Book a domestic vacation package roughly four to eight weeks before departure, and an international one three to five months ahead. Those ranges are where the guidance from Google Flights, Expedia and Going overlaps. The complication specific to packages is that the two halves move in opposite directions: airfare tends to reach its low point around five to six weeks out, while Hotels.com's 2026 Hotel Price Index puts the hotel sweet spot at just 8 to 14 days before travel. A package forces you to buy both at the same moment, so you cannot optimise both — and the flight, being the less flexible and more steeply rising component, is the one to time.
A package makes you buy both at the same moment. That is the whole problem, and almost nobody writing about booking windows acknowledges it exists.
No, and they are close to opposites. Airfare generally falls until roughly five to six weeks before departure and then climbs steeply as the cheap fare buckets sell out. Hotel rates frequently do the reverse: Hotels.com's 2026 Hotel Price Index identifies 8 to 14 days before travel as the value sweet spot, and NerdWallet's analysis of more than 2,500 room rates found that booking 15 days out beat booking four months out 66% of the time, with rooms averaging 13% less.
Illustrative shapes drawn from the published sources cited throughout this page, not from a price study of our own. The directions are what the data supports; the exact curvature is a drawing. Sources: Google Flights historical fare analysis as reported by The Points Guy, May 2026; Expedia 2026 Air Hacks Report; Hotels.com 2026 Hotel Price Index; NerdWallet hotel rate analysis; Going booking-window guidance.
Airlines sell seats in fare buckets, each with a fixed allocation. The cheapest buckets are small and sell first. As departure approaches and the cheap allocations empty, what remains is the expensive inventory the airline has deliberately held back for travellers with no choice — business travellers, people attending funerals, anyone who has to be somewhere on Thursday.
That structure means a sold-out cheap fare is gone permanently. There is no mechanism by which it comes back. It is why the final three weeks reliably cost more and why "waiting for a last-minute flight deal" is a losing strategy on most routes.
A hotel room is the most perishable inventory in commerce. An unsold room tonight earns nothing tomorrow — it is not held back for a better customer, it simply expires. That gives hotels a strong incentive to discount as arrival approaches and rooms remain unsold.
The industry data reflects it. SiteMinder reported the global average hotel booking window at 32.15 days in 2025, far shorter than the flight equivalent, and Hotels.com places the value sweet spot at 8 to 14 days out. Hotels are running toward the deadline; airlines are running away from it.
Timing guidance is only useful against a real number. Price it now, then decide whether to book or wait.
Time the flight. Airfare rises more steeply and less recoverably near departure than hotel rates fall, and a sold-out fare bucket cannot be retrieved while an unsold room usually can. Booking in the flight window means accepting a hotel price slightly above its floor, which is the smaller loss. If your trip is genuinely flexible and the hotel is the dominant cost, book separately and time each half on its own schedule.
Here is the trap in plain terms. Suppose you take the hotel advice literally and wait until twelve days before departure to book your package. You may well capture a good room rate. You will almost certainly pay a punishing airfare, because at twelve days out you are buying from the buckets airlines reserve for people with no alternative.
Now suppose you take the flight advice and book at six weeks. You get a reasonable fare. Your hotel rate is above its theoretical floor, because the property has not yet reached the point of discounting unsold rooms.
The second mistake is much cheaper than the first. Airfare in the final fortnight can be multiples of the same seat six weeks earlier. Hotel rates rarely move by anything like that proportion, and the wholesale rate inside a package partly insulates you from the difference anyway.
So the asymmetry decides it. Book packages on flight timing.
All-inclusive resorts in peak season. These genuinely sell out, and the hotel-side logic reverses: waiting risks not getting the property at all. Treat peak-season all-inclusives like holiday flights and book three to five months ahead.
Trips where the hotel dominates the cost. If accommodation is seventy percent of your total and the flight is short-haul and cheap, the arithmetic tilts back toward hotel timing — and toward booking separately so you can time each half independently.
Genuinely flexible travellers. If you can move your dates by a week, that flexibility is worth more than any booking-window optimisation. Move the dates first, then worry about when to buy.
This is our reasoning from the published sources, not a measurement. We have not run a fare study and do not present one.
Enter your departure date and trip type. This applies the overlapping guidance from the published sources and gives you an actual window in calendar dates.
Pick a departure date
Four to eight weeks before departure for domestic trips and three to five months for international ones. Those are the ranges where the published guidance from Google Flights, Expedia and Going overlaps, which makes them the most defensible windows available given that the sources disagree with one another. Add roughly two months for peak season, and treat major holidays as needing three to five months domestically and four to ten months internationally.
