Dynamic Pricing vs Fixed Fares: Why an Airline Seat Price Can Change Every Hour

Airline fares are dynamic because seats are perishable, demand is uncertain and each flight has a fixed capacity. The price shown at any moment is the outcome of a revenue-management system deciding which fare products to offer, how much inventory to expose and how strongly demand appears to be developing.

Dynamic pricing does not necessarily mean an airline changes the price for every individual search. More commonly, it changes availability across pre-filed fare products or applies an offer rule when the system receives new data about demand, inventory or competition.

Fixed fare versus dynamic fare

A fixed fare is stable for a specified period. It is easy to understand but cannot respond quickly to a flight that suddenly starts selling faster than forecast, loses a competitor or faces an unexpected demand shock.

Dynamic pricing updates prices or available fare classes based on state variables such as remaining capacity, days to departure, booking pace, market demand, forecast uncertainty, competitive fares and itinerary value. In mathematical terms, this is a finite-horizon stochastic control problem.

The state of the flight matters

Two flights on the same route can have very different prices because their booking state is different. A Monday-morning business flight with strong late demand may protect higher fare classes. A Saturday afternoon leisure flight with weak bookings may retain lower fare availability for longer.

A useful simplified state vector is:

• Remaining seats by cabin and booking class
• Time to departure
• Observed booking pace
• Forecast demand distribution
• Competitor price and capacity signals
• Cancellation and no-show expectations
• Network opportunity cost of the seat

Why price-response data are difficult to interpret

Airlines raise fares when flights are filling and reduce fares when demand is weak. That creates a classic endogeneity problem: price and demand move together because both are reacting to underlying market conditions.

A naive predictive model may observe high prices and high sales on holiday flights, then incorrectly infer that increasing the price caused sales to rise. The real driver may be the unobserved holiday demand shock.

A recent Operations Research paper illustrates this problem using airline ticket pricing. It proposes proxy-aided causal inference using variables such as ticket-page views and fuel costs to estimate the causal effect of price on sales when demand shocks are partly unobserved. Read the original study.

Demand arrivals are not always independent

Standard pricing models often assume that customers arrive independently. In reality, demand can be self-exciting: a promotion, event, social-media post or sudden change in travel conditions may create a burst of related searches and bookings.

Research on self-exciting demand processes finds that the optimal response depends on whether the market is in a growth stage or a saturation stage. In a growth stage, a higher level of excitement can support a higher optimal price; in a saturation stage, it can support a lower price. Read the original Operations Research paper.

Why a price can change after you search

The most likely explanations are that a lower booking class sold out, a fare rule expired, the revenue-management system recalculated availability, a competitor changed price, or a scheduled data update occurred. Persistent searches can affect website personalisation or marketing, but they are not the only—or usually the main—reason an airfare changes.

Data-science finding

Dynamic pricing should be treated as a causal decision problem, not just a forecasting problem. A model that predicts bookings well can still recommend poor prices if it cannot distinguish correlation from the true price elasticity of demand. Airlines therefore need experimentation where feasible, robust causal methods where experimentation is not feasible, and guardrails around availability, fairness and customer communication.

Glossary

Dynamic pricing

Adjusting prices or available fare products as demand, capacity and market conditions change.

Price elasticity of demand

The percentage change in demand associated with a percentage change in price.

Endogeneity

A statistical problem in which price is influenced by the same hidden factors that influence demand, making simple correlation misleading.

Causal inference

Methods designed to estimate what would happen under a different decision, such as a different fare, rather than merely predict observed outcomes.

Finite horizon

A decision period with a fixed end point; for airlines, the horizon ends when the flight departs.

Self-exciting demand

A process in which recent demand raises the likelihood of further demand, creating bursts rather than independent arrivals.

Opportunity cost of a seat

The expected revenue sacrificed by selling a seat now instead of saving it for a potentially more valuable later booking.

Want to apply these airline economics ideas yourself? OpenSky by Papaeya turns route planning, pricing and airline trade-offs into a playable strategy game.