I was trying to build a predictive churn model based on statistical analytical techniques. However, I have not been able to hit upon the right tool (logistic regression, cluster analysis, or PCA) or the right methodology. It would be very nice if you could provide some insights as to the modeling aspects.
I've likened predictive churn modeling to the alchemy practiced during medieval times when trying to find a formula which would create gold. Most predictive models are built on a combination of multivariate techniques which include regression and branching (CHAID) methodologies. For reference, I've attached one of my articles on this subject, which appeared last year on SearchCRM.
Dig Deeper on Customer loyalty and retention
Related Q&A from Michael Lowenstein
Find out if luxury retail stores should offer loyalty programs for their customers in this tip. Learn how some luxury retail brands build loyal ... Continue Reading
Get an experts take on raising the hurdle rates of a customer loyalty program and learn more about value-added benefits from Michael Lowenstein. Continue Reading
Learn a few ways to measure customer loyalty and find out about customer spend or "share of wallet," which expert Michael Lowenstein believes to be ... Continue Reading