Imagine that a company’s customer base is a cake. And a segmentation tool is the knife. Give each department the knife and they’ll cut the cake up in different ways: Sales by demographic, Loyalty by previous purchase, Marketing by channel, etc. With everyone cutting in different ways, the cake will end up a pile of crumbs. In the real world, this is driving an inconsistent customer experience across the customer journey through the different departments.
A best practice has emerged that looks at segmentation in a completely different way. It puts the customer at the centre of everything. To be more precise, the needs of the customer take centre-stage.
Many industries are developing ‘strategic’ segments according to customer behaviour, with the underlying logical and emotional needs driving the behaviour of the customer being key considerations. For airlines, this could mean dividing customers into common groups deriving from the customers’ perspective, namely purpose of travel.
AirBaltic worked with us to examine their segmentation using unsupervised clustering techniques to define how to ‘cut the cake’. Unsupervised clustering refers to the identification of naturally occurring clusters through machine learning techniques.
To start with, Amadeus R&D teams looked at AirBaltic customer profile data to identify patterns and useful variables. We can access this knowledge thanks to the Amadeus Customer Experience Management solution, which aggregates various data sources into unique customer profiles. Combined with the unsupervised algorithm, this wealth of data helped identify two specific clusters: business and leisure.
However, this wasn’t going to take us far enough. During a strategic segmentation workshop, the teams were able to define 10 different segments and the different variables that characterise them. By adding more variables such as advanced purchase or number of trips per year, Amadeus’ data science team fine-tuned the algorithm.
As a result, AirBaltic and Amadeus were able to identify eight different clusters, for example, business commuters or city breakers. We were then able to better understand the different traveller types within each cluster and to discover valuable information. For instance, we found out that city breakers tended to book at the last minute.
This segmentation method is an iterative process including the customer, a data scientist and a consulting approach. Other variables like booking channel or ancillary purchase would help us to better cluster travellers in the next iterations.
This understanding of the traveller and his or her motivations will help each airline department ‘cut the cake’ in a uniform manner. Segmentation is part of a sequenced process. The next steps include persona building and tactical segmentation alignment, where it’s recommended to add a human face and an emotional layer that will help turn these traveller types into real people. To learn more, visit the Amadeus Customer Experience Management for airlines website.