The $215 round trip fare to Atlanta looks good. You call the airline but discover that seats are no longer available at that price. You can buy a ticket for $274. Or you might call back next week and find the $215 fare available after all.
Confused? Blame it on technology -- and the airlines' urgency to squeeze the most money out of each flight.
On thousands of flights every day, computers constantly crunch data to help carriers figure out how many tickets to sell at what price and how many to overbook.
The trick for the airlines in allocating fares and seats is making sure they don't get too greedy -- and that they don't sell too cheaply.
"It's not a case of gouging the consumer. It's a case of managing the product," said Kenneth R. Stephens, president and chief executive officer of Behavioristics Inc., a small College Park firm that provides software for USAir Group's revenue management program.
"Just by selling the seats a little smarter, USAir saved $140 million last year," he said.
Pioneered by American Airlines nearly two decades ago, revenue management techniques have been adopted by a wide range of industries, including long-distance telephone carriers, cruise ship lines, hotels, even opera companies.
In the airline industry -- where profit margins are paper-thin even in the best of times -- the tool is critical, saving the nation's airlines hundreds of millions of dollars each year. Indeed, some analysts believe USAir's lag in developing a sophisticated revenue management system is partly to blame for the airline's huge losses since 1988.
Last week, USAir, the largest carrier operating at Baltimore-Washington International Airport, said it expected to make a profit this year, its first since 1989. And its yield, or income per passenger, has increased -- the result of a higher percentage of seats filled by people paying relatively higher fares.
"This is the airlines' main tool for optimizing their revenues," said Jon F. Ash, managing director of Global Aviation Associates, a Washington aviation consulting firm. "Absent a yield management system, you forfeit the ability to maximize revenues."
The reason is simple: once the plane leaves the gate, every empty seat is revenue lost forever.
Airlines need to sell as many seats as they can all the time. But they try to fill the plane with the greatest number of high-paying passengers before freeing up seats to those who pay less.
"People may find it frustrating to call one day and get one fare, then another fare another day," Rocky B. Wiggins, USAir's senior director of inventory management, said. "They say the airline is playing games with me. But in order to make money, we have to have a mix on the plane."
As a result, passengers on the same flight can pay a half-dozen or more different fares. And for many consumers, that's perplexing.
"Some of the savvy travelers understand, but most people don't," says B. J. Cook, assistant manager at Roland Park Travel in Baltimore.
She taps the keyboard on her computer terminal, searching for the best deal from Baltimore to Atlanta on Oct. 23. An array of 13 different fare categories darts across the screen, ranging from $219 to $813 for seats on the same Boeing 737.
What Ms. Cook doesn't see is how many seats are available at any given fare. That information is closely guarded by every airline and available only to insiders, like reservation agents.
But experience tells Ms. Cook what the computer doesn't.
On a Friday afternoon flight between Washington, D.C., and New York, for instance, airlines rarely sell discount seats. Likewise, Dec. 23 flights to San Juan, the Caribbean connecting point, are rarely discounted.
"But a flight to Minneapolis during the winter? A whole different story," Ms. Cook says.
Bookings change from day to day, even hour to hour. Typically, airlines will free up more discount seats if the plane is undersold closer to departure.
Ms. Cook, for instance, recently tried to book a $179 one-way excursion fare from Baltimore to Tampa on a 4:30 p.m. flight. No seats were available at that price, but she was able to nab the fare on an 8 p.m. flight. A week later, she discovered seats were available for the price she wanted on the flight she first sought.
Discounts are scarce
Confusing or not, airlines don't give up discount seats frivolously.
"You sell too many lower-fare levels, you don't make money," Mr. Wiggins said. "You don't meet your costs and you've taken up seats that last-minute, high paying passengers might need."
Without yield management, he says, prices would be much higher. "The system really benefits the traveling public to a great degree. . . . It allows the airlines to offer more cheap seats with confidence."
Typically, the least expensive tickets carry the most restrictions -- such as advance purchase requirements and no refunds. Passengers pay a premium to get tickets with no strings.
