Before it was absorbed by Alaska Airlines, Virgin America fashioned itself a business person’s airline, focusing on tech-forward travelers willing to pay hundreds or thousands of dollars to fly last-minute to and from San Francisco and Los Angeles.

Because techies, consultants, producers, directors, and actors in Silicon Valley and Hollywood could pay two to five times as much as leisure travelers, Virgin America set aside seats for them, hoping they would snatch them up a few days before departure and boost profits.

When Alaska’s revenue managers examined the data, they didn’t like what they saw. Sometimes, the high-value customers materialized. Other times, they didn’t, and the airline may have lost revenue it could have earned weeks earlier.

Alaska changed the strategy. On former Virgin America routes between major cities, it reduced the number of seats held for last-minute travelers, increasing availability for leisure passengers who book further in advance but demand cheaper prices.

“Our traditional comfort level was to take them further out and so we adjusted our booking practices there,” Alaska Chief Commercial Officer Andrew Harrison told investment analysts in January.

Alaska is one of two U.S. airlines to tweak its booking curve strategy recently. The other, United Airlines, went the other way, saying it had erred by leaving too few seats for big spenders. On core business routes, it’s probably tougher to score a deal on United than a few years ago.

The struggles to get it right at Alaska and United highlight an ongoing dilemma for revenue managers. Leisure travelers book early, and often demand deals. But most airlines don’t want to sell out at $49 or $99 fares. They must hold enough seats for the lawyers, consultants and bankers who will pay astronomical prices. But how many?

To decide, airlines look at basic indicators, including route, day of the week, and time of year. They also plug information about past bookings into sophisticated computer systems, which spit out recommendations.

But humans guide the computers, and some management teams have more tolerance for risking empty seats than others. Some carriers prefer to book more revenue early, giving them certainty, while others wait for big spenders.

“The whole concept of revenue management is that you get a call from a customer and you have to decide whether or not to take it,” said Tom Bacon, a consultant and former vice president for revenue planning at Frontier Airlines. “It is a bird in the hand issue. You can say, ‘Nope I am not going to take this booking at $79 because I know I can get a higher fare passenger to book if I just wait.’”

United’s Case

United’s journey to taking more last-minute revenue began after Scott Kirby joined as president in August 2016.

Early in his tenure, Kirby told investment analysts the airline had an antiquated revenue management system that could not predict with enough certainty which customers would show last minute, so it preferred to fill planes early.

And even when computer recommended United wait for better revenue, United’s revenue managers might override it. They feared having too many empty seats, Kirby said.

“It’s easy to sit around and say, ‘Oh my God, December load factor is off by x points,’” Kirby said three years ago, pitching the need for a better system to analysts. “[People say] ‘We’ve got to go open inventory or lower the fares.’”

Over the past three years, United had updated its revenue management platform, now called Gemini. United sets aside more seats for its most lucrative passengers, and sells fewer deeply discounted tickets months in advance. Also, United lets the system work as designed — with limited human interference.

“We continue to successfully shift our booking profile to reduce dependence on lower-yielding tickets booked further out from departure, while increasing our share of higher yielding business tickets generally booked closer in,” United Chief Commercial Officer Andrew Nocella said on the airline’s January earnings call.

Indications suggest it’s working. In 2015, United filled 83.4 percent of its seats, with an average yield of 15.72 cents per passenger mile. In 2018, load factor was 83.6 percent, with yield of 16.38 cents. In October, analyst Hunter Keay of Wolf Researched credited the system with “driving high loads and close-in yields, as designed.”

How does United know how many passengers will show?

It is a guess based on math. United may not sell all seats it leaves open. But the airline has calculated it will make more money attempting to sell them than if it put the seats on sale six weeks prior to departure.

“These models try to come up with the best possible guess or demand predictions, and it is a statistical estimation,” said Surain Adyanthaya, senior vice president of Strategy for PROS, a revenue management consulting firm. “It is a combination of the mathematics along with the strategy the airline wants to employ in that market.”

If the demand does not materialize, an airline can sell cheap seats last minute. But since no premium airline wants a corporate lawyer to snag a low fare three days before departure, the carrier may decide to leave the seat empty. (Another option to sell only basic economy fares cheaply just before departure, since business travelers rarely buy them.)

