but has barely touched breakfast, parking or late check-out.
The timing problem, which is really the whole problem
Most independent hotels do attempt to upsell. They do it at the front desk, at check-in, in the ninety seconds while someone is holding a suitcase and looking for their passport.
That moment is the worst one available. The guest is tired, there is a queue, and the person on shift has to decide whether to risk the interaction. Most days they do not. When they do, the guest says no because they have not thought about breakfast yet and do not want to make a decision standing up.
The window that works is before arrival, while the traveller is still planning the trip and thinking about logistics. They are choosing a train, checking what time they can drop bags, deciding where to eat on the first night. An offer that arrives then is useful information. The same offer at the desk is a sales attempt.
At C-Hotels Silt, a 73-room seafront property in Middelkerke, fewer than 6% of arriving bookings received any message at all across an eight-month window before launch. Breakfast and parking existed as products and were sold only if someone remembered to ask. After launch, more than 65% of arrivals were being reached, and in the best full month it passed 80%.
The five moments worth building
Our step-by-step setup guide covers the mechanics of building these campaigns. This is what each moment is for.
Booking confirmation. Not an offer, a mention. Early check-in available from €X, breakfast served until 10:30, parking on site. The traveller is not buying yet, but they now know these exist, which is what makes the later offer land.
Twenty-four to forty-eight hours before arrival. The strongest window for anything requiring planning: room upgrades, early check-in, transfers, a dinner reservation. The traveller is packing and working out logistics. Room upgrades belong here rather than closer in, because the decision involves money and comparison.
The evening before arrival. Breakfast. This is later than most hotels expect and it converts better, because the guest is now thinking about tomorrow morning specifically rather than the trip in general. It is also operationally sane, since your kitchen gets a count the night before rather than a surprise.
Day one of the stay. Restaurant, spa, experiences. The guest has arrived, formed an impression, and has an evening to fill. This is the only genuinely in-stay moment worth automating, and it should be light.
The night before departure. Late check-out and the transfer. Late check-out is the most under-sold product in independent hospitality, because it is only relevant for about twelve hours and the front desk is never asked at the right moment.
Two rules govern all five. Never offer something the guest has already bought, which is the failure that makes automation feel stupid rather than helpful. And suppress the offer when you cannot fulfil it, because a confirmed early check-in you cannot deliver costs more than the fee was worth.
Most of these arrive on WhatsApp, where what you can send and when is decided by rules worth knowing before you design the sequence. Our WhatsApp automation playbook covers them.
The numbers, and one number we are not going to use
You will find a claim circulating that automated upselling delivers up to a 250% increase in ancillary revenue. We are not using it. It comes from a vendor's own product page, it is an "up to" figure with no stated base, and every secondary source repeating it traces back to that one marketing claim rather than to any study.
Here is what is actually measured.
Mews data puts the average upsell value at online check-in at $48, converting at around 6%, which works out to roughly a 0.84% increase in total revenue. That is a real, specific, sober benchmark and it describes one mechanism: the offer presented during digital check-in.
Silt measured a different mechanism. Reaching guests across the pre-arrival window rather than at check-in, it booked over €24,000 in upsell revenue across three months: €15,000+ breakfast, €8,000+ parking, 450+ orders written straight back into the PMS. Spread across every arrival rather than only the ones who converted, that is €12 or more per arrival. In the best full month, better than one booking in eight bought something.
The gap between those two figures is the argument of this article. Both are automated. One asks once, at check-in, through a single screen. The other asks at the moment each product is actually relevant, through the channel the guest is already using, knowing what they booked.
What that looks like at 20 rooms
Take a 20-room property at 70% occupancy. That is 5,110 room nights a year. At an average stay of two nights, roughly 2,555 arrivals.
At Silt's measured €12 per arrival, that is €30,660 a year in ancillary revenue.
Set that against room revenue. At the European Q1 2026 average ADR of €107.80, those 5,110 nights produce about €550,858. The forecast 1.2% RevPAR growth adds roughly €6,610 to that.
The ancillary line is more than four times what a year of forecast RevPAR growth delivers. It also arrives at higher margin, and it does not require the occupancy or rate movement that the RevPAR figure is waiting on.
The other half of the same property's arithmetic is what it pays in commission on the room revenue itself, which we work through in the seven handoffs.
Two honest caveats. Silt is a 73-room property with breakfast, parking and a wellness area, so it has more to sell than a 20-room townhouse. And the two-night average stay is an assumption; substitute your own and the figure moves proportionally. The point is the order of magnitude, not the decimal.
Why this needs the reservation, not just a mailing list
Every moment above depends on knowing something.
The upgrade offer needs to know which room they booked, and that a better one is free on those dates. The breakfast offer needs to know they are not already on a rate that includes it. The early check-in offer needs to know whether the room can actually be turned over, and needs to stop going out once the day's slots are gone. The late check-out offer needs to know the departure date.
An email tool has a name and a date. It cannot do any of that, which is why the generic pre-arrival email performs badly and why most hotels sent it twice and gave up.
There is a second dependency that gets missed. The offer needs the property's own facts, not just the reservation: what breakfast costs today, when it is served, which room types can go early, what the parking fee is. If those live in a document that was last updated in March while the website says something else, the offer contradicts your own site and the guest stops trusting both.
So it needs two things in the same place. The reservation, from the PMS. And the property's current facts, from one knowledge base that your website reads from too. Integrations do not solve this, because an integration copies a fact between two systems that both still hold it and one of them is always behind.
What this does not fix
It does not raise your rate, and it will not rescue a property with an occupancy problem. If rooms are empty, ancillary revenue on a small base is a small number. The order here is occupancy first, then rate, then spend per guest.
It also has a ceiling set by what you actually sell. A hotel with no restaurant, no parking and no spa has breakfast, upgrades and check-in timing, which is a real but shorter list. Be realistic about the size of your own opportunity before building anything.
Your upsell offers, your website and your guest messages should all be reading the same facts, because a guest who gets two different breakfast prices stops believing either. Describe your hotel once, and AskWhisper runs the site and everything that follows it from one record.
Turn lookers into bookers. Just Ask Whisper.
