Using artificial intelligence to enhance customer segmentation
The aviation industry is a complex business. Significant time, effort, expertise and money go into the development of new destinations and maintaining the attractiveness of existing destinations. This makes it crucial to fly as many people as possible to every destination to recover costs and ensure profits. With delaware’s help, BrightAir took a cutting-edge approach to meeting the needs of specific target groups.
Know thy customers
BrightAir is active on a global scale. With the majority of their destinations in Europe and the Middle East, they recently also expanded their offering with transatlantic destinations. Already a longstanding delaware client, they approached delaware’s AI team in early 2017 with an interesting business case.
“They wanted to truly understand who their customers were,” explains Sven Arnauts, manager of the Customer Intelligence team at delaware. “It’s a deceptively simple question with a complex answer, as different target groups look for very different things when they travel. Each group experiences their travels differently – and has to be approached in a very specific way.”
After several intake discussions, the delaware team combined numerous sources of data into complete profiles for each customer and performed a customer segmentation exercise to cluster customers into specific profiles, each with their own interests and budgets. “After we finished this exercise, we had to identify when, where and how to best target them to optimize the value for BrightAir,” Sven continues.
Customer behavior reveals underlying preferences
The data revealed an important insight: even within each region, there are huge differences between individual customer segments. “We demonstrated that a general communication strategy per region is still not good enough to reach customers,” says Kevin De Beck, data scientist at delaware. “In order to drive ticket sales, BrightAir would have to focus on the level deeper being the customer’s behavioral characteristics, since those even differed within a region.”
However, this was no ordinary, everyday customer segmentation exercise. “We used a complex artificial intelligence algorithm, which takes into account many different aspects of customer behavior. The term ‘customer segmentation’ is misleading, as we didn’t start from business rules or traditional high-medium-low segmentations, which are based on assumptions and arbitrary choices. Our model is based on customer behavior itself, and is so multilayered that it took ten iterations to tune it.”
“For example, in multiple regions, we saw traditional families traveling with children. In one region, we observed a preference for a smooth and carefree travel experience, with customers booking their travels months in advance via a travel agency. In other regions, that same family group was represented, but behaved in a completely different way. For instance, they might only be interested in the cheapest tickets, or were much more susceptible to last-minute discounts they found online,” Kevin continues. “This completely shifted the way BrightAir viewed marketing personas and campaign management. There just wasn’t one group of ’family travelers’ or one group of ‘city trip travelers’, because within each group, we still found relevant sub-groups behaving completely differently.”
“We didn’t start from business rules or traditional high-medium-low segmentations, which are based on assumptions. Our multilayered artificial intelligence model is based on customer behavior itself.”
As a result, the robust insights generated by the model can be used as a foundation for BrightAir’s strategic roadmap for developing new destinations, not just to optimize its marketing activities.
“Developing new destinations and organizing flights is very cost-intensive,” Sven goes on to say. “Every flight is basically a sunk cost. If a destination isn’t popular, or if it doesn’t appeal to the target group in the right way, an airline risks significant losses. Each flight has a fixed price, excluding the costs of fuel, so the more customers you can schedule for a flight, the higher profit you can bring in.”
An ongoing exercise
Alongside delaware, BrightAir uses its new usage-based segmentation model to optimize communication. “But this isn’t a one-off project,” asserts Kevin. “Customers move between segments, and that’s where BrightAir must take action to build on the relationship. It’s an iterative process.”
Are you ready to discover what your customer data says about your company’s future? It’s not fortune-telling – it’s artificial intelligence. Our experienced team would love to hear from you.
Disclaimer: For reasons of non-disclosure, all identifiable information has been anonymized.