How airlines price flights using revenue management systems
Airlines adjust ticket prices thousands of times daily using algorithms that analyze demand, capacity and competitor fares. Here's how the systems that create price variation actually work.
A flight from LaGuardia to Los Angeles showing $420 at 6 a.m. and $680 at noon on the same day to the same passenger is not a glitch or an airline trying to trick you. It is the output of automated systems that airlines call revenue management, which makes pricing decisions thousands of times per day by analyzing current demand, seat availability, and what competitors are charging. The system does not set prices randomly. It follows mathematical logic aimed at filling every seat while charging passengers as much as they are willing to pay.
Revenue management has been core to airline economics since the 1970s, when deregulation allowed airlines to compete on price rather than government fares. Understanding how it works explains why identical seats on identical flights cost different amounts, and why the cheapest fares often appear weeks in advance or in the last few days before departure.
How the system sees inventory and demand
An aircraft flying from New York to Los Angeles has finite inventory: perhaps 180 seats. Once that flight departs, any empty seat is lost revenue. A revenue management system solves this by predicting how many passengers will book over the next weeks or days, then rationing cheap seats to earlier bookers and reserving expensive seats for last-minute buyers who have less flexibility.
The system starts with historical data: on this route, on this day of the week, in this season, how many people booked how far in advance, at what price? It layers in current signals: how many passengers have booked so far, how does that pace compare to the forecast, what are competitors charging? Systems compare actual booking pace to the forecast in real-time and adjust if bookings are running ahead or behind.
When bookings are ahead of forecast, cheap fares disappear quickly and expensive fares remain available. When behind forecast, the airline opens cheaper fares to stimulate demand. The passenger service system that processes reservations receives instructions on which fares to display and which to hold back, based on these calculations.
Segmenting passengers by booking behavior
Airlines do not charge everyone the same because passengers have different needs and time horizons. A business traveler booking on Wednesday for Thursday travel needs to leave that day and will pay a premium. A family booking in September for December travel did so months ago and paid much less. Revenue management systems segment passengers by when they book, what they do, and their loyalty status.
The system assigns different fare classes within the same cabin and controls how many seats each class can sell. Economy might have five or six fare classes: basic fares for early bookers, standard fares for mid-range, and premium economy seats that include checked bags or seat selection. When the airline has sold enough basic fares, the reservation system stops showing them and opens only standard fares or premium fares instead. No inventory is reserved on an empty seat; it is a matter of how many seats can be sold at each price tier.
This is called yield management: the airline aims to extract maximum revenue from each departure by ensuring that high-paying passengers can buy seats while not leaving seats empty if cheaper fares could sell them. The balance between these two goals is what creates price variation that passengers see.
The algorithms that set the prices
The core algorithm predicts future demand using historical patterns, current booking data, seasonality, and external factors. Airlines use statistical models to forecast how many people will buy in the next few days at each price point. Modern systems incorporate machine learning to improve these forecasts by detecting patterns in weather impacts, events, competitor actions, and market conditions that humans might miss.
The algorithm then solves an optimization problem: given predicted demand, remaining capacity, and time until departure, what price should the airline set to maximize total revenue? If the plane is half full and departure is three days away, the revenue management system will likely open discounted fares. If the plane is 85 percent full with 24 hours to go, it will close discounted fares and raise prices for remaining inventory.
These decisions flow downstream to distribution channels. The airline's website receives instructions on which fares to display. Global distribution systems used by travel agents receive the same instructions. Third-party booking sites see the fares the airline has published. From the passenger's perspective, availability seems random, but it is driven by this algorithmic allocation of inventory across fare classes and price points.
“When bookings are ahead of forecast, cheap fares disappear quickly and expensive fares remain available; when behind forecast, the airline opens cheaper fares to stimulate demand.”
From fixed fares to continuous pricing
Traditionally, airlines sold tickets at discrete price points: $100, $150, $200. If a passenger would pay $125 but those exact fares were sold out, the airline had to choose: show them the $100 fare and lose $25 in potential revenue, or show them $150 and they leave without buying. Modern revenue management systems are moving to continuous pricing, where airlines can offer any price on a curve rather than restricting themselves to preset fares.
Lufthansa began testing continuous pricing in 2020, removing reliance on fixed price points and instead calculating the optimal price for each specific booking request based on demand models, competitive fares, loyalty tier, and itinerary details. This approach aims to capture the revenue that fixed fare structures leave on the table.
Continuous pricing does the same work as traditional revenue management but with more granularity. Instead of allocating seats across five fare classes, it uses real-time calculation to set a unique or near-unique price for each booking. The result is more price variation than passengers are accustomed to seeing, because the system can optimize for smaller slices of passenger behavior and willingness to pay.
How the system began
Before 1978, U.S. airline fares were set by the Civil Aeronautics Board, the government regulator. Airlines could not compete on price; they competed on service and flight frequency. Deregulation in 1978 changed this completely. By 1983, airlines were increasingly free to set their own prices and could no longer rely on government-set fares to structure revenue.
Robert Crandall, president of American Airlines, pioneered yield management in response. His team built early computer systems to optimize prices by passenger segment and booking time, introduced the frequent flyer program to encourage repeat business, and implemented practices like overbooking and yield management that are still used today. What began as a competitive advantage for American Airlines became industry standard.
What began as a competitive advantage for American Airlines became industry standard. Today, every major airline operates revenue management systems as core infrastructure. The sophistication has grown from early rule-based systems to statistical models to machine learning, but the core principle remains unchanged: maximize revenue by filling seats and charging fares that reflect demand.
Related coverage: How Theatre Tickets Are Actually Priced; How Airline Status Actually Works Now.
