United’s Strategy to Reduce Overbookings: After a series of customer-service fiascoes, the airline is looking to big data for answers
United Continental Holdings Inc. has been in the news a lot this year, and not always for good reasons.
Chicago Department of Aviation officers forcibly removed a passenger from a flight in April, an incident that blew up on social media and raised broader questions about how the airline handles overbooked flights. In February, the airline delayed about 500 flights after suffering a second glitch in its computer systems in just over two weeks.
Nevertheless, United is benefiting from strong demand for air travel and, like other airlines, has thrived by offering a wider range of fare categories and collecting fees for increasingly popular extras like priority boarding and seat upgrades. In July, United, the No. 3 airline in the U.S. by traffic, said net income in this year’s second quarter climbed 39% from a year earlier, on a 6.4% increase in revenue.
Linda Jojo, United’s executive vice president of technology and chief digital officer, says the company is using data analytics to help manage overbooking situations more strategically, while also helping it target families, business travelers and vacationers with specific offers.
Ms. Jojo spoke with The Wall Street Journal about those efforts and others on the digital front, including how United is using mobile devices to improve its interactions with customers. Edited excerpts follow.
WSJ: How does technology help United address problems with overbooking?
MS. JOJO: The vast majority of [overbooking situations] are due to some kind of operational disruption that causes us potentially to fly a slightly different plane than the one we had planned. They might shift my flight from a 737-900 to a 737-800, which is fundamentally the same but not the exact seating situation, and that could cause an overbooking situation. That could happen fairly close to the time of departure.
The biggest thing we do from technology is help run the airline better [in those situations]. How do we help with anticipating issues so that if we do see a thunderstorm, we can help with proactively rescheduling and talking to customers about the fact that weather could disrupt their travel? If we can get to people early so they can change their travel plans, that’s one of the biggest things we do.
The situation of overbooking that I think people go to first is when we oversell a flight. The way we do that is looking at information about when people actually book a flight and show up. This is where technology can help make that analysis much more precise.
We know that people flying between two business locations on a Monday morning are probably going to show up. We know people who are flying to a resort location are probably going to show up. But there’s a certain percentage of people who don’t set their alarm right, or hit traffic and don’t make that early-morning flight. We also know the last flight of the day, everybody’s there to get back to their city. So what we’re doing is helping further refine the analytics to understand which flights have a higher probability of no-shows. That’s basically one or two [seats oversold]. If the people who bought those last two seats in that oversell situation are actually there and get on the plane because two other people changed their plans for whatever reason, we’re actually helping those customers.
WSJ: Some airlines have said they are going to stop overbooking. Is this something you plan to do?
MS. JOJO: In April, we announced 10 new initiatives to improve the customer experience, including reducing overbooking, which has helped to dramatically reduce involuntary denied boardings by more than 80% from last year. Overbooking is an example of where today’s new techniques of machine learning and analytics can help us stay away from black-and-white rules and look at past behaviors, such as which flights have the most number of historical no-shows, and current events—is there something unique going on in a city—and situationally determine if a specific flight should allow overbooking or not.
WSJ: What’s a tech project you’re working on now?
MS. JOJO: When someone is booking a seat on our airplane, what are some of the things they’re likely to want to purchase in addition to that seat? We’re trying to look at different situations: Is this a business traveler, a leisure traveler, a family? Based on those different personas, we’re looking to offer different bundles of features.
We’re still working out what’s the best bundle for a family, for example. Is it Wi-Fi on the plane? Probably not, that’s more for a business traveler. Is it checked baggage and how we could bundle that? Is it time in the lounge? Is it prepaid food? It’s all of the things you could look at to try to see what customers would be likely to buy in certain situations.
Seat map on mobile
WSJ: What role do data analytics play?
MS. JOJO: One thing all airlines have is a lot of data. We are looking to get the information and insights into the hands of our employees. Now, passengers might see our flight attendants with mobile devices on the plane. One use is to let customers buy food and beverages in the economy cabin. But the other thing on that device is a seat map with some very-easy-to-look-at views of our passengers. We can see what loyalty status they are, and we can see what their next connection is if they have one. In the seat map there’s a little airplane, and it turns red if they have a tight connection. If we get to a place where we’re landing at a hub and trying to get people to where they have to go, we know who the passengers are that need to get off that plane and get to their gate.
WSJ: What else is happening on the mobile front?
MS. JOJO: If there’s a large thunderstorm going through an airport causing a lot of flights to be changed, you used to see long lines of people at a gate or customer-service area. We were limited by the number of computers that were in that room. We now have mobile devices and mobile printers that have been issued to all of our customer-service reps so they can swarm an area and service our customers, change their flights if they need to, check information if they need to, or print out boarding passes and bag tags. So we’re now not limited by the infrastructure in the airport.
WSJ: What’s the biggest challenge to implementing all this technology?
MS. JOJO: What’s somewhat unique about airlines is we are operating all the time. We have an airplane in the air all the time, and so we don’t have a period of time where we can take systems down and try things. Thinking back to some of the jobs I’ve had in other industries, you often looked to long holiday weekends as a gift to the IT department to make changes. [At United] that’s actually often our busiest time. It’s how you make these changes in a way that doesn’t disrupt the operation.
WSJ: Are you using biometrics?
MS. JOJO: Biometrics and those kinds of things are big, large, innovative ideas that will at some point impact enough of our customers that it will make a difference. But there’s so much infrastructure required by other people to actually make that happen. I like to focus on smaller, incremental things that can impact a broader level of customers.
We’re looking at making our app a little more context-sensitive. If we know you’ve already checked in, is there a way to make our app more user-friendly at that period of time? If we know that you’ve gone through security, are there things that we can do? We’re looking at those kinds of things. Our ideas are around creating a more frictionless travel experience that could include things like improving the boarding process and providing better, more specific information about the status of your flight.
WSJ: What is your team focused on over the next few years?
MS. JOJO: What we’re trying to do is expand our thinking to the customer journey. The customer starts thinking about their flight on United from the time they check in until the time they get to their destination, whether it’s home or the hotel or the office. How we can create a frictionless experience from the beginning to the end of that process is where I’d like to see us go.