It's no secret that the best low-fare travel times are mid-week and Saturday. Even low-cost airlines, like Southwest, that offer the simplest fare structure juggle seat availability.
"You have to limit the number you put on at any given price to turn a profit," said Joseph W. Lamkin, area marketing manager for Southwest in Baltimore. "That means the flights most easily filled, like Friday night to Chicago, will have the fewest number of $89 seats."
The staggering number of daily flights -- 20,000 nationwide -- also reveal why effective revenue management can be so lucrative. USAir, for instance, flies 5,000 flights a day, many with three and four segments.
"You can add one passenger per airline and make a huge difference in revenue," said Richard D. Niggley, executive vice president for Aeronomics Incorporated, an Atlanta-based firm that provides yield-management systems for companies around the world.
Likewise, charging one passenger $20 more on 1,000 flights a day is significant. "The numbers roll up very quickly," he said.
Each flight has historical patterns and a host of variables -- such as time of day, day of week, origin and destination -- that influence demand.
Three hundred days before a flight departs, USAir's revenue management system, known as Excalibur, begins to track the flight. It starts with a bank of historical data on that flight. Before it actually takes off, Excalibur will review the flight 45 different times.
"Closer to departure, those review points get closer and closer together," says USAir's Mr. Wiggins.
Every night, between midnight and 3 a.m., information, gleaned from the reservation system, is fed into the system. The so-called adaptive neural network processes the data and revises its predictions.
"If we do all this accurately and the plane still doesn't generate enough to make money, then we have to think about eliminating that flight," says Mr. Wiggins.
But the computer is only part of forecasting flight demand.
At USAir's marketing services office in Winston-Salem, N.C., 70 revenue analysts, assigned to geographical areas, work in cubicles before powerful computer terminals. They constantly review Excalibur's predictions for flights.
"They're always asking themselves, 'What do I know that's not forecast here?' " Mr. Wiggins said.
Typically, Excalibur knows how the Christmas holidays affect demand for Florida flights. But it doesn't know that Hurricane Luis is sitting 200 miles off Puerto Rico. Analysts, however, know that such conditions could reduce the demand for seats and increase the no-shows.
Annual events, like the Kentucky Derby and the Super Bowl, create historical patterns in the data bank. But airlines rely on their analysts to plug in data that predicts demand for special events, like the pope's visit to Baltimore.
"Mathematical algorithms guide you, but no data system can predict the full impact of that," said Mr. Niggley.
Not only does the system allow airlines to change seating availability, it often helps them determine where to shift aircraft. For instance, during the World Cup USA 1994, United Airlines and others constantly updated its system with the latest information about game results. That helped the carrier shift planes onto high demand routes as fans migrated to the next city with a winning team.
Likewise, local economic conditions, such as a plant closing, may affect demand for air travel in certain markets.
The system also tracks what other airlines are doing. An unanticipated fare sale by a competitor can leave an airline with half-empty planes unless the airline quickly adjusts its own fares.
SG By examining the history of no-shows on every flight, airlines also
decide how much to overbook a flight to minimize the loss of revenue. The percentage of overbooking varies flight by flight.
While the computer system reveals much about travel patterns, it also shows a lot about people and cultures. Typically, no-shows will be much higher, for instance, on a flight coming out of Las Vegas on a Tuesday afternoon than one leaving Des Moines. Gamblers often find it difficult to walk away from table just because their plane is waiting.
No-shows may be relatively high on New York-to-Washington flights but not out of Minneapolis. "Folks in Minneapolis don't just no-show, they cancel," says Mr. Stephens of Behavioristics.
With the power of computers -- and generations of technology yet to come -- revenue management techniques may someday allow consumers to bargain with the airlines for a fare.
"Right now the industry is still structured around filing fares and publishing fares, Mr. Wiggins said. But someday, he predicted, fares will be negotiable.
"Eventually, we will get to that day," he said.