Dynamic Pricing for Hotels: How AI Sets the Right Rate Every Night

After more than twenty years working in hotel revenue management and distribution across Thailand and Southeast Asia, I can tell you the single most expensive habit in this industry: setting a rate in January and leaving it there until something forces a change. It happens in properties of every size, from boutique resorts in Rawai to multi-property groups in Bangkok. A rate gets set, it feels reasonable, and then the market moves underneath it for months while nobody adjusts.

Dynamic pricing is the discipline that fixes this. And in 2026, it is no longer a manual spreadsheet exercise reserved for hotels with a dedicated revenue team. AI now does the heavy lifting, watching demand signals around the clock and adjusting rates with a consistency no human can sustain. But the technology is only as good as the thinking behind it, so it’s worth understanding what dynamic pricing actually is before you trust a system to run it.

Static Pricing Is Quietly Costing You Money

A static rate is a rate that doesn’t respond to demand. It’s the same on a dead Tuesday in low season as it is on a peak Saturday when half of Phuket is sold out. On the quiet night, you’re priced too high and the room goes empty. On the busy night, you’re priced too low and you’ve handed money to a guest who would happily have paid more.

Both of those are losses, and the second one hurts more than most owners realise. An empty room is an obvious loss. An underpriced room that sells looks like a win on the surface, which is exactly why it goes unnoticed. You only see the revenue you captured, never the revenue you left on the table.

Dynamic pricing closes both gaps. The rate moves up as demand builds and eases down only when the data genuinely calls for it. Done properly, it lifts your average daily rate (ADR, the average price you achieve per occupied room) without you having to chase occupancy through discounting. Across our active full-service clients, intelligent yield management of this kind delivers an average ADR improvement of around 12% within the first year, and it does so through better timing, not lower prices.

What “Demand-Driven” Actually Means

The core idea behind dynamic pricing is simple to state and harder to execute. Your rate should reflect demand, not the calendar.

In practice, an AI-driven system watches a handful of signals continuously. It tracks your occupancy pace, meaning how quickly rooms are selling for a given date compared with how they sold for the same date in previous years. It watches the booking window, which is how far in advance guests are reserving. It monitors competitor positioning, so you’re not sitting 30% below the property next door when demand is strong. And it reads broader market signals, from flight searches to local events, that tell it when a date is heating up.

When pace runs ahead of the historical baseline for a date, the system lifts your rate and pulls back the early-bird discounts, because demand is clearly there and you don’t need to bribe anyone to book. When pace lags well behind with only a week or two to arrival, it doesn’t immediately slash the rate. Instead it looks at whether the problem is genuinely a lack of demand or simply a conversion issue, which is a distinction most manual pricing decisions skip entirely.

That distinction matters more than almost anything else in revenue management, so it deserves its own section.

The Mistake That Destroys ADR: Discounting First

When occupancy looks soft, the instinct is almost always to drop the price. It feels like the responsible thing to do. In most cases, it’s the wrong lever.

Demand is not linear. At low occupancy, price elasticity is high, which means a price drop can genuinely stimulate bookings. At high occupancy, elasticity is low, so raising the rate barely dents demand at all. A good pricing system understands this curve and reasons about where you sit on it before touching the rate. A panicked owner staring at a half-empty calendar does not.

The bigger problem with reflexive discounting is what it teaches your market. Once guests learn that your rates soften as the date approaches, they stop booking early and start waiting. You’ve trained your own demand to arrive late and cheap, and that’s very hard to undo. This is why our standard position is to add value before cutting rate wherever possible. A free breakfast, a room upgrade on availability, a late check-out, or an F&B credit will fill a soft date while protecting your headline rate and your brand. A 25% discount across every channel fills the same date while quietly resetting what your rooms are worth.

AI helps here precisely because it removes the emotion. It doesn’t panic at a quiet week. It diagnoses, then acts.

The Direct Channel Has to Win on Price

There’s one rule in dynamic pricing that overrides almost everything else, and it gets violated constantly: your direct booking rate must never be more expensive than your OTA rate for the same room on the same night.

If a guest finds your room cheaper on Booking.com or Agoda than on your own website, you’ve lost twice. You pay 15 to 25% commission on a booking you could have taken directly, and you teach the guest that the OTA is the smart place to book you. We call a rate parity violation of this kind a priority-zero issue, meaning it gets fixed before anything else, because it silently collapses your direct conversion and your paid advertising returns at the same time.

A well-configured pricing system protects this automatically. It maintains a 5 to 10% direct price advantage as standard, so your own channel is always the best place to book. That advantage is affordable precisely because a direct booking costs you far less to acquire than an OTA one, and it comes with something an OTA booking never gives you: the guest’s data, and the chance to win their next stay directly.

Where AI Genuinely Earns Its Place

I’m cautious about AI hype, and I’ve written before that practical automation beats theoretical intelligence every time. Dynamic pricing is one of the clearest cases where the technology earns its keep, for a straightforward reason. Pricing well requires watching many signals, across many dates and room types, continuously, and adjusting without emotion. That is precisely the kind of work humans do poorly and machines do well.

A revenue manager checking rates once a week will always be a step behind a system checking them constantly. The same logic that lets us run AI-led dynamic pricing for events and attractions applies directly to hotel rooms, where the demand curves are different but the discipline is the same. The point is not that the machine replaces judgement. It’s that the machine handles the relentless, around-the-clock monitoring so that human judgement can be applied where it actually matters: strategy, positioning, and the calls that genuinely need a person.

The Practical Starting Point

If your rates today are largely static, you don’t need to leap straight to fully automated pricing to see a result. The first step is simply to make your rate responsive: tied to pace, segmented by room type and booking window, and protected by a consistent direct price advantage. The automation then makes that discipline sustainable, applying it every night across every date without a person having to remember to.

What you should not do is treat dynamic pricing as a discounting engine. Used that way, it just helps you give margin away faster. Used properly, it’s the difference between a property that prices on the calendar and one that prices on demand, and over a full year that difference is substantial.

At The Percentage Company, dynamic pricing isn’t a standalone feature we bolt on. It sits inside a connected revenue system, linked to your booking engine, your channel mix, and your guest data, so that every rate decision protects your direct share and your margin at the same time. If your rates haven’t moved in months and you suspect that’s costing you, we’d be glad to take a look and show you where the gaps are.