New software. Same OTA share. What's missing?
You rebuilt the website. You launched a new booking engine. Maybe you even built your own AI assistant.
Then you looked at the channel mix. The booking portals still bring the same share. Your average daily rate has barely moved.
Nothing went wrong with the technology. The product stayed the same.
New shop window. Same shelf.
Open APIs and MCP servers mean any hotel can now build what used to take a big development project. That's real progress.
But the guest still sees the same room categories, the same descriptions and the same price as on the booking portal.
So why should anyone book with you directly? And when an AI agent searches for a place to stay on behalf of its human, why should it pick you?
If your product looks the same everywhere, there's nothing to choose you for. Guests book where they always book. Agents compare "Deluxe" with "Deluxe" and decide on price.
Categories were the right answer when rate grids and booking screens had no room for more. The world has changed since.
"Superior" is not a reason

Most hotels already have real differences: the quiet side, the view, the separate sleeping area, the terrace. They vanish inside Deluxe, Premium Deluxe and Superior Premium Deluxe.
The guest wanted a weekend away. You gave them homework.
Then come the emails every receptionist knows: "Does an extra bed fit in this room?" "Can we still have some privacy from the kids?" "And what's actually the difference between Superior and Deluxe?"
"A lot of times, people don't know what they want until you show it to them." — Steve Jobs
A family rarely goes looking for a "separate sleeping area". They're looking for a holiday where the parents still get an evening of their own. Show them that, and they'll see the value straight away. They'll also pay for it.
Simplify the choice, not the product
Many hotels have just cut their categories to make booking easier. Before AI, that was the right move. A booking page could only show so much, and guests had to sort through it all themselves.
That's no longer true. A family of four is looking for five nights in August. The system already knows who's travelling, for how long and when. Out of 220 products, it shows the three that fit them, and says why.
Simple now means a big choice behind the scenes, and only the right few options in front of the guest.
Cutting categories hides what makes your rooms worth more. Matching shows each guest exactly that.
Guests don't sleep in categories. They sleep in rooms.
That's the idea behind Dynamic Inventory: we made rooms sellable.
Each room or apartment is sold by what it really offers. One apartment can be a family stay, a long stay or an outdoor-living stay. It is still one unit with one availability, so when one product sells, the others update.
Booking portals keep receiving your categories, so you stay easy to compare there. Your own website sells what the portals can't show: the exact stay a guest is looking for. Booking direct finally means a better product, not just a lower price.
Categories sell the hotel's definition of premium. Dynamic Inventory sells the guest's.
The proof
At one vacation resort using GauVendi, where guests stay four nights on average, looking at bookings made since January:
13 room categories became 220 sellable products.
87 in every 100 bookings were for products defined by features or experiences.
48 in every 100 bookings were priced above the category equivalent, on average €150 more per stay, about €36 more per night.
Across all bookings: €72 more per stay on accommodation alone, €92 with extras.
It's the same rooms, with better products.
AI is only as smart as your labels
AI runs on chips, models and data. The chips and models are now available to everyone. The difference comes from data.
But data on its own does nothing. Give the best chef in the world a fridge full of unlabelled containers, and they can't cook anything good.
AI is the chef. Your data is the fridge. What's missing are the labels.
Palantir, one of the most valuable software companies in the world, built its business on this idea and calls those labels an "ontology". Its chief technology officer told investors that real value from AI is "only possible with Ontology."
A typical hotel fridge says Deluxe, Premium Deluxe, Superior. That tells an AI agent almost nothing. Dynamic Inventory gives it a language of experiences: what each room offers, who it suits and what it costs. Hotels with clear labels get recommended. Hotels with "Deluxe" get compared on price.
The data nobody else can collect
Once guests can choose, every choice teaches you something a category hotel never learns:
who pays extra for the balcony, and how much,
who lets the hotel choose the room in exchange for a lower price,
which name makes a guest pay more for the very same room.
Sold as "Superior Double", a top-floor room with a reading chair and an old-town view is one room among many. Sold as "Verena's favourite room", with a line from the owner about why she loves it, it has a story, and some guests will pay more for that story.
Nobody can predict this. Not your team, not a consultant, not an AI.
This knowledge only exists once guests actually buy, and they can only buy it if the product exists.
The AI will keep getting cheaper for everyone. This knowledge won't. It belongs to you.
Four checks before your next technology project
Define the difference. What can guests buy from you that they can't compare one-to-one elsewhere?
Give direct a real reason. A better product, not just a lower price.
Simplify the choice, not the product. Show each guest the few options that fit, in words they, and AI agents, understand.
Learn from what sells. Which products, features and names guests choose, and what they pay.
Build something worth buying
If the channel mix hasn't moved after your last project, the software isn't the problem. The product on the shelf is.
So ask yourself one question: What can a guest, or their AI agent, find on my website that they can't find anywhere else?
Want to know what your own rooms could sell? Contact us for more information.



Comments