Clustering is an unsupervised learning technique that could be used for customer segmentation in the airline industry. In data analysis, clustering is a technique for grouping data points with shared characteristics and attributes into distinct but related clusters.
Clustering can be used in the context of customer segmentation for an airline to classify passengers into subsets according to their preferences in flight routes, the frequency with which they fly, and other personal characteristics. This will help you categorise your clientele into more manageable subsets so that you can better understand their needs and tailor your marketing and support strategies to them.
Common techniques for dividing up clientele into smaller groups include K-means clustering and hierarchical clustering. To use K-means clustering, we divide the data into k groups and then find the smallest possible sum of squared distances between each data point and the centre of its cluster. To create a hierarchy of clusters, hierarchical clustering iteratively merges and splits existing clusters until a stopping criterion is met.
An in-depth literature review on airline customer segmentation would involve a thorough examination of the studies’ methods, findings, and overall conclusions. There is value in looking at other research that has used similar clustering techniques so you can compare and contrast the findings and draw your own conclusions.
K-means clustering was used by Chen et al. (2018) to categorise customers of a low-cost airline according to their flight patterns, ticket-buying preferences, and demographics. According to the data, there are three distinct types of customers: those who prioritise convenience, those who prioritise value, and those who prioritise price.
Clustering for product segmentation is another way in which a product management company can categorise their products according to shared features and buyer preferences. Example: Kim et al. (2020) used hierarchical clustering to classify products sold by a home appliance retailer according to sales and customer ratings. Products were found to fall into three distinct price points, which we labelled as “high-end,” “mid-range,” and “budget.”
These case studies illustrate how cluster analysis can be used to categorise customers and products across industries, yielding useful information for advertising and product development decisions.
