Travel Craft
Fares move because seats are sold in batches at different prices
Airline and rail pricing is a system for rationing a fixed number of seats over time, and understanding the batches explains most of the behaviour that looks random.
By Pranav Kulkarni3 min read

A single flight is sold at many prices at once
A seat is a perishable product: once the aircraft or train departs, an unsold seat is worth nothing for ever. Operators therefore try to sell each seat for the most that someone will pay for it, and the tool for doing so is to divide the cabin into price bands sold in sequence.
These bands are inventory rather than product. The seat itself is identical; what differs is the price attached to a particular allocation, and when one allocation is exhausted the system offers the next one up. That is the whole of why the price you saw last week is gone.
It also explains why fares appear to jump rather than to climb. Movement between bands is a step change, and the size of the step depends on how the bands were set, not on how many days remain.
Time is a proxy, not the mechanism
The familiar advice that prices rise closer to departure is a description of a correlation rather than a rule. Prices rise because seats sell, and seats usually sell as departure approaches, so the two tend to move together without one causing the other.
When they come apart the effect is visible. A departure that is selling poorly can have cheap inventory released into it late, and travellers who watch a specific route for long enough will see fares fall in the final weeks often enough to know the rule is not absolute.
The opposite also happens: a route with strong demand can exhaust its cheap allocations months out, so booking early on a popular seasonal route is not caution but the only way to access the low bands at all.
Business and leisure demand are separated deliberately
Much of the structure exists to distinguish between someone who must travel on a particular day and someone who is choosing to. That is why the cheapest fares carry conditions that a business traveller cannot accept: no changes, a fixed departure, a Saturday in the trip, a long connection.
Those conditions are not arbitrary inconveniences. They are filters, designed so that the passenger who would have paid more does not qualify for the low fare, and each one exists because it correlates with a type of demand.
Understanding that changes how you read a fare table. The expensive ticket is not better transport; it is the same seat with the filters removed, and whether that is worth the difference depends entirely on how likely your plans are to change.
The folklore is mostly unproven
A large body of advice circulates about clearing cookies, browsing privately, booking on a particular day of the week or at a particular hour. Operators generally deny that individual browsing history raises the price shown, and independent attempts to demonstrate the effect have not produced a consistent result.
The honest position is that this is unresolved rather than settled in either direction. Prices genuinely do change frequently, which makes it easy to see a pattern in what may be ordinary inventory movement, and a single anecdote cannot distinguish between the two.
What is not in dispute is that the same journey is often priced differently on different sales channels, and that some markets are priced separately from others. That is a real effect with a commercial explanation, and it is worth checking more than one channel for that reason rather than for the folklore one.
What actually gives you leverage
Flexibility is the only lever that reliably works, and it works because it lets you shop across many separate inventories. Being able to move a day either side, to use a second airport or to substitute a train for a flight multiplies the number of price bands you can reach.
Watching a specific route over several weeks before you need it is the other useful habit. It gives you a sense of the normal range for that journey, which is the only way to know whether a price in front of you is good, since a figure with nothing to compare it against tells you nothing.
Beyond that, a certain amount of acceptance is warranted. The system is designed by people with far better information than any traveller has, and the aim is to pay a reasonable price without spending a week on it rather than to beat a pricing engine at its own game.
Common questions
Is there a best day to book?
Claims of a single best day have been made repeatedly and do not hold up consistently across routes, markets and years. Any effect that did exist would be arbitraged away quickly once widely known. Route and season matter far more than the day of the week you happen to buy.
Why did the price go up while I was checking out?
Usually because the last seat in that band was sold, either to someone else or by being held in another session, and the system moved to the next allocation. It is also possible for a displayed fare to be stale by the time you commit, since inventory changes continuously.
Do rail operators price the same way?
Those running yield-managed fares do, and the same band logic applies. Networks with regulated or distance-based fares do not, and on those the price is stable regardless of when you buy, which is why booking behaviour that makes sense in one country can be pointless in another.
Consumer editor, The Next Postcard
Pranav joined to cover cities, slow routes, rail & road and stayed for the awkward questions and is happiest when a piece answers the question completely.





