Clustering is one of my favourite analytic methods: it resonates well with clients as I’ve found from my consulting experience, and is a relatively straightforward concept to explain non technical audiences.
Earlier this year I’ve used the popular K-Means clustering algorithm to segment customers based on their response to a series of marketing campaigns. For that post I’ve deliberately used a basic dataset to show that it is not only a relatively easy analysis to carry out but can also help unearthing interesting patterns of behaviour in your customer base even when using few customer attributes. ...