How far ahead to book, according to the data

Airfare advice circulates widely and most of it is either outdated or was never true. The pricing systems behind it are more explicable than they appear.

How airline pricing works

Seats on a flight are divided into fare classes, each with a limited number of seats and a different price.

Revenue management systems adjust how many seats are available in each class continuously, based on how bookings compare to the forecast for that flight.

Which means the price you see is a function of how full the flight is relative to expectation at that point before departure, rather than of a fixed schedule.

The practical consequence: prices do not decline steadily as departure approaches and do not rise steadily either. They move in steps as fare classes sell out or are reopened.

What the analyses show

Studies of large fare datasets by booking platforms and by researchers converge on a few findings.

There is generally a window during which average fares are lowest, and it is wide rather than a specific day.

For domestic travel in most markets, that window is commonly identified as somewhere between one and four months before departure.

For international travel, longer — commonly two to eight months depending on the route and season.

Booking very far ahead is generally not cheapest, because airlines release inventory at higher fare classes before adjusting.

Booking within two weeks of departure is generally most expensive, particularly on routes with substantial business demand.

The myths

That there is a best day of the week to book. This was based on a period when airlines filed fares on a weekly cycle. Pricing is now continuous, and analyses find negligible differences by booking day.

That clearing cookies or browsing privately produces lower fares. Repeatedly tested and not supported.

Fares change frequently for reasons unrelated to the individual viewer, which is why people believe they have observed the effect.

That the day of travel matters more than anything. It does matter — midweek departures are generally cheaper than Friday and Sunday — and it is one factor among several.

That a fare will drop if you wait. Sometimes, and studies of fare movements generally find that waiting past the optimal window costs more often than it saves.

What actually reduces the fare

Flexibility on dates. The largest single lever.

Fare calendars showing a month at a time make the variation visible, and shifting by a day or two frequently changes the price substantially.

Flexibility on airports. Secondary airports in a metropolitan area frequently price differently, though ground transport costs and time should be counted.

Avoiding peak periods, which is obvious and worth quantifying: school holidays, major festivals and public holidays produce predictable price increases in each market.

Considering one-way tickets on two carriers, which is frequently cheaper than a return on one, particularly on short-haul routes with low-cost carriers.

Checking nearby dates around a fixed event, since arriving a day earlier or leaving a day later is often substantially cheaper.

Price tracking

Several services monitor fares on specified routes and alert on changes.

These are genuinely useful for a flexible traveller and less so for someone with fixed dates, since the alert may be for dates that do not work.

Some services publish predictions about whether a fare will rise or fall. Their accuracy is moderate and they are more useful as a prompt to decide than as a forecast.

The cancellation rules

Worth knowing because they change the calculation.

In the United States, regulations require that a booking made at least a week before departure can be cancelled without penalty within twenty-four hours of purchase.

Which means a fare can be secured while checking whether a better option exists.

Rules elsewhere vary, and the fare conditions rather than the airline's general policy determine what applies.

The practical approach

Set alerts as soon as dates are known.

Check the fare calendar for the month rather than a specific day.

Book when a fare appears that is reasonable relative to what you have seen, within the window described above, rather than attempting to identify the bottom.

And accept that some fares will fall afterwards, which is the cost of not spending the intervening weeks monitoring.