Pricing rhythm
How Often Should a Hotel Change Its Rates? A Mechanical Answer, Not a Rule of Thumb
Whether you’re running 40 rooms off Sukhumvit in Bangkok or 35 villas on Koh Samui, you’ve probably asked yourself some version of this question: how often should we actually be changing our rates? Once a week feels manageable alongside everything else a general manager or owner-operator is already doing. Once a day feels excessive for a property without a dedicated revenue manager.
The honest answer isn’t a frequency at all. It’s what’s moving underneath your rate between the day you set it and the day a guest actually books — and for most independent hotels and resorts of this size, quite a lot moves in a week.
What actually changes between Tuesday and Thursday
Start with the things that shift daily, not weekly.
Your compset moves. Compset — short for competitive set — is the small group of comparable properties you price against. A boutique hotel in Bangkok might track eight to twelve nearby properties; a resort on Koh Samui might have only three or four genuine comparables within reach. Either way, that set doesn’t hold still. A comparable property drops its rate for a specific date because it’s fallen behind on pickup, or raises it because a room category just sold out. If you set your rate on Monday and don’t look again until next Monday, you’re static against a moving target for six days.
Pickup pace changes. Pickup is simply the rate at which rooms for a given date are being booked. Faster-than-usual pickup for a Friday three weeks out is a signal you can hold firm or raise the rate. Slower-than-usual pickup is a signal to move before the date gets close and the room sits empty. Pickup shifts day to day — sometimes hour to hour around a booking-engine promotion or an OTA placement change — not on a weekly cycle.
An event lands with no warning. A conference gets added to a Bangkok hotel’s calendar. A public holiday shifts. A flight route change alters how many overseas visitors are due into Phuket in six weeks. None of these wait for your review day.
Lead time keeps changing what “close to the date” means. Lead time is the gap between booking and arrival. For a city hotel this is often short — many bookings land inside two to three weeks of the stay. For an island resort it’s often long — three to six months isn’t unusual, especially for peak-season dates booked from overseas. A rate that’s a few days stale matters differently depending on how much runway you have left.
Where a city hotel and an island resort diverge
This is the part that’s easy to skip and shouldn’t be.
For a Bangkok or Chiang Mai boutique hotel, the short booking window means staleness compounds fast. If most of your bookings land inside two to three weeks of arrival, a rate you set on Monday for a date three weeks out is already operating on borrowed time by Thursday — a third of that booking window has passed on an unreviewed rate. Weekday/weekend swings, corporate demand, and trade-fair spikes all move quickly and locally, and a weekly review means you’re often reacting to a compset picture that’s several days old for most of the week.
For a Phuket, Koh Samui, Koh Phangan or Krabi resort, the long booking window changes the arithmetic without removing the problem. A single day’s staleness on a date six months out matters less than it does for a city hotel three weeks out — there’s time to correct it. But the picture inverts as the date approaches, and it inverts hard during shoulder season and monsoon months, when demand is thin, weather-driven, and volatile. A rate that’s wrong by a meaningful margin during a low-occupancy week isn’t diluted across dozens of comparable bookings the way it might be in high season — it’s one of a handful of decisions that week, and it’s more visible, not less.
Neither property is well served by “review once a week.” The city hotel is behind because dates arrive too fast for weekly review to keep pace. The resort is behind because, close to a low-demand date, a week without a look is a week without a correction on the days that actually needed one.
What a week of staleness actually costs
No industry figure can tell you what this is worth to your property — only your own numbers can, so use them here.
Case one: rooms sold too cheap. Say a city hotel has six rooms still open for a Friday night three weeks out. Midweek, pickup for that date runs noticeably faster than usual — two rooms book in a day where normally it’s none or one. That’s a signal to raise the rate. If the rate isn’t revisited until the next scheduled review, those six rooms may sell at the original price instead of a corrected one. If the gap between the two is 300 THB a night — put in your own number — that’s 1,800 THB understated on a single date, from a single week’s delay, in one direction only.
Case two: rooms left empty. Now the resort case, during a shoulder-season week with three rooms unsold. Pickup has gone quiet, but the rate wasn’t adjusted down, so a compset property with a lower rate wins the booking instead. Guests don’t complain about your rate — they simply book elsewhere, and you never see the booking to know it existed. Assume the same 300–400 THB gap, and the property doesn’t just lose margin, it loses the booking entirely. And that’s only the nights you’d have noticed to check; dates you never looked at again aren’t in this count at all.
Run this on your own occupancy and rate for a slow week and a busy one, for a near-term date and a far-out one. The point isn’t the specific number — it’s that the gap exists in both directions, on ordinary weeks, not just headline ones.
Why weekly review is already behind
None of this requires prices to change constantly to make the point. It requires the information behind the price to be checked constantly, because the compset, the pickup pace, and the calendar don’t wait for a scheduled review. A system that checks and republishes rates only once a week is, by the end of that week, working from information that’s up to six days old on every date it hasn’t looked at since. That’s the argument for continuous repricing — an AI agent that never sleeps, checking and adjusting rates as pickup and compset data change, rather than reconstructing the picture from scratch once a week.
This is also where operator judgement stays essential, not optional. Continuous repricing can’t see the corporate client you’re deliberately holding a room for, or the local event your compset hasn’t priced in yet. It should be able to be overridden — a rate a system proposes is a starting point a human can adjust, not a final word.
One more thing worth saying plainly: the system doing this runs on hardware at the property rather than sending guest data to an outside cloud service. That’s a real architectural difference from most cloud-hosted revenue tools, not a compliance claim — Thailand’s PDPA doesn’t require on-premise hosting, and we wouldn’t tell you it did.
What engaging looks like
If this matches what you’re seeing in your own booking pace, the first conversation is short: what your compset looks like, how your booking window behaves across the year, and what your current rate-review process actually involves day to day. From there, a pilot means connecting to your property management system and channel manager and watching the system propose rates alongside whatever you’re doing today, before anything changes automatically. No commercial terms are set until that conversation has happened.
Frequently asked questions
Isn’t changing rates that often going to upset guests who already booked?
No — rate changes apply to unsold inventory going forward, not to guests who’ve already booked. Existing reservations keep the rate they booked at; only the remaining open rooms for a given date are affected by a new rate.
We only have one person handling rates part-time. Doesn’t more frequent review just mean more work for them?
That’s the point of automating the checking, not the deciding. A system that continuously checks compset and pickup and proposes a rate removes the manual labour of hourly or daily review; the person still approves or overrides, but isn’t the one doing the checking.
Our booking window is 4-6 months out for peak season. Does any of this matter that far ahead?
Less on any single far-out date, more as the date gets closer. Far-out dates can tolerate a slower review cycle; the same date inside a month, especially during shoulder or low season, behaves like the short-lead-time case and benefits from more frequent attention.
Does automated repricing mean rates change unpredictably, making it hard to plan cash flow?
Rate changes are driven by pickup pace and compset movement, both of which are visible and explainable at any point — you can always see why a rate moved. It isn’t randomness; it’s the same logic a revenue manager would apply, run more often than a person can manage manually.
What happens if the system gets a rate wrong for a specific date?
It can be overridden. The system doesn’t have visibility into things like a held room for a regular client or a local event not yet reflected in compset data, so operator override on specific dates is expected, not a failure case.