Speaker 1: What is customer segmentation? Customer segmentation provides insight into the landscape of the market, revealing customer characteristics that can be used to group customers into segments that have something in common. This process is also known as clustering, and the techniques used to develop these models are called clustering algorithms. For example, let's consider the following case. Each point on the scatter plot represents where a customer lies with respect to their age and income. Notice that there are five distinct segments in the data set. There are also some points with extreme values, which may be interpreted as outliers. Clustering algorithms find these segments in the data and label each record with the cluster or segment that it belongs to. Based on an understanding of the variables that characterize each cluster, you can assign a name or meaning to each of these clusters. For example, the cluster in the lower left corner that contains young individuals with a low income might be labeled students, or the cluster of young individuals with high income may be labeled yuppies. These labels describe or summarize the characteristics of the cluster and can be used to define products and services or offers for each customer segment. Customer segmentation can be applied with two objectives. The first is to segment the customer base into smaller groups. This will help in tailoring services and products offered to each group. And the second is to generate a new index, for instance a segment number or label, to be used in other models or as a predictive variable. For example, when we use the variable segment number, the value equivalent to the student segment will always have young age and low income. Data for customer segmentation can include many attributes. For customer segmentation to be effective, data needs to contain demographic information, such as age, gender, marital status, income, and much more. The data can also contain transactional information, such as the products purchased, the dollar volume purchased, number of items purchased, or time of day they were purchased. Cluster models use these variables to create segments that contain customers with similar attributes. In order to have a successful outcome from segmenting customers, the goal needs to be stated up front. Examples include segmenting data to identify the best group of customers to sell a new product to, creating a successful marketing campaign, or optimizing the sales channel mix. By segmenting customers into different groups, businesses can focus not only on the overall behavior of the customers but their future profitability, allowing companies to better focus the resources on the most profitable customers. ANGOS Data Science Platform offers you segmentation functionality found in Knowledge Seeker, Knowledge Studio, and Knowledge Reader software. To find out how ANGOS software and services can help you with customer segmentation, contact us at info at angos.com.