Discovering Customer Journey Maps using a Mixture of Markov Models

Customer Journey Maps (CJMs) summarize the behavior of customers by displaying the most common sequences of steps they take when engaging with a company or product. In many practical applications, the challenge lies in automatically discovering these prototypical sequences from raw event logs for thousands of customers. We propose a novel, probabilistic approach based on a mixture of Markov models and show it can reliably extract CJMs with just one input parameter (and potentially none).