Cluster analysis and segmentation represents a purely data driven approach to grouping similar objects, behaviors, or whatever is represented by the data.
Traditional marketers segment customers based on easily identified traits, e.g., age, zip code, gender. Machine learning can take into account all available attributes and information, e.g., website visit history, social network activity, mobile access, purchase history as well as ‘traditional’ attributes. The result is far more insightful, actionable and valuable.
Machine learning clustering algorithms can easily identify consumer behavior that are very challenging for traditional marketing efforts e.g., that a retail customer is going to become a parent – the result can drive focused marketing efforts targeted at the new parent to drive new revenue sources.
Far more obtuse conclusions are possible – e.g., noticing a change in web browsing behavior, combined with reduced purchases might point towards a group of customers who are considering leaving the company. This information could be used to target these customers for customer retention activity (it’s much easier to keep a customer than find a new customer.)