Short answer: an airline does not set one price for a flight. It loads a ladder of fares and controls how many seats it will sell in each rung. The number you see is simply the lowest rung still open when you search — and rungs open and close as the flight fills, as competitors move, and as the departure date approaches.
A flight is not one product
Think of a 180-seat aircraft as 180 opportunities to sell to very different customers. The family booking a holiday four months out is price-sensitive and flexible on dates. The consultant booking on Tuesday for a Wednesday meeting is neither. A single price would either leave money on the table with the consultant or price the family out entirely.
Revenue management resolves this by segmenting the same physical seat into fare classes — the letters you occasionally see on your booking (Y, B, M, H, Q, and so on). Each carries a price and a set of conditions: advance purchase requirements, change fees, refundability, minimum stay, baggage allowance.
Buckets, not a dial
Here is the part most people get wrong. Airlines rarely change the fare amounts themselves day to day. What they change is availability — how many seats are authorised for sale in each bucket.
| What you experience | What actually happened |
|---|---|
| The price jumped overnight | The cheapest open bucket sold out; the next rung up is now the lowest available |
| The price dropped a week before departure | The flight is booking behind forecast, so cheaper buckets were reopened |
| Two people see different prices | They searched at different moments, or for different itineraries, cabins or points of sale |
The forecasting system continuously compares actual bookings against the expected booking curve for that flight. Ahead of curve, close cheap buckets and protect seats for late, high-value demand. Behind curve, release inventory to fill the aircraft.
What genuinely moves the number
- Days to departure. Late bookings skew business and inelastic, so cheap inventory is deliberately withheld from them.
- How full the flight already is. Load factor against forecast is the single biggest input.
- Competition on the route. A rival dropping a fare on the same city pair can be matched within hours by automated fare filing.
- Day of week and season. A Friday evening departure and a Tuesday lunchtime departure on the same route are separate markets.
- Capacity. An extra daily frequency, or a competitor exiting, resets the whole pricing environment.
- Currency and point of sale. The same seat sold in different markets can sit in a different fare structure.
What does not move it
Two myths deserve a quiet burial.
Cookies and "they saw you searching". Fares are filed and distributed through global systems, and availability responds to bookings, not to your browsing history. Repeated searches do not train the system to punish you. If you want to test it, search in a private window and compare.
A single golden booking window. The "book exactly 54 days out" advice is an average of averages across millions of itineraries. It tells you almost nothing about your specific route on your specific date. A peak-season leisure route and a thin business route have opposite curves.
What this means for booking
- For peak leisure dates — school holidays, festivals, long weekends — the cheap buckets go early and rarely come back. Book early.
- For thin off-peak routes, airlines often reopen cheap inventory close to departure to fill the aircraft. Waiting can work, but it is a genuine gamble.
- Be flexible on the day, not the route. Shifting a departure by 24 hours usually moves you into a completely different demand pattern, which does far more than any browser trick.
- Set a price alert and understand what it tells you. A falling fare means a bucket reopened; that inventory can close again just as quickly.
The takeaway
Airline pricing feels arbitrary because you are seeing the output of a system that is optimising the revenue of the whole aircraft, not quoting you the value of one seat. The price is a snapshot of remaining inventory in a market that is moving continuously. Once you see it as buckets opening and closing rather than a price going up and down, the behaviour stops looking random.