I’m working on a project involving the change in customer attrition (a binary variable that can be 0 if the person stays or 1 if they leave) over the last two years. I’d like to find a way to determine which customer segments are driving the change. For example if the attrition rate (% 1s out of the total population) has increased 5 points overall in two years, for which segments of customers has it increased more or less?
What type of statistical method could I use to model the change in attrition over time?
I’m not necessarily interested here in modeling attrition with a logistic regression, as that would be a snapshot in time rather than a look at what are the significant factors in influencing a trend in attrition over time.