Cheap flights: how the price you see is actually set

Fare buckets: why one flight has twenty prices

Every seat on a commercial flight is sold out of an inventory bucket identified by a single letter — the booking class. Y is generally full-fare economy; the discounted classes below it run through letters like B, M, H, Q, V, S, N, O and so on down to the deepest promotional class the airline has chosen to open. Each class holds a limited number of seats and each is tied to a fare basis code carrying its own rules on advance purchase, minimum stay, changeability, refundability, mileage accrual and upgrade eligibility.

The airline's revenue management system decides, continuously, how many seats to leave open in each class. When the cheapest class sells out, the next fare up becomes the lowest available. Nothing was "raised" — a bucket simply emptied. This is why a fare you looked at yesterday is gone today, and why refreshing does not bring it back.

It also explains something travellers find counter-intuitive: a flight can get cheaper close to departure. If the revenue system concludes a departure is going to fly with empty seats, it can reopen a discounted bucket. That is a gamble, not a strategy — the far more common outcome inside two weeks is that only the expensive classes remain. We would not advise anyone to wait on it for travel that matters.

Fare buckets also drive things that have nothing to do with price. Upgrade eligibility is set by booking class, not by what you paid: two passengers who paid the same amount can have different upgrade rights because they were ticketed in different classes. Mileage accrual works the same way. If either matters to you, say so at the point of quoting — see flight upgrades.

Advance-purchase windows and when they bite

Discounted fares carry an advance-purchase condition: the ticket must be issued a stated number of days before departure. The common thresholds are 21, 14, 7 and 3 days, and they operate as cliffs rather than slopes. At 22 days out the fare exists; at 20 days out the same fare basis is simply not available for ticketing and the system prices you into the next class up.

Two practical consequences follow. First, if you are close to a threshold, the deadline is on ticketing, not on holding a reservation — an unticketed booking does not preserve the fare. Second, the "book early" advice is only half right. Airlines release cheap inventory progressively, so buying eleven months out often means buying before the discounted classes have opened. Squeeze those two mechanics together — advance-purchase thresholds pushing you earlier, progressive inventory release pushing you later — and the practical range is one to three months ahead on US domestic and two to six months on long-haul, stretched further for peak holiday travel.

Myths we get asked about, answered plainly

Tuesday at 1am. This came from an era when fare filings were loaded in batches, so a wave of new sale fares would appear at predictable times. Modern revenue systems reprice continuously. There is no purchase hour that reliably beats another.

Incognito mode and clearing cookies. Fares come from airline inventory systems and are returned the same regardless of your browsing history. If a price moves between two searches, a booking class changed. A browser cookie is not an input to fare construction: the number you are shown is the filed fare for whichever booking class is open at that instant, and the reservation system that returns it never sees your browsing history at all.

Searching from another country's site. Point of sale genuinely can matter, because fares are filed by market and currency. But acting on it means buying in a foreign currency, absorbing a card foreign-transaction fee, and holding a ticket sold under another country's consumer rules — which is exactly the wrong trade when something goes wrong. We do not recommend it for US-originating travel.

Booking two one-ways is always cheaper. Sometimes on low-cost carriers, rarely on legacy long-haul, where round-trip fare construction and minimum-stay rules often price below two singles. It is worth testing, not assuming, and an advisor can test both in one pass.

Hidden-city and throwaway ticketing: why we decline

Hidden-city ticketing exploits the fact that a fare to a further point sometimes prices below a fare to the connecting point. Throwaway ticketing exploits round-trip pricing by discarding the return. Both work arithmetically. Both breach the contract of carriage you accept when you buy the ticket.

The failure modes are real and they compound. You cannot check a bag, because it will be tagged to the ticketed final destination. Missing a segment normally cancels every remaining segment on the ticket, including your return. Loyalty accounts can be debited or closed. And because the ticket is sold by an agency, the airline can raise a debit memo against that agency for the fare difference — which is why no reputable agency will knowingly construct one.

There is a legitimate version of the same idea, and it is simply asking for it: open-jaw itineraries, where you fly into one city and out of another, and stopovers, where you deliberately break the journey at a connecting point for a day or a week. Both are priced, ticketed and fully protected. Many long-haul fares permit a stopover at the carrier's hub for a modest fare uplift or none at all. Ask the advisor to price the open-jaw before you consider anything clever.

When a consolidator fare genuinely beats the published price — and when it does not

A consolidator fare is a net fare an airline sells to an agency under contract, usually to move volume on a route where it is under-selling. The discount is real, but it is concentrated exactly where that logic applies: long-haul international economy and premium cabins, where the seat is expensive enough that an unsold one hurts; itineraries involving two or more carriers, where no single airline's own website can build the fare; and dates outside the last three weeks before departure, before the carrier's own revenue system starts taking whatever it can get.

It shows up least on US domestic short-haul, on low-cost carriers that do not discount through agency channels, and on routes where the published market is already fiercely competitive. It also comes with tighter rules than the published fare: often no voluntary changes, refunds limited to recoverable taxes, no upgrade eligibility, sometimes reduced mileage accrual, and a requirement that the issuing agency handle any reissue.

The correct comparison is therefore never fare against fare. It is total cost of the trip against total cost of the trip, including a checked bag, a seat assignment, and a realistic probability that you will need to change something. We do that arithmetic out loud. If the published fare wins, we tell you and point you at it — see how our booking process works for what happens next.

What to actually do

Decide first which of your constraints are real. Fixed date and fixed airport is a different search from flexible date and any airport within two hours' drive, and the price gap between those two searches is usually larger than any other saving available to you. Then decide how much flexibility you need to buy: read the change and cancellation rules of the fare, not the marketing name of the fare family — what each fare type actually returns to you is set out in our refund and cancellation policy and on the cancellations page.

Then price it properly. Add the checked bag if you will check one — our baggage reference sets out the typical bands. Add the seat if sitting together matters, which is covered in prebook seats. Add any service fee, published in our fee schedule. Then compare that number, not the headline, because a fare that looks $40 cheaper and carries a $70 bag is not cheaper. And if you are landing somewhere that needs a vehicle, put the car rental on the same file, so a retimed arrival is checked against the counter's opening hours while the ticket is still being worked on rather than after you land.