Off-Season: Hold Your Rate or Drop It?
In a soft market, hold. Cornell measured lodging price elasticity at −0.14 — discounting rarely creates demand. Drop only close to arrival, and fence it.
In a soft off-season, hold your published rate. Cutting it rarely creates new demand — Cornell measured the price elasticity of lodging demand at −0.14, meaning a 1% rate cut moves demand about 0.14%. The genuine exception is distressed inventory close to arrival: a specific night that will otherwise go empty, where anything above your variable cost is money you would not have had. Discount that night behind a fence — advance purchase, non-refundable, minimum stay, or a package — so the low rate cannot leak into your high season.
The question that never gets answered
Every low-season thread goes the same way. Somebody asks whether to drop rates in January, and within four replies it is about closing a floor, rotating rooms to keep them aired, shutting the restaurant, and putting the housekeepers on reduced hours.
All of that is real and useful. None of it is pricing. Here is the pricing half.
First, work out which kind of quiet you are in
The word "off-season" covers two situations that need opposite responses.
| Soft market | Dead market | |
|---|---|---|
| What it looks like | Bookings still arrive, just fewer and later | Almost no search, almost no bookings |
| Your comp set | Also quiet, but transacting | Also empty, some closed entirely |
| Cause | Fewer trips being taken to your area | The demand driver is gone — no season, no events, no business travel |
| Does price move volume? | Slightly. Elasticity is low but not zero | Barely. Nobody is choosing between hotels |
| Right move | Hold, work channels and length of stay | Hold the rate, cut cost. Consider closing floors |
Check your comp set before deciding. If competitors are selling at all, you are in a soft market and price positioning still matters. If the whole street is empty, discounting is a transfer from your P&L to nobody.
Why discounting into a soft market usually fails
Two Cornell datasets say the same thing from different directions.
Demand does not respond much to price. Canina & Carvell (Cornell, 2003) studied 480 hotels over eleven years and put the price elasticity of lodging demand at −0.14. A 1% cut in rate produced roughly a 0.14% increase in demand. What actually moved demand was income and GDP — the economy, not your rate. Nobody takes a trip they were not going to take because your room is $15 cheaper.
Where discounting does move volume, it moves it at a loss. Enz, Canina & Lomanno analyzed 67,008 hotel observations from 2001 to 2007 (Competitive Hotel Pricing in Uncertain Times, Cornell Hospitality Report Vol. 9 No. 10, 2009), comparing hotels against their own comp sets:
| Rate vs comp set | Occupancy vs comp set | RevPAR vs comp set |
|---|---|---|
| 20–30% lower | +15.2% | −12.2% |
| 10–15% lower | +4.5% | −8.3% |
| 5–10% higher | −0.5% | +6.7% |
| 20–30% higher | −10.9% | +10.3% |
For independent hotels specifically, pricing 20–30% below the comp set showed RevPAR 16.31% lower; pricing 5–10% above showed RevPAR 4.12% higher.
That study period included the 2001–2003 downturn — this is not boom-time data. And the authors state the caveat themselves: it shows correlation, not causation. Hotels that discount heavily may already have a problem that caused the discount.
Note also what the occupancy column says. Discounting does fill rooms — you will see the occupancy move, which is exactly why it feels like it worked. The revenue per available room is where it shows up.
When a lower rate genuinely is right
Holding rate is the default, not a religion. Three situations where a lower number is the correct answer.
Perishable inventory. A room night is gone at midnight and cannot be recovered. Unlike stock in a warehouse, unsold inventory has no salvage value. That is a real argument for taking a lower price rather than nothing.
Distressed inventory close to arrival. Inside about seven days, with a specific night you can see will not fill, most of the demand for that date has already booked. Holding rate no longer protects a future booking, because there is barely any future left for that date. The Cloudbeds 2026 State of Independent Hotels panel (90 million bookings, 180 countries, 2025 data) puts the independent booking window at 40 days; SiteMinder's Hotel Booking Trends panel (2025 data) puts the global window at 32.15 days. The two panels disagree — but both say the great majority of bookings for a date have landed by the last week.
Covering variable cost on a night that would otherwise be empty. If your cost per occupied room — housekeeping, linen, amenities, in-room utilities, card fee, commission — is $25, a room sold at $70 contributes $45 toward fixed costs. Zero contributes nothing. That is a genuine gain.
The test for all three: is this a specific night close in, or is this a season? Discounting a distressed Tuesday is revenue management. Discounting January is repositioning.
