Subscribe to DSC Newsletter



I am trying to build a churn model using survival analysis for our top tier customers.  The thing is that we are not a subscribtion based company, which makes the defintion of churn/defection hard to define.  I can try to define a churn customer as a customer whose annual spent with us drops by 10% or 20%.  However, what is the OPTIMAL way of defining a churn customer?  Is it a business decision that each company has to make or can it be formulated as an optimization problem and be solved using true optimization techniques?  Thanks.

Views: 501

Reply to This

Replies to This Discussion

My 2 cents:

I have faced this problem utilizing survival analysis itself -- at some point in time since last purchase, when does the probability of purchase dip below a business rule (say 10%). I would then deem the customer attrited.

Second, I have tested an intervention (a discount offer) at various times since last purchase and measured where the incremental lift versus a control group was the highest. Then shape of the smoothed curve can help inform about organic demand...

Love to hear others opinions.


On Data Science Central

© 2021   TechTarget, Inc.   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service