Analytics Helps Southwest Airlines Rev Up Results

Operating an airline efficiently and profitably is a huge challenge. Business results depend on a mind-bending array of factors, including scheduling aircraft and crews, procuring fuel and making real-time adjustments due to weather and other factors.

At Southwest Airlines, which offers scheduled service to 97 destinations and operates 700 aircraft and 3,500 flights daily, business optimization is at the center of everything. “It’s unwise to make local decisions without having a big-picture view of things,” says Rusty Burlingame, senior operations research developer.

Over the last two decades, Southwest Airlines has used optimization modeling and analytics software from FICO to assist with important and sometimes critical decision making. “Optimization helps us trim between 3 and 5 percent from big-ticket items and more than 5 percent in several areas,” he reports.

Yet, even relatively small savings add up to major dollars in areas such as fuel costs, which total as much as $5 billion annually, and labor costs for 50,000-plus employees, which are even higher. “It’s an extremely competitive industry, and we have to make the best of our resources,” Burlingame points out.

Analysts Can Make the Right Decision on the Spot

For instance, the airline, which buys fuel from several sources that bid at different prices, examines spot pricing at different stops during the day. The typical aircraft averages approximately seven flights each day.

“If a particular fuel station is less expensive relative to the type and weight of the aircraft, we can buy extra fuel and carry it down-line so that we don’t have to purchase it later,” he explains. In fact, analysts can make the right decision on the spot using the FICO Optimization Modeling tool—depending on whether the aircraft is an older model or a newer plane that is extremely fuel efficient. The data is typically updated every 15 minutes.

Another area in which Southwest uses the software is gate optimization. With the software, analysts can determine the right gate to send a plane to when it arrives.

“The goal is to minimize gate conflicts and tarmac delays and to ensure that planes are at a gate that allows a fast turnaround,” Burlingame says. “Passengers hate gate conflicts and changes. Even if you arrive 25 minutes early and sit on the tarmac for 10 minutes, they often wind up annoyed. Anything we can do to prevent these issues is beneficial for everyone.”

In the future, the airline is looking at plugging in all gate transfer data for passengers and including that in the decision about what gate to assign an aircraft.

Southwest Airlines stores most of its data in an enterprise data warehouse. It depends heavily on APIs to feed data from various external and internal data sources. This high level of flexibility has proved revolutionary. Last November, during a major blizzard, it used the system to deliver immediate insights about aircraft, crews and passengers.

“Combined with more advanced passenger communications capabilities, we are able to run the business in a way that wasn’t possible in the past,” Burlingame says. “We have analytics and dashboards that put resources to use in the most efficient way possible.”