Airbnb Last Minute Discount Strategy: When to Cut Rates

You have twelve nights still open inside the next week. Your base rate is $200. Airbnb keeps nudging you to switch on a last minute discount, your neighbours are all running one, and the obvious move is to knock 20 percent off and watch the calendar fill. Here is the problem with the obvious move. On a night that already had a 60 percent chance of selling at $200, a 20 percent discount has to lift your booking probability by fifteen full percentage points just to break even. Not to make money. To break even. Most hosts have never run that number, which is why most hosts are quietly paying for occupancy they already had.

This article does the arithmetic that the rest of the internet skips.

What an Airbnb Last Minute Discount Actually Does

A last minute discount is a standing rule, not a decision. That distinction is the whole article, so it is worth sitting with for a moment.

In your Airbnb pricing settings you choose a window between 1 and 28 days before check-in and attach a percentage or fixed reduction to it. From that moment on, every night that is still open when it enters the window gets the discount automatically. You are not choosing to discount a specific weak Tuesday. You are pre-committing to discount every night that happens to be open, including the ones that were about to sell at full price anyway.

Airbnb applies the reduction in a defined order. According to Airbnb’s own documentation on rule-sets, nightly and weekend pricing resolves first, then length-of-stay discounts, then early-bird and last minute discounts. Any additional promotion you layer on top calculates off the rule-set price rather than your original rate. A rule-set that drops a $120 night to $100, plus a 20 percent promotion, produces $80, not $76.

Two mechanical details are worth knowing before you touch anything.

• Last minute discounts are unavailable while Smart Pricing is switched on. Smart Pricing overrides rule-sets entirely, so you have to choose one system or the other. We have written separately about what Smart Pricing gets right and where it costs you money.
• At a discount of roughly 10 percent or more, Airbnb may show a struck-through original price in search results. That visual treatment is a real part of the value you are buying, and it is why a 9 percent discount and an 11 percent discount are not two points apart in effect.

Why this is suddenly a bigger decision than it used to be

Booking windows in North America have compressed hard. Reported figures vary by source and by market, but the direction is not in dispute: guests are booking closer to arrival than they were three years ago, urban markets are running median lead times in the high teens, and destination markets with strong seasonal peaks still hold windows two to three times longer. Airbnb itself leaned into the trend in its 2026 Summer Release by adding seasonal and last minute settings designed to help hosts fill late gaps.

More of your inventory now sells inside the discount window. Which means the standing rule you set once, and probably have not looked at since, is touching a much larger share of your revenue than it did when you set it.

The Break-Even Lift: The Only Formula That Matters

Every guide on this topic tells you to ask whether the night would have sold at full price. None of them tell you what to do with the answer. Here is the arithmetic.

Call p the probability that an open night sells at your full rate before check-in. Call d the discount you are considering, as a decimal. A discount is worth running only if it lifts your booking probability by at least:

Required lift = p × d / (1 − d)

The derivation is short. Expected revenue with no discount is p times your rate. Expected revenue with the discount is the new probability times your discounted rate. Setting those equal and solving gives a required post-discount conversion of p divided by (1 minus d). Subtract the original p and you have the lift you need to buy.

Numbers make it concrete. Here is what a 15 percent and a 25 percent discount each demand, expressed in percentage points of extra conversion.

Read the bottom right cell again. To justify a 25 percent discount on a night that already had an 80 percent chance of selling, you would need the discounted version of that night to convert 106.7 percent of the time. There is no such number. The discount cannot pay for itself at any conversion rate whatsoever.

That is not a judgement call or a matter of taste. It is a ceiling, and it generalises.

The Discount Ceiling Rule

Required conversion is p divided by (1 minus d). That expression exceeds 1 whenever p is greater than (1 minus d). Rearranged into something you can use on a Tuesday morning:

Your maximum arithmetically defensible discount is 100 percent minus the share of open nights that sell at full price in that window.

The ceiling column is the point of no return. The practical column is where the required lift stays small enough that a normal market can actually deliver it. Anything between the two is theoretically possible and operationally reckless.