Links to Expedia vacation packages. We may earn a commission on qualifying bookings.
Most travel sites pick whichever window suits their argument and present it as settled. It is not settled, and the disagreement is more useful than any single number.
| Source | Recommended window | What they measure |
|---|---|---|
| Google Flights historical fare analysis | 38 days out | Search queries across all carriers; cheap zone roughly 28–61 days |
| Expedia 2026 Air Hacks Report | 15–30 days | Expedia's own bookings, skewed toward leisure and package travel |
| Going | 1–3 months | Curated route set, hunting outlier fares rather than averages |
| Overlap of all three | 4–8 weeks | The range no source disagrees with |
| Source or region | Recommended window | Note |
|---|---|---|
| Going, general international | 2–8 months | Peak season needs the longer end |
| Expedia 2026, international economy | 31–45 days | Reported $190 saving versus six months ahead |
| Europe from the US | 3–5 months | Add two months for summer |
| Asia | 4–6 months | Longer for cherry blossom season |
| Latin America and Caribbean | 2–4 months | Longer over the winter holidays |
| Australia and New Zealand | 5–9 months | Longest windows of any major region |
Compiled from Going booking-window guidance, the Expedia 2026 Air Hacks Report, and aggregated fare analyses. Regional figures are published third-party guidance rather than our own measurements.
They are not measuring the same thing, and once you see that the disagreement stops being confusing.
Google Flights observes an enormous volume of searches across every carrier and route, including all the fares nobody bought. Its picture is the broadest and least filtered.
Expedia observes its own completed bookings. That population skews toward leisure travellers, package buyers and people who booked reasonably late — which mechanically pulls its optimal window closer to departure.
Going is not measuring averages at all. It watches a curated set of routes for unusually low fares. Its guidance describes where outliers appear, not where the mean sits.
So Expedia's shorter window and Google's longer one are not contradictory. They are answers to different questions.
Three things, and they are worth more than the disagreement.
Eleven months out is not cheapest. Airlines open booking around 330 days ahead, and every source agrees that the opening price is set before competition establishes the market rate. Booking the day the schedule opens is the most common expensive mistake among organised travellers.
Inside two weeks is expensive. On flights, near-universally. The cheap buckets are gone.
The middle holds the value. Every source puts its recommendation somewhere between roughly two weeks and five months, and the overlaps are narrow enough to be actionable.
That is why we publish an overlap window rather than picking a favourite study. It is the range that is defensible no matter which dataset you trust.
A real price today beats a rule of thumb about next month.
Google Flights measured shopping midweek instead of at the weekend as producing fares about 1.9% cheaper. Expedia's 2026 report names Friday as cheapest, roughly 3% below Sunday.
Two sources, two different days, both effects tiny. The myth dates from when airlines filed fare sales by hand early in the week. Continuous automated pricing ended that years ago.
Verdict: a rounding error. Book when you are ready.
For flights, they rise. Cheap fare buckets sell out and are not replenished, so the final three weeks are structurally expensive on most routes.
For hotels, they often do fall — NerdWallet found booking 15 days out beat four months out 66% of the time, by about 13% on average. But that analysis covered 2019 to 2021, and travel patterns have shifted since.
Verdict: true for rooms, false for seats, and a package is mostly governed by the seat.
It is the largest lever on the page. Google data has put Monday to Wednesday departures around 13% below weekend departures, and Tuesday, Wednesday and Saturday recur as the cheapest days to fly across independent analyses.
Thirteen percent against roughly two percent for the shopping day. If you move one variable, move this one.
Verdict: the most underused saving in travel.
Midweek. Google data has shown Monday through Wednesday departures running around 13% cheaper than weekend departures, and Tuesday, Wednesday and Saturday recur across independent analyses as the cheapest days to fly. This effect is roughly six times larger than any difference in the day you buy the ticket, which measures at about 2 to 3%. Shifting your departure day is the single most effective timing decision available to most travellers.
| Lever | Approximate effect | Source |
|---|---|---|
| Travel day: midweek versus weekend | ~13% | Google, 2025 |
| Hotel: 15 days out versus 4 months | ~13% | NerdWallet, 2,500+ rates, 2019–2021 |
| Booking day: Friday versus Sunday | ~3% | Expedia 2026 Air Hacks |
| Booking day: midweek versus weekend | ~1.9% | Google Flights |
| Year-on-year airfare change, 2026 | −3.4% | NerdWallet Travel Price Index, BLS data |
Effects are drawn from separate studies with different methodologies and are not directly comparable. They are listed together to show relative magnitude, not to be added.
