Booking lead time math: why an empty calendar isn't an emergency

Two summers ago I looked at my July calendar on the 20th of May and saw four nights booked out of thirty-one. I panicked, dropped the nightly rate $25, and felt responsible for doing it. July finished at 73% occupancy — exactly where it lands every year — and almost all of it booked inside the final three weeks, at a rate I'd just kneecapped. I gave roughly $300 to travellers who would have paid full price, in exchange for nothing. The calendar wasn't empty. It was on schedule. I just didn't know what my own schedule looked like.

What booking lead time actually is

Lead time is one subtraction: check-in date minus booking date. A guest who books on June 1 for a June 5 arrival has a 4-day lead time. One who books in February for August has a 180-day lead time. Your listing has a distribution of these — some last-minute, some far out — and the shape of that distribution is the single most useful thing you can know about your own demand.

The number most hosts quote is the median lead time. Mine is about 9 days for a city one-bedroom; a beach house that fills with week-long family trips might sit at 60. But the median hides the part that matters, which is the pace: not "when do bookings arrive on average" but "how much of my final occupancy is already on the books at each point before check-in." That curve is what tells you whether today's calendar is healthy or sick.

Here's the trap in one sentence. A calendar that is 30% full looks identical whether you're cruising toward a great month or heading for a disaster — because "30% full" is a snapshot, and a snapshot has no idea how far away check-in is. The same 30% is triumphant at T-7 and alarming at T-7 only if you know where 30% should be at T-7 for your listing. Without the curve, you're reading a thermometer with no numbers on it.

Read your own pickup curve

A pickup curve is the answer to: "at X days before check-in, what fraction of my eventual bookings do I already have?" Build it from your own history and it turns the scary snapshot into a number you can judge.

Here's the curve for a real-shaped city one-bedroom over a typical 30-night month that finishes at 22 nights booked (73% occupancy):

Read the right-hand column, not the middle one. At T-30 this listing has 9 nights booked — a calendar that looks 30% full and feels thin. But 9 nights is 41% of where the month ends up. It is on pace. The same listing at T-14 with 16 nights looks healthier, and it is, but it's actually the identical trajectory — it was always going to be there. The panic at T-30 and the relief at T-14 are the same month seen at two points on one curve.

To build yours: pull your last twelve months of reservations, and for each completed month, record how many nights were on the books 30, 14, and 7 days before each night's date. Average across months. You'll get a curve like the one above, and it will be specific to your listing, your market, and your channel mix — which is the whole point. Generic "the average host books X days out" numbers are useless because your listing isn't average. If your bookings live in one calendar across platforms, this is a five-minute export; if they're scattered across three platform dashboards, pulling every reservation into one place is the prerequisite to even seeing the curve.

The panic-discount trap, with the math

Walk through what my $25 cut actually cost. Thirty-one nights in July, ADR $150, and history says the month finishes at 22 booked. At T-40 I had 6 nights on the books — the early-planner tail. Twenty-five nights still open.

The best case for the discount still loses to doing nothing. Cutting the rate didn't summon demand that wasn't coming — this market books late regardless of price, so the cut just re-priced bookings I'd have won anyway. In the most charitable version, I bought 2 extra nights of occupancy for $150 of lost margin on the other 16. That's paying $75 a night to fill a night that, even sold, nets less than the margin I burned to get it. The honest version is the first row against the second: $400 set on fire for the feeling of having done something.

This is the asymmetry to internalise. Raising occupancy from 71% to 91% feels like progress and can quietly lower what you earn per available night — the same trap I unpack in the RevPAR post. A late-booking market punishes early discounting specifically because the discount lands on inventory that wasn't at risk.

When an empty calendar is the signal

None of this means never discount. It means discount against your curve, not against your nerves. The pace test is one comparison: take where you are now, find that lead-time bucket on your curve, and ask whether you're above or below the share you'd expect.

You're at T-14 with 9 nights booked. Your curve says T-14 should be at 16 — 73% of final. Nine nights is 41% of a 22-night month, which is the T-30 number. You are a full two weeks behind your own pace. That is a real signal, and it earns a real response: a targeted drop on the specific open nights, a relaxed minimum stay to catch gap-fillers, or a last-minute discount that platforms will actively surface. The difference between this and my July mistake is that here the curve is telling you the month is genuinely underbooked, not just early.

The discipline is boring and it works: compare to your curve before every pricing move. If you're on or above pace, the open nights are not an emergency — they're the future bookings doing exactly what they always do, arriving late. If you're below pace, act, and act on the open nights specifically rather than re-pricing the whole month.

What lead time tells you besides price

Pace is the headline, but the lead-time distribution drives three other decisions hosts usually make on vibes.

Cancellation risk scales with lead time. A booking made 120 days out has 120 days for plans to fall apart; a booking made 5 days out is people who've already bought flights. A calendar that's "full" four months ahead is softer than one full three weeks ahead, even at the same occupancy percentage. Don't treat far-out bookings as money in the bank, and think twice before blocking around them or turning down nearer, firmer demand to protect them.

Minimum stays can flex by lead time. Far-out demand skews toward planned trips and longer stays; last-minute demand skews toward one- and two-night gap-fills. That's an argument for a longer minimum on the distant months and a relaxed one as check-in approaches — exactly the gap-night recovery logic in the orphan-night post. Setting a 3-night minimum eight months out costs you nothing; keeping it 6 days out strands the single nights nobody else will take.

Your calendar window should match your curve. If 95% of your bookings land inside 90 days, opening availability 24 months out mostly exposes you to low-commitment far-future bookings at prices you'll regret by the time they arrive. Open roughly 12 months, and price the distant tail deliberately high — those bookings are optional, so only take them if they pay a premium for locking you up early.

Airbnb and Booking.com don't book on the same clock

Lead time isn't just a property trait — it's a channel trait. Booking.com skews shorter: more mobile, more last-minute, more one- and two-night business and transit stays. Airbnb skews longer: more vacation planning, more groups, more "we're booking the summer in March." The gap isn't small; in many city markets a Booking.com-heavy listing fills a week or two later than the same unit sold mostly on Airbnb.

The practical consequence: your pace threshold has to be channel-weighted. If Booking.com is half your mix, your "normal" T-30 number is lower than a pure-Airbnb host's, and your panic point should sit later. Build the curve from your reservations across your channels, not from a platform's blog post about average traveller behaviour. A listing fed by one merged calendar makes this trivial; three separate dashboards make you eyeball it, and eyeballing is how the May-20th panic happens.

One opinionated take

Stop looking at how full your calendar is. Start looking at how full it is for this many days before check-in, against the only benchmark that means anything — your own history. Occupancy-at-a-glance is the metric that makes hosts discount in May for a July that was always going to sell. The pickup curve is unglamorous, it lives in a spreadsheet, and it will save you more money this year than any pricing tool you can buy, because it stops you from solving a problem you don't have.