Creating a Predictive Churn Model : Part 1

A Predictive Churn Model is a tool that defines the steps and stages of customer churn, or a customer leaving your service or product. Having a predictive churn model gives you awareness and quantifiable metrics to fight against in your retention efforts. Without this tool, you would be acting on broad assumptions, not a data-driven model that reflects how your customers really act. But with an evolving churn model, you can fight for retention by acting on the metrics as they happen. This gives you the ability to pattern habits of customers who leave, and step in before they make that decision.

The first step in creating this model is understanding your customer behavior that comes from customer data points. Without a strong base of customer data, it’s going to be impossible to understand how you can better serve them. You need to be able to evaluate the triggers that caused them to ultimately leave your company. So what data do you need? Let’s look at some popular items that might help you think of your own company and what data you have, or could get easily.

Purchase History

[WAIT] Are you doing everything you should be doing for online marketing?. See  The Plan ››

  • Frequency of Purchase
  • Date of Last Purchase
  • Value of Purchases
  • Payment Method
  • Balances / Store Credit
Customer Information
  • Zip Code
  • Income
  • Gender
  • Occupation
  • Do they get your Mailings?

Customer Interactions

  • Service Questions
  • Store Visits / Online
  • Complaint Resolutions
  • Complaint Priority
Products
  • Type of Product
  • Variety of Products
  • Coupon Usage

The more data you have in your evaluation, the more accurate your model can be. It’s important to know as much as you can about the events that lead to your customer leaving. But once you have customer data in a format that you can easily manipulate and query, you can start to look for trends in you data that equate to customer churn. Identify all the churned customers first, then look at that group to establish any patterns. In the second part of this article we will look at how to use that data to start asking the right questions to see the churn happening, and in the last part of this series we will look at how we take that information and flip it around to put our predictive churn model into action.

Sign up for our monthly newsletter to get notification of all New North Insights.

Read Part 2 of this series >>

 

 

New Call-to-action