Fence the discount so it cannot leak
The danger is not the money you lose on the discounted night. It is that a low public rate becomes your anchor — visible to guests, to repeat customers, to your comp set, and to the OTA algorithms that rank you.
A fence is a condition the guest has to accept to get the lower rate. It separates the price-sensitive guest from the one who would have paid full rate.
| Fence | How it works | What it protects |
|---|---|---|
| Advance purchase | Book 21+ days out to get the rate | Late, high-value bookers still pay BAR |
| Non-refundable | Lower rate, no cancellation | You keep the revenue; cancellations do not eat the gain |
| Minimum length of stay | Rate applies on 2+ or 3+ nights | Raises total revenue per arrival, spreads cleaning cost |
| Package value-add | Same rate, breakfast or parking included | Headline rate never moves — you give value, not price |
| Midweek-only | Rate available Sunday–Thursday | Your strongest nights stay protected |
| Closed-user-group / member rate | Visible only to signed-in or emailed guests | Never enters your public rate history |
The package option deserves the most attention in an off-season. Adding $18 of breakfast to a $120 room is a better trade than dropping to $99, because the perceived value moves more than the cost does, and your rate integrity survives into spring.
One more reason to fence: non-refundable rates matter more than they look. The Cloudbeds panel put the OTA cancellation rate at 21.8% against 10.6% direct. An unfenced OTA discount can be cancelled back out from under you.
Decision table: what to do by market condition
| Market condition | Signal you can check | What to do |
|---|---|---|
| Soft market, 60+ days out | Comp set transacting, your pickup slow but present | Hold. Too early — most bookings have not happened yet |
| Soft market, 14–30 days out | Comp set still selling, you are behind them | Hold BAR. Open a fenced advance-purchase or LOS rate |
| Soft market, inside 7 days | Specific nights clearly will not fill | Drop those nights only, fenced and non-refundable, above variable cost |
| Dead market, whole season | Comp set empty or closed, no search demand | Hold rate, cut cost. Close floors, reduce roster. Price will not save it |
| Comp set undercutting hard | Several competitors 20%+ below you | Hold. Cornell says the discounters are buying occupancy with RevPAR |
| Comp set selling out | Sold-out flags appearing near you | Raise. Even in low season, a compressed night is a compressed night |
| One-off event in a dead week | Concert, conference, sports fixture | Price the night, not the season. Off-season rates should never apply on it |
| You are below your rate floor | Rate under CPOR plus commission | Never sell. An empty room costs you less than a room sold below cost |
The anchoring damage of a published low BAR
Your BAR — best available rate — is the standard public rate anyone can book without conditions. Cutting it does four things that outlast the off-season.
It becomes the reference price for anyone who looked at you in January and comes back in June. It sits in your OTA rate history, where returning guests see the drop. It signals to your comp set that you are the one who blinks, which invites them to follow you down. And it re-sorts you into a lower rate band on OTA filters, where you compete with a cheaper tier of hotel.
Fenced rates do none of that. The condition is doing the work, so the headline number survives.
Common mistakes
Publishing a "winter rate" for the whole season. A season is not a demand signal. There will be sold-out nights inside your off-season — a conference, a wedding, a fixture — and a flat seasonal rate will sell them cheap.
Discounting at 60 days out. Both booking-window panels say most bookings arrive inside six weeks. An empty February calendar in December is not evidence of anything.
Matching a competitor's discount. You do not know whether their rate is distressed inventory, a fenced package, or a mistake. Cornell's numbers say the person cutting hardest is usually the one giving up the most RevPAR.
Cutting BAR when a fenced rate would do. Advance purchase, LOS and packages get you most of the volume with none of the anchoring damage.
Forgetting commission in the floor. A $70 rate at 18% commission nets $57.40. Check the net, not the headline, before deciding it beats an empty room.
Assuming low season means low ADR is fine. CoStar reported US full-year 2025 occupancy at 62.3%, ADR $160.54 and RevPAR $100.02 — the first annual occupancy and RevPAR decline since 2020. A soft national market is exactly when rate discipline separates properties.
The bottom line
Check whether your market is soft or dead before touching a number. In a soft market, hold your published rate — elasticity is too low for a discount to create demand, and Cornell's comparison data shows the discounters buying occupancy at the cost of RevPAR. Drop only for specific nights close to arrival that will otherwise go empty, only above your variable cost, and only behind a fence. Never let an off-season rate become your public anchor for next season.
The work that makes this manageable is checking, every day, which specific nights are actually distressed rather than just early. Our occupancy forecasting shows expected fill night by night, so you can tell a Tuesday that will not fill from a Tuesday that simply has not filled yet.