Notice what the ceiling does as check-in approaches. A night still open at 25 days out, in a market where most demand has not arrived yet, has a high p and therefore a low ceiling. The same night still open at 3 days out has a low p, because the fact that it survived this long is itself evidence of weak demand, and therefore a high ceiling. This is why a tiered ladder outperforms a flat discount. Not because tiering feels sophisticated, but because the ceiling genuinely rises as the date approaches.

Worked Example: Twelve Open Nights, Two Discount Depths

Take a two-bedroom listing with a $200 base rate. On the first of the month you have twelve nights still open inside your seven day window. From two years of your own calendar history you know that roughly seven of any twelve such nights end up selling at full price, and five go dark.

Baseline, no discount: 7 nights × $200 = $1,400.

Scenario A: a 15 percent discount

Discounted rate is $170. The seven nights that would have sold still sell, now at the lower rate. Assume the discount and the struck-through price in search convert two of the five dark nights.

• 7 nights × $170 = $1,190
• 2 additional nights × $170 = $340
• Total: $1,530. Gain over baseline: $130. Nine nights sold.

Scenario B: a 25 percent discount

Discounted rate is $150. Deeper cut, better conversion. Assume it now rescues three of the five dark nights instead of two.

• 7 nights × $150 = $1,050
• 3 additional nights × $150 = $450
• Total: $1,500. Gain over baseline: $100. Ten nights sold.

The deeper discount sold one more night and earned $30 less. It also handed a $50 refund to each of seven guests who were going to book regardless, which is where the money went: $350 of giveaway against $450 of rescued revenue.

Run the break-even test on both and they pass. Full-price sell-through here is 7 of 12, or 58.3 percent. Scenario A needed a 10.3 point lift and got 16.7. Scenario B needed 19.4 and got 25.0. Both clear the bar, which is exactly the point worth being precise about: the break-even formula tells you whether a discount beats no discount, not which depth beats another. It is a floor test. Choosing between depths requires the full calculation above, and the full calculation frequently favours the shallower one.

This is the same structural trap as chasing occupancy at the expense of rate, which we unpacked in detail in our piece on ADR versus occupancy for Airbnb hosts. A fuller calendar and a smaller deposit are entirely compatible outcomes.

Building the Ladder From Your Own Data

You need one number you almost certainly do not have written down: for each lead-time band, what share of nights that were still open at the start of the band went on to sell at full price. Export twelve months of reservations, tag each booking with its lead time, and compare against the nights that stayed empty. An afternoon of work, and it replaces every generic percentage table on the internet with your actual market.

Here is what a ladder looks like once the ceiling rule is applied, using a $200 base rate and sell-through figures typical of a mid-sized US market with moderate compression. Substitute your own.

Look at the last column. The required lift sits between 2.9 and 4.4 points at every rung. That is the signature of a ladder built correctly: each tier is asking the market for roughly the same favour. A flat 20 percent discount across the whole 28 day window would demand a 16.3 point lift at the top rung and 3.0 points at the bottom, which is a rule that is simultaneously far too aggressive early and slightly too timid late.

The 22 to 28 day rung sets nothing at all, deliberately. In most markets that band is still inside the meat of the booking curve, so there is nothing to rescue yet. If your booking pace is genuinely running behind at 28 days out, the answer is a base rate problem, not a discount problem.

Setup checklist

• Export twelve months of reservations and compute full-price sell-through by lead-time band. Do this per season, not as a single annual average.
• Apply the ceiling rule to each band and set the discount well inside it, not at it.
• Confirm Smart Pricing is off, since it overrides rule-sets and blocks last minute discounts.
• Cross-check that the discounted rate at your deepest rung still clears your variable cost per night, then confirm it also clears the cannibalisation test above, which binds first.
• Exclude your known peak dates, festival weekends and holidays from the rule entirely, using a separate rule-set or blocked dates.
• Check the struck-through price actually appears in search at your 10 percent and deeper rungs. If it does not, the discount is buying you far less than you think.
• Re-derive the sell-through numbers every six months. Compression is still moving.
• Log what you changed and when, so that the revenue difference three months from now is attributable to something.

Your Floor Rate Is Not the Constraint You Think It Is

Almost every guide on this subject ends at the same place: work out your break-even cost per night and never discount below it. That advice is correct and nearly useless, because the floor is hardly ever what stops you.

