Boundless customer segmentation with machine learning.
Sending relevant messages to customers using customer segmentation is not new to today’s marketer. Archetypes or personas are often used to create customer segments. Customer data is enriched with demographic characteristics and insights such as urbanity, type of household and level of prosperity are looked into.
But how does this work if your customers live in different European countries. Countries where such characteristics are unknown, you have no credentials and you lack local insights?
During a two-day AI Hackathon, organized by the DDMA, our data team developed a machine learning model focused on a new way of segmenting. This model creates clusters by recognizing patterns in the data that people can easily overlook. Each cluster is unique but also complex, the above animation simplifies each cluster with a color.
After an in-depth analysis of the characteristics of each cluster, labels are assigned and the cluster is given a face and a consumer profile is created. Consider, for example, the “Urban professional”.
By visualizing data and displaying the profile on a map, we make the penetration rate transparent and this helps to prioritize the marketing and sales activities and the internal conviction of the chosen strategy.
Machine learning modeling is not only valuable for finding detailed mutual patterns, it helps to exclude human views and prejudices for determining consumer profiles.
Want to know more about how machine learning can help segment your existing customers and find new target groups? Then contact us now!
