Analysis of Customer Churn 

Purpose

To understand the customer relationships in your retail company, it is important to understand the reasons why customers stop bringing their custom to your company. If you know the reasons why customer relationships have ended, you can predict whether there is a risk of your customers taking their business elsewhere and, if necessary, – that is, if it is important to you to continue the relationship – take measures to retain customers who are likely to move away (“Customer Retention Management”). To do this, start by taking a period of time in the immediate past and analyze which customers have stopped visiting your business and their reasons for doing so. Secondly, determine which of your present customers these reasons also apply to, that is, determine which customers are likely to churn. For this step you can also use statistical and data mining procedures. The period of time for which you are analyzing customer churn is the analysis period and the period of time for which you are predicting is called the prediction period.

This method requires you to know which customers have stopped returning to your company during the analysis period. Since this is not immediately evident in the retail industry (for example, by the termination of a contract) you are forced to determine which customers you are going to regard as churned on the basis of (transaction) data from contact with the customer.

To determine which customers have taken their custom elsewhere during the analysis period, we need two key figures:

·        The first key figure measures customer activity during the analysis period (for example, by sales during this period)

·        The second key figure measures whether the customer was active before (for example, through sales over a period of time prior to the analysis period).

Include both these key figures in a query in SAP BW as the basis for an ABC Classification. In this ABC analysis class the customers for whom the first key figure is 0 as churned, and the others as active. You then save this information as either a customer attribute or as movement data.

Using the information from a customer-only information system you can now, for example, determine the customers who have stopped returning to your business during the analysis period, but who in the period before were classed as frequent customers. You can then try to win these customers back using suitable measures.

To establish the reasons for customer churn, you generally need detailed information regarding your customers’ activities. For more information on these analysis methods, see Analysis Processes in a Customer/Article Information System.

If you use an SAP CRM System, you can save the results of a customer churn analysis to SAP CRM. You do this by defining a marketing attribute for the business partner for the churn indicator in SAP CRM and describing it from SAP BW. You can then have the results of your analyses flow into marketing campaigns when you are defining target groups. You can also use a modified version of this procedure to ensure that customers who have been inactive for a long period of time are excluded from marketing campaigns.

When determining the Customer Lifetime Value (CLTV) a criterion is also defined for the point when a customer is classed as lost. If you use CLTV you should check whether the different definitions of customer churn are consistent.

Process Flow

...

       1.      Define the customer churn index in SAP BW using key figures and create a query in which you calculate the churn index for each customer.

       2.      You define a new ABC classification “Churn Analysis” in which the query you created in step 1 delivers the key figures.

       3.      You make the result of the churn analysis available again in SAP BW. You can now add a churn analysis.