Yearn, Learn, and Turn to Cut Churn

Spending money to acquire customers,  only to lose them,  is both frustrating and expensive.   Wouldn't it be great if you could automatically predict which of your customers was at risk of leaving and automatically take steps to retain them?  Yeah, that would be great.  But, how would you go about it? 

  • Who do I want to retain?
  • What data do I need?
  • How do I process the data to score customers for attrition risk?
  • What do I do with the scores?
  • How do I contact customers?
  • What do I say and how much am I willing to spend to retain those customers? 

Right  -- here's what we've learned about designing retention programs:  

1. Who do I want to retain?

You probably want to retain profitable customers who you did not charge-off.

2. What data do I need?

A simple MCIF (Marketing Customer Information File) will do.  These files contain sufficiently detailed information to enable good prediction of churn; once you've built your churn file.   Building a churn file is a topic for another blog.  

3. How do I process the data to score customers for attrition risk?

You'll need to run the churn file through a classifier that will assess the risk of each customer to attrit. This process is tricky so - make sure you've got a good data scientist at the helm.    

4. What do I do with the scores?

You'll need to assign thresholds to the scores and map score brackets to specific actions -- which are pushed to the CRM and marketing systems.  

5. How do I contact customers?

You'll need to contact customers through all channels: banking centers, call centers, email, online and mobile banking in a coordinated manner.  

6. What do I say and how much am I willing to spend to retain those customers?

You'll need to decision each customer based on his/her value to the institution, the cost of replacing that customer,  and also contemplate the probability that the prediction is incorrect into the incentive-to-stay offer; which, ideally -- is paid out over time.  Once the offer is set -- the sales script becomes pretty obvious. 


Retention is one of the highest value metrics an institution can drive because it grows enterprise value by building long-term profitable relationships.  But, it's a tricky process with many steps requiring careful procedure development and execution.   Getting it right, however, immediately adds material revenue to the quarterly results and long-term value to the enterprise so -- it's worth the effort.  

If you'd like to learn more about how we handle retention, get in touch -- we'd love to talk with you.