Corpus ID: 39119115

Discovering Customer Journey Maps using a Mixture of Markov Models

@inproceedings{Harbich2017DiscoveringCJ,
  title={Discovering Customer Journey Maps using a Mixture of Markov Models},
  author={Matthieu Harbich and Ga{\"e}l Bernard and P. Berkes and B. Garbinato and P. Andritsos},
  booktitle={SIMPDA},
  year={2017}
}

Matthieu Harbich, Gaël Bernard, +2 authors P. Andritsos

Published in SIMPDA 2017

Computer Science

View Paper

ceur-ws.org

Save to Library

Create Alert

Cite

Launch Research Feed

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).