The classic rule of thumb — shop in the opposite season — is roughly right for peak travel, and roughly wrong for everything else. Here is the version that accounts for both.
| Travelling | Domestic: start looking | International: start looking | Why |
|---|---|---|---|
| Spring, off-peak | Late January | November | Low demand; the standard window applies |
| Summer, peak | February to March | December to January | Add two months to the standard window |
| Autumn, shoulder | July to August | May | The best value period of the year |
| Thanksgiving | June to August | January to July | Demand concentrated on a few dates |
| Christmas and New Year | July to September | February to August | Cheap inventory disappears and is not replaced |
| Spring break | October to December | September to November | Family travel clusters tightly |
Derived by applying the published booking windows above to each travel period. Holiday guidance of three to five months domestic and four to ten months international comes from Going.
Holiday travel is the one case where the usual warning against booking too far ahead does not apply, and the reason is structural rather than seasonal.
On an ordinary Tuesday in March, demand is spread across many possible dates. If one flight fills, travellers shift to the next day and airlines respond by managing inventory across the whole period.
At Christmas, demand is concentrated on perhaps six specific dates. Nobody can shift. Cheap inventory on those dates sells out and is never replenished, so prices only travel in one direction. The same applies at Thanksgiving and to a lesser extent spring break.
This is also the period where the flight and hotel timing conflict matters least, because holiday hotel rates do not soften near arrival either. Everything is expensive late.
None of this timing work compares to simply travelling when other people are not.
Autumn shoulder season — roughly late September into November in the northern hemisphere — carries most of the weather and a fraction of the price, on both halves of the trip simultaneously. Airfare is lower because demand is lower. Hotel rates are lower for the same reason. Neither effect requires you to guess a booking window correctly.
A perfectly timed booking in July might save you ten percent against a badly timed one. Moving the trip to October can save considerably more than that, and it does so without any monitoring, alerts or cleverness at all.
If you have any flexibility on when to travel, spend it there before you spend a single minute optimising when to buy.
General windows are a starting position. These are the cases where following the general advice will actively cost you money.
The single most concentrated demand event in leisure travel. A narrow set of dates, a finite number of resort rooms, and an audience that largely cannot move its plans.
The general Latin America window of two to four months is wrong here. Treat late December as needing four to eight months, and understand that at the popular all-inclusives you are competing for availability rather than price. If the property matters more than the saving, book when the calendar opens.
A two to three week window, forecast publicly, that draws visitors from every continent at once. Published guidance puts it at four to ten months ahead, against a normal Asia window of four to six.
The complication is that the bloom dates shift year to year and are not confirmed until close to the season, so you are booking a window rather than a date. That is a genuine risk and one worth accepting deliberately rather than by accident.
Room rates here are driven by the convention calendar rather than the season, and they swing more violently than almost anywhere in the United States. The same room can triple across two adjacent weeks with no seasonal logic at all.
That volatility rewards checking the calendar before fixing dates, and it makes fixed booking-window advice close to useless. Shift your dates by three days and you will frequently beat any timing strategy.
Abundant accommodation, competitive short-haul routes and low seasonal concentration mean prices here are less time-sensitive than the three-to-five-month guidance suggests.
These trips also bundle poorly, for the reasons set out on our page on when bundling wins. Booking the two halves separately, later, is usually the better play.
Mountain accommodation behaves unlike urban hotels. Inventory is genuinely limited, the season is short, and the good properties in the good weeks sell out rather than discounting.
Treat a ski package like a holiday booking — three to five months — and be aware that lift tickets, equipment and lessons sit outside the package entirely and want budgeting separately.
Substantially. Regional windows differ — Europe from the US around three to five months, Asia four to six, Latin America two to four, Australia and New Zealand five to nine. But specific events override the regional average entirely. The Caribbean over the winter holidays, Japan during cherry blossom season and ski resorts in peak weeks all need booking far earlier than their regional guidance suggests, because those are availability problems rather than pricing problems.
There is a strategy that works and a habit that looks like the same thing and quietly costs money. Worth separating them.
The strategy: decide your number first. Before you begin searching, work out what you would be content to pay. Not the lowest imaginable price — the price at which you would consider the trip good value and stop thinking about it. Then book when you see it.
This works because it converts an open-ended search with no stopping rule into a decision with one. Without it, every price you see becomes a reason to wait for a better one, and there is no point at which you conclude you have finished.
