Overview

Skill Level: Any

Prerequisites

This tutorial is intended for users of both WCA and WMI. WMI Standard Edition is required.

Step-by-step

  1. Add columns to your WCA data

    Include the following source columns in the WCA application.

    Column Description
    CHURN_IND

    Indicates customer churn or attrition. This must be a numeric value in WCA. Valid values are:

    0 – they are still a customer

    1 – They are no longer a customer

    TOTAL_PURCHASE_AMT Numeric value indicating the total value of all items purchased to date.

     

     

  2. Add the new columns to your WMI import

    Use the configuration API to identify the new columns to import. CHURN_IND and TOTAL_PURCHASE_AMT are required targets of the models.

    Attrition also requires two or more more valid predictors, which can be any string or numeric colums.

    For example:

     curl -X PATCH -–header 'Content-Type: application/json' 
    -–header 'Accept: application/json'
    --header 'Authorization: APIKey yourapikeyvalue'
    -d '{ \"value":["CHURN_IND","TOTAL_PURCHASE_AMT,predictor1,predictor2,..."]\}'
    'https://wmi-rest-api-r1.watsonmarketinginsights.ibm.com/api/v1/config/wcaInputColumns'

Expected outcome

This procedure demonstrates how to add additional behavior models to WMI by adding columns to your WCA database. When using this method, all models are run including the engagement models.

You can also enable additional behavior models by importing data in a delimited file. Refer to Adding attrition, CLTV, and pulse analysis models using a delimited file.

Join The Discussion

Your email address will not be published. Required fields are marked *