Hotel operators have spent the better part of a decade buying software on the promise that it will grow revenue, and a large share of that spend has produced dashboards instead of dollars. The technology works. The assumption underneath it, that installing a system is the same as changing how a property sells, does not. Marta Penda, who works in travel technology, frames the problem in terms most vendors avoid: software does not manufacture guests. “Technology doesn’t create demand,” she says. “What it does is remove the friction between the demand a hotel already has and the revenue it captures.” That distinction sounds modest until it is applied to a profit and loss, because it reroutes the entire conversation away from feature lists and toward the specific places where money currently leaks out of a property before it reaches the bank.
Where The Money Moves
Penda identifies four mechanisms, and none of them involve filling rooms that were never going to sell. Capturing more direct bookings lowers acquisition cost on every reservation. Pricing closer to real demand replaces last month’s assumptions with this morning’s. Selling more per guest through upgrades, add-ons, and pre-arrival offers raises the value of a booking already made. And freeing staff time from administrative work puts people back into selling. Each one attacks margin rather than volume, which is why the standard scorecard misses them. “It’s not RevPAR alone,” Penda says. “It’s net revenue after distribution cost too.”
That second measure is the one that should make owners uncomfortable. A property can post a strong RevPAR figure while quietly surrendering a growing share of it to distribution, and the headline metric will never flag the problem. Penda’s most pointed observation concerns the economics of loyalty: a repeat guest booked direct is the most profitable guest a hotel will ever have, “as long as you haven’t spent a lot on digital marketing to get that guest.” The caveat matters more than the claim. Direct booking is not automatically cheap. It is cheap when the guest relationship, not the advertising auction, is doing the work. Pre-arrival upselling sits in the same category, generating revenue with almost no added cost, which is precisely why leaving that capability dormant is such an expensive form of inaction.
The 90 Days Nobody Owns
Ask where the industry’s real weakness lies and the expected answer involves missing features or immature products. Penda rejects that framing outright. “The gap isn’t in the product roadmap,” she says. “It sits in the 90 days after go-live.” The capability exists in most properties. What goes missing is the follow-through, and it goes missing in a way that generates no alarm, no error message, and no line item. “Maybe nobody notices that the upsell engine was never switched on or that rates haven’t been updated in hours.”
The structural cause is a misalignment of purpose between the two parties signing the contract. “Vendors sell features and hoteliers need outcomes,” Penda says, and the commercial relationship, as commonly constructed, rewards the first while assuming the second will follow. Her remedy shifts work to before the signature rather than after it: agree on the revenue baseline before the contract is signed, define success in the hotel’s own numbers, and then, the part she flags as most important, stay accountable for that outcome. Technical accuracy is the wrong success criterion. “The best tech partners should change habits.” Hoteliers who accept that standard will find their vendor shortlist getting shorter, which is the point. A partner unwilling to be measured against the hotel’s own revenue figures is selling a licence, not a result.
The same logic exposes why unified systems matter more than their reputation suggests. When a property management system, channel manager, and booking engine share one source of truth, inventory errors and rate disparity disappear. When they do not, those errors persist silently across channels. As Penda puts it, they cost hotels more than they realize, which is another way of saying the losses never appear as losses.
Automate The Pricing, Keep The Context
The debate over AI-driven dynamic pricing tends to be posed as a choice between machine and human, and Penda declines the premise. Her position on the division of labour is unusually direct: most daily pricing decisions should be automated. No human can reprice every room type across every channel several times a day, and anyone who can is expensive. Automation also removes a category of error that manual repricing quietly generates. The machine is simply better at pricing thousands of dates, reacting in real time, and spotting patterns in data.
What the machine cannot see is everything outside the dataset. “Algorithms only know what’s in the data. They don’t know that a major client is negotiating a group block, that a competitor just renovated, or that the owner wants to reposition the brand.” That context is human, but Penda’s instruction is to feed it into the system rather than hold it back as justification for manual override. The commercial skill being redefined here is not price-setting. It is knowing which facts the algorithm lacks and getting them in front of it.
The next 12 to 18 months put a deadline on all of this. Penda expects AI assistants to become the main booking channel, and her preparation checklist is about fundamentals. Rates, availability, room descriptions, and policies must be structured and consistent everywhere. Systems must be open and connected, able to share real-time data through integration. And the direct offer has to stand up on its own. “If AI is comparing options, a clear reason to book direct matters.” Properties with messy data are about to discover that an assistant cannot recommend what it cannot understand.
Follow Marta Penda on LinkedIn for more insights on travel technology, hotel revenue strategy, and direct booking performance.
