It would have been an ordinary business flight. William Harris and four coworkers needed to make a quick trip from Birmingham, Ala., to Charlotte, N.C., this past winter. Harris, a national account manager for BellSouth, checked with several different airlines. The best price he could find? A $950 ticket. “I could have flown round trip to Europe for $300, but I couldn’t get to Charlotte for three times that price,” Harris says. “It just didn’t make sense for us to spend $5,000.” They drove the 400 miles instead.
Call it the mystery of the Saturday-night stay.
Airfare rules have always been puzzling. Some fare differences are obvious: First class costs more than coach. Others are harder to understand — so hard that airlines have invested millions in massive computer systems to determine who gets which fare.
The process is called revenue management, a science the airlines have been honing for 30 years. Executives and researchers wield intricate forecasting models, searching for strategies that will match different types of customers with fares they’re willing to pay.
It is a delicate balance. While airlines aim to fill every seat, they also seek to ensure that some full-fare seats are available for last-minute travelers.
The easiest way to do that is to divvy up seats by price and availability: a certain number can be sold within 21 days before a flight, within 14 days and so on. That split into “buckets” is likely why a ticket agent tells you a flight is sold out — but check back in a week. A last-minute ticket can often be exponentially more expensive, airlines maintain, because you’re not just paying for the flight — you’re also paying for them to shoo away other customers and save your seat until the moment you buy it.
“Sometimes that gamble doesn’t pay off,” says Columbia University professor Garrett van Ryzin, who studies revenue management models. “Sometimes they reserve five seats or 10 seats for people at the last minute and they take a loss.”
Row 17, Seat B economics
The concept has some legs. In 1972, an employee of BOAC (the predecessor to British Airways) named Kenneth Littlewood devised a decision rule for the process of selling cheap airline tickets. Littlewood’s concept was straightforward: More people will buy discount tickets than full-fare tickets, so sell discount tickets only until your revenue matches the potential revenue you’d earn from a smaller number of full-fare tickets. The trick, of course, was to determine how many full-fare tickets you would be able to sell.
Over time, other airlines joined the game and revenue models became more complex. They decided it was necessary not only to charge different fares but also to set up restrictions, or “fences,” that would separate buckets using basic psychology about travelers and their perceptions of value. The Saturday-night rule, for example, was created as a simple method to weed out leisure travelers from the lucrative business travel market. Nonrefundable tickets were created for a similar reason.
The guy in the next seat may be on the same flight as you, but airlines certainly don’t see it that way. They view these things as different products, says van Ryzin. “If you’re willing to buy 30 days in advance and fly at off-peak times and put up with nonrefundability, that’s a different product.”
By the early 1990s, the range of fare options was dizzying. American Airlines, for example, improved its predictive models to the point where it operated with at least six or seven buckets per flight just for coach seats. Those divisions comprised not only fare codes — the “Y” that represented the serious cash for a full fare, the “M” or “Q” that showed a cheap flight — but also the dates when those fares were available to a customer.
None of these divisions were meant to be obvious when you booked that trip to see your folks. What customers saw, though, was an inconsistent and seemingly random set of fare options. A cottage industry of travel advice was born, trading hints on how to beat the system.
But in a turn that might bring joy to the disgruntled flier’s heart, the old models are breaking down. Price ratios between the lowest and highest fares have spread as far as 20 to one, which is a difficult split for anyone to explain. The full-fare traveler has become increasingly difficult to woo, in part because of the economy and in part because of low-fare competitors.
The simplicity of the low-fare world actually threw a huge wrench in the gears of revenue management. Low-fare carriers do have models to manage revenue, but they’re very simple: Prices tend to step up steadily as a flight date nears, and travelers who want flexibility — a refundable ticket, perhaps — pay more.
“Once you go to the low-cost airline model, you don’t have fare buckets anymore, you have one continuous pricing model,” says Andy Storey, managing director of European operations for the Rubicon Group, which helps such airlines as British Airways and Delta manage revenue.
Despite what you might think, the impact of low-fare airlines on major carriers isn’t in the lowest fares: It’s in the middle range. A traveler can book three days out on Southwest or Britain’s EasyJet and get a mid-price fare, while a full-fare ticket on the competition costs hundreds more.
That, says Rubicon president Steve Swope, is because major airlines usually have two different departments handle fares. Pricing managers scour competitors’ rates and route popularity to set actual prices. Meantime, revenue managers tweak their intricate mathematical models to decide how many customers will qualify for the various fares. With those tasks separated, Swope suggests, it’s hard to match prices with the number of customers who will qualify for them.
“When they get down to that last seat or the last two seats at $1,500, that makes sense. But try offering the last 50 seats for $1,500? That’s a mistake,” says Swope. “The argument that the airlines are making that ‘I was reserving those seats for a higher fare’ doesn’t hold up.”
As an alternative, Swope points to models used by his other clients: rental car companies. While car rentals also step up their “price points” for different customers, they have far more flexibility — partly because their models are based on renting almost every car in the fleet, not applying fixed prices to the cars. And because car rentals have much smaller ranges for pricing, they have to be more flexible.
Even execs look for bargains
Someone else also gets credit for the changes: business travelers. Enormous gaps in ticket prices existed for years largely because corporate travel departments were willing to pay those top rates. Airlines would use those high-priced tickets to offset the discount fares they promoted. But the shrinking economy tightened travel budgets and companies scrambled for discounts — forcing executives to stay over on a Saturday night, or buying blocks of cheap, nonrefundable tickets and throwing some away. And many realized the cost savings should be permanent.
In many cases, that meant corporate bean counters were slashing what Montrose Travel President Joe McClure calls their “magic number” — the amount of extra money a company would allow employees in booking preferred flights. While some of McClure’s clients once allowed hundreds extra to be spent so an employee could get their preferred frequent flier miles or a better flight schedule, magic numbers have shrunk to as little as $50 over the lowest offered fare. Business travelers’ subsidy of leisure travelers has mostly ended.
“The majority of our travelers understand the Saturday-night stay rule, they understand advance purchase and they understand the difference between a leisure and business fare,” says McClure. “They’re tired of being gouged and they’re tired of being taken advantage of.”
A ticket buyers' market
Actually, many airlines are already changing their methods to accommodate the shift. American Airlines’ latest system, devised by Sabre Inc. out of the research that once expanded on Littlewood’s theories, has largely abandoned fare codes in favor of a “bid price” — the amount American has to earn on each seat in order to sell it. The new system accommodates the proliferation of new ways to buy plane tickets, including discounters and online markets such as Priceline and Hotwire, which sell “distressed inventory” that airlines would otherwise write off. If any seller can match the bid price, the ticket can be sold.
“Some consumers are much more price sensitive, and there are easier ways to shop price today than there were 10 years ago,” says Barry Smith, Sabre’s chief scientist.
To researchers like Smith, it became obvious over time that the old method — locking buckets of seats into specific fares — left airlines with empty seats, losing money. Regular travelers would see the empty seats, look at the extra zeroes on their tickets, and be puzzled. A quick glance at a low-fare airline’s Web site, and puzzlement turns to anger.
That lack of flexibility in the old models was precisely what William Harris discovered. Granted, he wanted to fly on a weekday without remaining for the weekend, and to book within a month of the flight. In fact, he would have paid extra to indulge his last-minute planning.
“I think if they had told us it was $400 a person to fly,” he says, “we would all have gotten on the plane and gone.”