How far in advance should I book a domestic flight?

Roughly three weeks to two months out covers most of the low range. Google Flights data reported for US domestic travel puts the typical low point around 38 days before departure, with a broad low band from about 21 to 52 days.

How far in advance should I book an international flight?

Earlier and wider than domestic: commonly two to eight months out, and toward the longer end for long-haul, peak season, or routes with limited seats in the cheaper fare classes.

Is it ever cheaper to book last minute?

Occasionally, when a specific flight is selling badly. It is not a strategy, because airlines fill their cheapest fare buckets first and what remains close to departure is disproportionately expensive.

When should I book for Christmas or Thanksgiving?

Earlier than an ordinary week. Google's data puts the domestic low point around 45 days before Thanksgiving and around 58 days before Christmas, both with wide ranges around them.

Does booking very early get the best price?

Usually not. Fares released eleven months out are often set conservatively high, and the cheaper buckets tend to open later. Booking extremely early buys certainty and seat choice rather than the lowest fare.

How far in advance should you book a flight?

There is a real booking window. It is wider and blurrier than most headlines admit, and the useful version of the advice comes with a caveat that the headlines leave out.

The domestic window

For US domestic trips, most published analysis converges on somewhere between three weeks and two months before departure.

Google Flights, working from several years of its own data, has put the typical low point around 38 days before departure, with a broad low range running from roughly 21 to 52 days out. That is the shape worth remembering: not a day, a band about a month wide, with a gentle low point in the middle rather than a cliff on either side.

Holiday travel shifts earlier. The same analysis puts the domestic low point around 45 days ahead of Thanksgiving and around 58 days ahead of Christmas, again with wide ranges around each.

The international window

International travel moves earlier and spreads wider: commonly two to eight months before departure, and toward the longer end for long-haul routes, peak season, or routes where only a few carriers operate and the cheap fare classes are thin.

The reason is inventory rather than anything mysterious. A route with one daily flight has far fewer discounted seats to distribute than a route with six, so those seats are gone earlier, and the window for catching them opens and closes sooner.

Both ends of the range are bad

Fares generally get worse in the final one to three weeks. Airlines sell their cheapest fare buckets first, so what remains close to departure is disproportionately expensive, aimed at travellers whose dates cannot move. Last-minute drops happen when a specific flight is selling badly, but betting on that means betting a particular aircraft will underperform, which is not a plan.

Booking extremely early is a subtler trap. Fares typically load about eleven months ahead and are often set conservatively high, with the cheaper buckets opening later. What very early booking buys you is certainty and seat choice. It does not usually buy the lowest price.

Where the rule stops working

Here is the caveat, and it is larger than the rule.

Every number above is an average over a vast number of routes with completely different competitive structures, seasonal shapes and demand forecasts. The average is real. It is also close to useless for your specific flight, because your route may behave nothing like the centre of that distribution. A monopoly route with one carrier and a contested route with four do not share a booking curve, and no general window describes both.

The deeper problem is that a calendar rule answers the wrong question. “Book six weeks out” cannot tell you whether $450 is a good price. On a route that normally sells for $300, $450 is poor no matter how well you timed it. On a route that normally sells for $900, $450 six weeks out is excellent, and so is $450 five months out, and so is $450 nine days out.

Timing is a weak proxy for the thing you actually want to know, which is where this fare sits relative to what this route normally costs. The proxy is worth using when you have nothing better. It is not worth using instead of the real thing, and the real thing is not hard: there is a ten-minute method in how to tell if a flight price is good, and the mechanics behind the curve are in why flight prices change.

How to use the window properly

Treat it as a default for when you know nothing else. If you have no information about a route, starting to look about five weeks out domestically, or three to five months out internationally, puts you in roughly the right neighbourhood.

Then replace it with something better as soon as you can. Spend ten minutes establishing what the route normally costs, and the window becomes what it should have been all along: a rough guide to when to start paying attention, not a rule about when to buy.

So the honest answer to “when should I book” is another question. Do you know what your route normally costs? Until you do, any date on the calendar is a guess with good manners.