Why flight prices change so much

Short version, for anyone who wants it and nothing else: an airline does not set one price per flight. It divides each cabin into a set of fare buckets, prices them differently, and uses software to continuously adjust how many seats sit in each bucket as departure approaches. The price you see is whichever bucket is currently open to you. That is why the number moves, and why it moves without any connection to who you are or what you searched yesterday.

The rest of this explains how that works, and disposes of a few myths that survive mainly because the truth is less satisfying.

Fare buckets

Every seat in an economy cabin is physically the same. Commercially, they are not.

Airlines file fares in classes, identified by single letters, and each class carries its own price and its own conditions on changes, refunds, baggage, and mileage accrual. A single economy cabin might contain ten or more of these. The cheapest are heavily restricted; the most expensive are flexible and refundable. Two passengers in adjacent seats can easily have paid a factor of four apart from each other, and both paid the correct price for the class they bought.

What the airline manages is not the price of the flight but how many seats are allotted to each class at any given moment. When the cheapest class sells out or gets closed off, the displayed price jumps to the next one up. That is the mechanism behind almost every sudden increase people notice, and it is a step change rather than a smooth drift, which is why fares often sit still for days and then move sharply.

Revenue management

The system deciding how many seats sit in each bucket is called revenue management, and its objective is not to charge as much as possible. It is to fill the aircraft while capturing as much value as the demand curve allows.

It works from a forecast. For a given flight on a given date, the software has an expectation of how bookings should accumulate over the months before departure, built from historical data on that route, that season and that day of week. Then it compares reality to the forecast and reacts.

Bookings running ahead of forecast means demand is stronger than expected, so cheap buckets close early and the price rises. Bookings running behind means the opposite: cheap inventory reopens, and sometimes a fare sale appears. This is the entire explanation for the seemingly irrational behaviour of a fare dropping six weeks out on one route while rising on another. The two flights are tracking differently against their own forecasts.

Two consequences worth internalising.

Prices usually rise close to departure, because the cheap buckets sold first. Google Flights data reported for domestic US travel puts the low point around 38 days before departure, with a broad low range roughly 21 to 52 days out. What remains inside the final two or three weeks is disproportionately the expensive classes, aimed at business travellers who cannot move their dates.

Late drops are a signal of weak demand, not generosity. When a fare falls close to departure, it means that flight is not selling. That happens, but planning around it means betting that a specific flight will underperform, which is not a bet with good odds.

What makes one route expensive and another cheap

Revenue management sets the movement. Route economics set the level the movement happens around.

Competition. The single largest factor. A route with three carriers and a low-cost operator behaves completely differently from a route one airline flies once a day. Monopoly routes are expensive and stable; contested routes are cheaper and volatile.

Capacity. Seats per week, which is a function of aircraft size and frequency. A route served by a widebody twice daily has far more cheap inventory to distribute than one served by a regional jet.

Seasonality. Demand on leisure routes swings enormously by month, and the fare band swings with it. A summer Mediterranean route and the same route in February are, for pricing purposes, unrelated.

Directionality. The same city pair can be priced differently depending on which end you start from, because the point of sale market matters. Origin markets with weaker purchasing power or stronger competition are often cheaper, which is why a round trip starting in one city can undercut the identical itinerary starting in the other.

Connections. A connecting itinerary competes against every other routing between the same two cities, so airlines discount connections to win traffic they would otherwise lose. That is why a one-stop can undercut a nonstop on the same airline.

Things that do not affect your price

Three pieces of folklore that outlive every attempt to kill them.

Incognito mode and cleared cookies. They do nothing. Sophia Lin, a director of product management for travel at Google, told Travel + Leisure in June 2025 that incognito mode, browsing history, search history and switching devices do not affect the prices shown in Google Flights. Fares on busy routes change often enough on their own that anyone who searches twice and sees two numbers can construct a story about surveillance. The story is wrong, and clearing your browser before searching is wasted effort.

Booking on a Tuesday. This came from an era when airlines filed fare changes on a weekly cycle and discounts often landed early in the week. Fare filing does not work that way now. Day of purchase is close to irrelevant; day of departure is not, and midweek departures do tend to be cheaper than Friday and Sunday ones.

Searching too many times. No major airline or metasearch engine has been shown to raise a fare because an individual looked at it repeatedly. The prices moved because prices move.

Why nobody can tell you the best day to book

Every general booking rule is an average over an enormous number of routes with wildly different competitive structures, seasonal shapes and demand forecasts. The average is real. It is also close to useless for the specific flight in front of you, because your route’s behaviour may sit nowhere near the centre of that distribution.

A rule can tell you that domestic fares tend to bottom out around five weeks before departure. It cannot tell you whether $612 is a good price for your route, because it does not know your route. Those are different questions, and only the second one determines what you pay. There is a method for answering it in how to tell if a flight price is good, and a fuller treatment of the booking window in best time to book a flight.

Glossary

Fare class. A single-letter code identifying a fare’s price and conditions within a cabin. Economy typically contains many of them.

Fare bucket. The pool of seats currently available at a given fare class. Airlines adjust bucket sizes continuously; this is what makes displayed prices move.

Revenue management. The airline discipline, largely automated, of deciding how many seats to sell at each price point in order to maximise revenue for a departure.

Yield. Revenue per passenger per mile. The metric revenue management is optimising against.

Load factor. The percentage of available seats sold on a flight. High load factor with low yield can be worse for an airline than the reverse.

Point of sale. The market a ticket is sold in. Fares for the same itinerary can differ by point of sale, which is why a route can be cheaper starting from one end than the other.

Fuel surcharge. A carrier-imposed fee, separate from the base fare and from government taxes. On long-haul tickets it can exceed the fare itself, which is why a fare that loses its surcharge becomes a mistake fare.

Knowing the vocabulary will not get you a cheaper ticket. It will stop you looking for explanations in the wrong place, which is most of what separates people who consistently pay less from people who clear their cookies.