“We don’t actually achieve 100 percent load factors despite demand that might suggest we could get 100 percent load factors,” Bacon said. “That’s because we have set aside those last three seats for high fare passengers with only a 60 percent chance that they will be filled. The probabilities and statistics suggest that is a better result.”

Different Philosophies

United and Alaska probably should have disparate philosophies about when to book more revenue, since they have different business models.

United has hubs in five of the top six U.S metropolitan areas, as measured by the U.S. census bureau. It carries a disproportionate number of business travelers, many of whom book in premium cabins. Alaska has just one major hub, in Seattle, along with smaller ones in Los Angeles, San Francisco and Portland, Oregon.

Alaska, which is more of a leisure airline than United, is not as confident last-minute business travelers will show, Shane Tackett, executive vice president of planning and strategy, told investment analysts on the airline’s recent earnings call.

“The trade-off is if you hold out for yield or not, and then you sort of hope that it comes and when it doesn’t, you feel bad about your strategy and when it does you feel great,” Tackett said. “We traditionally like to manage that just further out the booking curve.”

For airlines, early-booking passengers have benefits, too, Adyanthaya said. They may not help a carrier produce massive margins, but they “give you certainty to keep the planes full and keep the machinery running,” he said.

In addition, Adyanthaya said, some low-cost-carriers find early-booking passengers more valuable than ever. When customers book three months in advance, an airline has 90 days to ask them if they want to add hotel rooms, rental cars, seat assignments, baggage and onboard meals to their bookings. All are high-margin items.

“That changes the equation a little bit,” Adyanthaya said. “Now you want the planes as full as possible. There is a bigger cost to spoilage, since you are not getting the potential ancillary revenue from having a full flight.”

But it can be more complicated, Bacon noted. Some low-cost-airlines call themselves “spill carriers,” thriving by selling to customers who cannot afford a legacy airline’s last-minute fare.

These airlines may save tickets for the end of the booking curve, waiting until full-service carriers hike their prices. If a major airline asks for $400 one-way a next-day flight, a spill carrier might charge $300.

Market dynamics play a role, too. Most airlines will sell more tickets early in the booking curve on pure leisure routes, as vacationers rarely book late.. United’s strategy on Los Angeles to Honolulu is likely much different than from Los Angeles to Newark.

Where an airline is based also matters, Adyanthaya said. In some countries, airlines place an emphasis on filling every seat. In others, that’s less important.

“You’ll see some premium international carriers that really strive to have the best possible product and they value their product very highly and they are willing to wait in the booking cycle to get the demand for their flights and risk some spoilage, whereas the North American and European carriers are more aggressive in just filling the planes,” he said.

Other Factors

U.S. airlines have reported record or near-record profits since 2013, so carriers can afford to be choosy about which passengers they want.

What about during rougher periods?

Other factors come into play, Bacon said. In 2008 and 2009, when he was running Frontier’s revenue management department, the airline was in bankruptcy. It needed cash, and Bacon had to find it.

“We were managing cash flow every week, monitoring whether we had enough cash to keep the airline going, so booking early was pretty important to us,” he said. “Every change was a crisis. We wanted to be aggressive and get the bookings in early.”

Sometimes, the situation can be out of an airline’s control. During the financial crisis, Frontier noticed passengers weren’t booking Christmas season flights months in advance as usual. The airline, Bacon said, was fearful it would miss out on one of its most lucrative periods.

That’s not what happened. Perhaps fearing the crisis would worsen, leisure travelers waited. But as the holiday season approach and travelers had to decide whether to skip travel or go on the usual trip, most booked.

“The whole booking curve had changed,” Bacon said. “Our revenue was right on track even though we hadn’t got the bookings as well as expected.”

Kirby said he’s seen similar changes, which is why he tells his revenue management team to trust the computer. Just as long-term investors should not panic and sell all their stock when the market drops for a couple of weeks, revenue managers should not change focus because they’re not selling as many tickets as expected, he said.

“Every single time the revenue environment has turned in my career, there was a month or two where you were looking out and saying, ‘Oh my God,’ and worried about it,” he said. “But we had faith in the system and had faith in the forecast and it turned,” Kirby said.

Photo Credit: Alaska Airlines has changed its revenue management strategy for former Virgin America routes. Pictured is a former Virgin America Airbus A321. Alaska Airlines