Why Do Airline Prices Change All The Time?
An intro to Yield Management Pricing
Most people have had the experience of going to book a flight, whether for vacation or business, and finding that the price is significantly different from other times you’ve flown the same route. Why do airline prices change so much and so often?
The airline industry was one of the first to invest heavily in complex computer-driven pricing models. The industry refers to these optimization models as Yield Management or Revenue Management, though the latter term is used across many industries to refer to optimizing pricing and profitability in a much more general sense.
The reason airlines invested so heavily in developing sophisticated pricing models springs from the nature of air travel itself. Think of how a commercial flight works. It has a specific number of seats, and it leaves as a specific time and follows a specific route. The cost of operating a flight is quite high. And the cost of flying a half full plane is only slightly less than flying a full one. So, it’s very important for an airline’s profitability that it fill as many seats as possible while getting the highest possible price for each seat filled.
Start out by thinking of the basic problem an airline has. For a given flight, say from my own Columbus, OH to Daytona Beach, FL for Spring Break, the airline has already made a decision about whether to have a flight and what size plane to use. Say they have a 737 with 170 seats on it.
They could fill all those seats immediately if they offered them at $100, but they would surely lose money at such a low price. If they priced them at $800 per seat, they would fly with a mostly empty plane. The airline is trying to figure out the right prices to offer at the right times to fill the plane without leaving money on the table by over discounting.
Airlines start out with a regional and seasonal demand model of how quickly flights between given points are likely to fill up at different times of the year. For instance, that Columbus to Daytona Beach flight would have been particularly expensive this week, because this is the week when Ohio State has its spring break. Tens of thousands of college students wanting to hit the beaches during the same week is predictably higher demand, and so airlines would have set prices higher from the beginning.
Part of that seasonal demand model is an expectation of how quickly a flight should fill up. The airlines’ computer systems will be constantly checking to see whether the flights are filling up as fast as they ought to be. If sales are slow, they will automatically offer more discounted tickets.
The market for airline tickets is what is called a highly transparent pricing market. This means that it is easy for customers to see the prices of different companies and compare them. When someone logs onto Travelocity or some other travel portal, they see the prices being offered by nearly all the airlines simultaneously. Airlines can also see each other’s prices in real time.
This means that it’s easy for airlines to see their price position versus other airlines, and if an airline lowers their price, it’s easy for a potential customer to see the value of that discounted fare.
Because the market for airline tickets is so transparent, even fairly small changes in price can have significant effects on the demand for tickets on a given flight. Customers might make buying decisions based on a difference of just a few dollars.
Just as airlines will discount if sales are behind on track for a full flight, similarly if demand for a specific flight increases, airlines will notice that the flights are filling up faster than they expected and they will increase the prices.
There’s not a significant advantage to an airline selling out too early, and there is a significant advantage to getting the maximum revenue for each flight, so if the data starts to tell airlines that lots of people are urgent to get to a particular place on a particular day, they will automatically increase their prices to take advantage of the demand.
I saw this happen in real time some years ago when the company I was working for scheduled a national sales conference in Palm Springs. This meant that in the hours after the company email went out telling everyone to book tickets, several thousand people from all over the country simultaneously started booking tickets into that one relatively small airport on the same day.
Over the next 48 hours, I watched the price of tickets to Palm Springs for that day more than double. The airlines didn’t have to know that a Fortune 500 company had scheduled a sales conference near a smaller regional airport. Their demand models told them that flights were filling up much faster than they normally would, and the automated pricing model then increased the price in order get as much revenue as possible in the situation. (After watching the tickets for the sales conference days go from the low $300s to over $800 over the course of a couple days, I went to my boss and pointed out it would save the company money if I few out to California a couple days early to visit my family first.)
There are a number of other types of sales which have somewhat similar dynamics. Hotels vary their prices based on demand to sell more rooms at low demand times and get the most for their rooms at high demand times. Tickets for sporting events and concerts vary to some extent based on demand, though the real extremes in those markets come from the secondary market.
If you were hosting an event of some sort, and you found that it was not filling up as fast as you had planned, you might discount the tickets for the event.
However, airlines have a uniquely challenging business model with their different routes, classes of ticket, and the high cost of each flight. And so airlines have invested in this kind of pricing technology to a unique degree.