The habit: waiting for the floor. Fares do not descend tidily toward a knowable minimum. They move in both directions on a schedule set by inventory systems nobody outside the airline can observe. The fare that was $420 this morning and $480 this afternoon has not told you anything about tomorrow.
The most expensive outcome in travel is not paying slightly above the optimum. It is watching a good fare disappear while waiting for a better one, then paying substantially more three weeks later.
Establishing what normal looks like. Two weeks of watching a route tells you whether $520 is high or low for it. That context is worth more than any published booking window, because it is specific to your trip.
Catching genuine outliers. Fare alerts exist because occasional mistakes and sales produce prices far below the normal band. Those are worth being ready for.
Refundable hotel rates. This is the one case where monitoring after booking pays. Book a refundable room early to secure the property, then check again closer to arrival. Given that hotel rates often soften late, rebooking at a lower rate is a real and low-risk option — and one that packages generally remove.
Predicting the direction of a specific fare. Nobody outside the airline can do this, and the tools that claim to are producing confident-sounding output from the same public data you can see.
If a service tells you to wait with high confidence, treat that confidence as a marketing decision rather than a measurement.
Only if you are still outside your booking window. Once you are inside it, waiting is more likely to cost than save, because airfare tends to rise from that point as cheap fare buckets sell out. The practical method is to decide what you are content to pay before you start searching, then book when you see that number rather than waiting for a floor you cannot identify in advance. Prices move in both directions and no public tool can reliably predict which way a specific fare will go.
Shorter booking windows, later discounting, and rates that expire nightly.
Every number on this page comes from a named third party, with the organisation and where possible the date and sample attached. We have not tracked fares, we do not hold a price dataset, and we are not going to imply otherwise by writing in a way that suggests the findings are ours.
What we have done is read the major published sources, notice that they disagree, work out why, and publish the overlap. That synthesis is genuinely ours. The underlying measurements are not.
The two-line chart near the top of this page illustrates directions the sources support — airfare falling then spiking, hotel rates continuing to fall. The exact curvature is drawn rather than plotted. We label it as illustrative on the page itself rather than in a footnote nobody reads.
The NerdWallet hotel analysis covers 2019 to 2021. It is the most rigorous public comparison of advance versus last-minute hotel rates we could find, and it predates substantial changes in travel patterns. We cite it because it is the best available, and we flag the date because you should weigh it accordingly.
Expedia's data describes Expedia's customers. We cite it and we also point out that this population skews toward leisure and package travel, which plausibly explains its shorter recommended window. We are an Expedia affiliate; that is disclosed on every page, and it does not change what the caveat should say.
Averages hide enormous route-level variation. A window that is right on average can be wrong on your specific route in your specific week. Treat every figure here as a starting position rather than an instruction.
None of it beats flexibility. Moving your dates outperforms timing your purchase, and we would rather say so than sell you a system.
Eighteen questions with sources attached. Where the honest answer is that nobody knows, that is what we say.
See also: whether to bundle at all and the bundle versus separate calculator.
Four to eight weeks before departure for domestic trips, three to five months for international. Those windows are where the published guidance from Google Flights, Expedia and Going overlaps, which makes them the most defensible ranges available given that the sources disagree with one another. Peak season and holiday travel need the longer end of each range, and in some cases more.
No, and this is the single most important thing to understand about package timing. Airfare generally falls until roughly five to six weeks before departure and then rises steeply as cheap fare buckets sell out. Hotel rates often behave the opposite way: Hotels.com's 2026 Hotel Price Index identifies 8 to 14 days before travel as the sweet spot, and NerdWallet's analysis of over 2,500 room rates found booking 15 days out was cheaper than four months out 66% of the time. The two components move in opposite directions.
Time the flight. Airfare rises more steeply and less predictably close to departure than hotel rates fall, and a sold-out fare bucket cannot be recovered while an unsold room usually can. Booking in the flight window means you accept a hotel price slightly above its theoretical floor, which is the smaller of the two losses. If your trip is genuinely flexible and the hotel is the dominant cost, booking separately and timing each half independently beats any package.
Three to five months for most destinations, with regional variation: Europe from the US around three to five months, Asia four to six, Latin America two to four, and Australia or New Zealand five to nine. Add roughly two months for peak-season travel such as summer in Europe or the Caribbean over the winter holidays. Going's guidance runs wider at two to eight months, which is compatible with these ranges rather than contradicting them.