Price out the marginal cost of one additional occupied night on a listing where the cleaning fee is charged separately and covers the cleaner in full. Linens and consumables, call it $12. Incremental utilities, $8. A wear and tear reserve, $10. Airbnb’s host service fee at 3 percent of the booking subtotal, which on a $150 night is $4.50. Total marginal cost: roughly $35.

So on pure cash terms, a $40 night beats a dark night. The floor is $35 on a listing with a $200 base rate. You will never get anywhere near it, because the cannibalisation arithmetic stops you at $150 or $170.

The floor only becomes the binding constraint in two situations. The first is when your cleaning fee does not cover your cleaner, which turns every marginal booking into a subsidised one and is a cleaning fee structure problem rather than a pricing one. The second is a single orphan night wedged between two bookings, where the night either sells or is worth precisely zero and the cannibalisation risk is close to nil. Orphan nights are the one case where the aggressive move is usually right, and they deserve their own treatment, which is why we gave them a dedicated article.

Where Last Minute Discounts Backfire

Peak dates inside the window

A standing rule does not know that the second weekend in October is your town’s marathon. It discounts that Saturday alongside a dead Wednesday in February. On a compression date, full-price sell-through is often above 90 percent, which puts the ceiling under 10 percent and makes almost any discount value-destroying. Carve peak dates out explicitly.

Training your repeat guests to wait

This one is real and rarely quantified, so quantify your own exposure rather than accepting a number from a blog. If repeat and referred guests are 18 percent of your bookings, and a visible standing discount persuades half of them to delay until the window opens, you have permanently handed back roughly 1.8 percent of total booking revenue with nothing in return. Those guests were always going to book. Run that calculation with your own repeat rate before you set anything deeper than 15 percent.

Confusing a discount with a rate cut

A 10 percent last minute discount and a 10 percent base rate cut both display $180 on a $200 listing. They are not the same instrument. The discount touches only nights that reach the window still open, roughly a quarter to a third of inventory in a compressed market. The base cut touches everything, including the booking made 70 days out by a guest who never saw a competing price. If your problem is that the whole calendar is soft, a discount is the wrong tool and you should be looking at your base rate and your minimum stay settings first.

Using it to paper over a positioning problem

If your listing is already priced below its comp set and still not selling, the issue is photography, reviews, amenities or search placement. Discounting a listing that is already cheap tells the market something you do not want it to hear, and it does nothing about the actual cause.

Three Levers That Beat Discounting

Before you widen a discount, exhaust the changes that cost you nothing per booking.

Minimum stay. Late bookers overwhelmingly want one and two night stays. A three night minimum sitting on a Tuesday four days out is not protecting your rate, it is hiding your listing from the only people still shopping. A lead-time ladder on minimum stay, loosening as the date approaches, converts more nights than a price cut and costs nothing on the nights that were selling anyway.

Advance notice. A two day preparation requirement quietly deletes same day and next day bookings from your calendar. In a market where a meaningful share of reservations now land inside 48 hours, that setting can be more expensive than any discount you are contemplating. Shorten it to same day with a cutoff time your cleaner can actually meet.

Instant Book. Late bookers are making a decision in one sitting, often on a phone, often comparing three tabs. A request-to-book listing asks them to wait for a stranger to reply. Many will not.

All three change conversion without touching the rate that your full-price bookers pay. That asymmetry is the whole reason to try them first.

Frequently Asked Questions

Conclusion

A last minute discount is not a favour you do for guests who were going to book anyway. It is a bet that a specific price reduction will buy a specific amount of extra conversion, and that bet has a break-even you can compute in about a minute. Required lift is p times d over one minus d. Your ceiling is 100 percent minus your full-price sell-through. Above the ceiling, the discount cannot win at any conversion rate, which makes the choice arithmetic rather than instinct.

Get the sell-through numbers out of your own calendar. Build the ladder so that each rung asks the market for roughly the same lift. Carve out your peak dates. Then check the exits before you widen anything: minimum stay, advance notice and Instant Book will often convert the same nights without charging your full-price guests for the privilege.

If you want a second set of eyes on your ladder, your base rate, or the dynamic pricing strategy underneath both, get in touch and we will look at your actual numbers.