Predicting Customer Behavior 

Purpose

You can use this business process to uncover patterns in customer behavior such as buying and churn behavior and deploy these insights to make predictions about future customer behavior. You can use the data mining methods Scoring and Decision Tree.

 

Using this business process, you can determine, for example, whether a customer is likely to terminate their contract. You can use this prediction as the basis to decide, for example, whether and how customers are to be divided into target groups or whether and how they are to be addressed in a marketing campaign.

Process Flow

...

        1.      System provides customer master data and customer profiles (SAP CRM)

        2.      System provides the transaction data (SAP R/3)

This data such as customers’ payment behavior is collected in the operational processes.

        3.      System provides the transaction data (SAP CRM)

This data such as activities and sales orders is collected in the operational processes.

        4.      System updates data in customer knowledge base (SAP BW)

By regularly extracting data from SAP CRM and SAP R/3 into SAP BW, the system updates the data in the customer knowledge base in SAP BW. This way, the data can be applied in different analyses and for data mining methods.

        5.      Create and train model for scoring or decision trees (SAP BW)

You create a model for the data mining method Scoring or Decision Tree and train the model using historic data. Both of these data mining methods are used to classify your customers according to known or unknown criteria.

        6.      System determines propensities and scores (SAP BW)
When the model is trained, the system looks for definite patterns in the data. In the case of scoring, for example, you assign to each customer a score that is calculated from specific attributes or behavior attributes and the respective weighting for those attributes. In the case of decision trees, the system determines which customer attributes can be associated with a particular type of customer behavior or propensity.

        7.      Apply model to new data to predict customer behavior (SAP BW)

To predict customer behavior, you apply the model on other data such as data relating to customers or prospects for whom you only have incomplete data.

        8.      Analyze decision trees or scoring results (SAP BW)

In the Analysis Process Designer (APD), you analyze the prediction results.

        9.      Update customer knowledge base (SAP BW)

You update the prediction results in the customer knowledge base.

     10.      Transfer customer propensities and scores (SAP BW)

You transfer the determined customer propensities and scoring results to SAP CRM.

     11.      System updates the customer behavior information (SAP CRM)