No. Google Flights measured shopping on Tuesday, Wednesday or Thursday instead of a weekend as producing prices only about 1.9% cheaper on average. Expedia's 2026 Air Hacks Report actually names Friday as the cheapest booking day, roughly 3% below Sunday, attributing the shift to reduced end-of-week business demand. The myth dates from an era when airlines filed fare sales manually early in the week; continuous automated pricing erased it years ago.
Midweek, and this matters far more than the day you book. Google data has put Monday to Wednesday departures around 13% cheaper than weekend departures, and Tuesday, Wednesday and Saturday recur across independent analyses as the cheapest days to fly. A 13% effect on the travel day dwarfs a 2% effect on the shopping day, so if you can only move one variable, move when you fly.
For the hotel half, sometimes. For the flight half, almost never. Airfares generally rise in the final three weeks as cheap fare buckets sell out, while hotel rates frequently soften close to arrival because an unsold room earns nothing. In a package these two effects work against each other, and the flight effect is usually larger, so last-minute packages are on balance a poor bet. The exception is genuine distressed inventory at a resort with rooms to fill.
Most major US airlines open bookings around 330 days, or roughly eleven months, before departure. Being able to book that early does not mean you should. Prices at the opening of the window are set before competition has established a market rate, and every major dataset agrees that booking eleven months out is not the cheapest option. The competitive pricing that produces low fares needs time to develop.
Because they measure different populations. Google Flights sees search queries across every carrier and route. Expedia sees its own bookings, which skew toward leisure and package travellers. Going watches a curated set of routes hunting for outlier fares rather than averages. None is wrong; they are describing different slices of the same market. A site that presents one as the definitive answer is either unaware of the others or hoping you are.
Shorter than most people expect, and shortening. SiteMinder reported the global average hotel booking window at 32.15 days in 2025, with Ireland the longest at 46 days. Hotels.com's 2026 Hotel Price Index puts the value sweet spot at 8 to 14 days before travel. Hotels manage inventory that expires nightly, which pushes discounting late; airlines manage fare buckets that sell out, which pushes prices up late.
It is a reasonable strategy for hotels and a poor one for flights. A refundable room booked early costs a little more but lets you capture a drop if rates soften near arrival, which they often do. Airfare rarely falls in the final weeks, so waiting on a flight usually costs money. Within a package neither option is usually available, because the booking is a single contract governed by the stricter of the two components.
Three to five months ahead for domestic travel and four to ten months for international. Holiday travel is the case where booking early genuinely pays, because demand is concentrated on a handful of specific dates and cheap inventory disappears rather than being replenished. The common advice to book in the opposite season is a reasonable rule of thumb: shop for Christmas travel around July, and for summer travel around Christmas.
There is no reliable calendar date at which packages get cheaper. Package pricing follows the underlying inventory of both components, so it moves with the airline's fare buckets and the hotel's unsold rooms rather than to a schedule. What does exist is a window, roughly four to eight weeks out domestically, where both components are usually reasonably priced at once. That is a window, not a moment.
Lower. Airfares in 2026 have been running around 3.4% below 2025 levels according to the NerdWallet Travel Price Index, which uses Bureau of Labor Statistics data. Analysts have also projected softer total travel demand, which generally means more unsold seats and more opportunities for fares to fall. That is a mild tailwind rather than a transformation.
Frequently, and this surprises people. Airlines set opening prices before competitive pressure has established the market rate for a route, so fares eleven months out are often above what the same seat costs three months out. Hotels behave similarly, publishing rates far ahead that they later discount to fill rooms. Both the earliest and the latest ends of the booking timeline tend to be expensive; the value sits in the middle.
You usually cannot know, and chasing certainty is how people miss good prices entirely. Fares do not descend tidily toward a knowable floor; they move in both directions on a schedule nobody outside the airline can see. The practical approach is to decide the price you would be content to pay before you start looking, and pay it when you see it. The most expensive habit in travel is refusing to book while waiting for slightly better.
Somewhat. All-inclusive resorts sell a larger share of their inventory through packages and further ahead than city hotels, and popular properties during peak weeks genuinely do sell out. That pushes the sensible booking window earlier than the general hotel guidance would suggest. Treat all-inclusive peak season more like holiday flight timing: three to five months rather than the eight to fourteen days that suits an urban hotel.
For hotels, yes, if you booked a refundable rate. For flights and packages, usually not, because the terms rarely let you act on what you find and the exercise mostly produces regret. If you do check and find a meaningful drop on a refundable hotel booking, rebooking is straightforward. Otherwise, the sensible move after booking is to stop looking.
Inside your window?
Open Expedia ↗Affiliate link. We may earn a commission on qualifying bookings, at no extra cost to you.