Real time behavior is the best predictor of relevance to airline customers

When I travelled for business, my needs were like many other business travelers: I had to fly on weekdays, stay downtown and set my plans only a few days before departing. When flying with my family, we’d plan more in advance, we’d need to rent a car with a car seat, and stay at a less expensive hotel with easy access to the I-5 interstate.

What I needed for each trip was dependent upon why I was going that particular time. Not the last time I went. Nevertheless, my in-the-moment needs were and are predictable with some applied data science. With a prediction like that, airlines can help meet their customers’ immediate needs with relevant offers of rental cars, hotels, or even a combination of options related to the flight itself, like checked bags and the ability to change flights.

I had the pleasure of speaking about prediction and relevance at theHamburg Aviation Conference , along with Michael Farrugia, the CTO ofPlanitas Airline Systems , a company specializing in airline business analytics.

Importance of Merchandising

It’s been wonderful to meet the other participants, too. I’ve still got a bit of my outsider perspective. Before joining the airline industry a few years ago, I learned the importance of merchandising… of showing customers how relevant your products are to their needs. Predicting those needs is a very important part of doing that. Airline industry talk often leads to a view of merchandising as a great Moloch, requiring great sacrifices in hopes of supernatural outcomes.

Not true. There are several techniques airlines can use to be relevant now. Michael and I focused on one of particular power and achievability: the prediction of why someone is travelling as they are making their airline booking. You could call this “segmentation,” but traditionally, in marketing, segmentation refers to characteristics of peoples’ lives: household income, number of children, technology adoption. We used the concept of “cohort,” the grouping of customers by their similar intentions in the moment. It’s like in Ancient Rome: a cohort military unit was made up of several hundred men all trying to do the same thing at the same time (picture the opening scenes from the movie Gladiator) even though they came from diverse backgrounds.

By looking at the details of a customer’s booking, while they are in the process of making it, an airline can use algorithms to predict why they’re traveling, assign the customer to a cohort, and make meaningful recommendations in real-time. That reduces the customer’s time to search out solutions to their problems, makes the customer’s experience better when interacting with the airline, and, of course, the airline earns the resulting increase in revenue. It’s a win-win scenario.