Use Case: Telecommunications Industry Churn Example - Teradata Analytic Apps - Vantage Analyst

Vantage Analyst with Machine Learning Engine User Guide

Product
Teradata Analytic Apps
Vantage Analyst
Release Number
1.1
Published
December 2019
Language
English (United States)
Last Update
2020-08-06
dita:mapPath
ezh1551894635141.ditamap
dita:ditavalPath
wsp1565965728073.ditaval
dita:id
B035-3805
lifecycle
previous
Product Category
Teradata Vantage™

In the telecommunications industry, addressing account closure, or churn, is a massive cost-saving effort. Using Path and Model, analysts in this industry can target ways in which to improve retention by understanding customer behavior.

The initial step involves creating an event table to integrate interactions and transactions involving the customer. By capturing the events you can view the customer’s journey, which may have involved visiting a store, going to the website, calling the support line, upgrading service, and canceling service.

Using Path, you can now click on the events to answer business questions, such as:
  • What paths are my customers taking on the website?
  • What paths are my customers taking before calling the support line?
  • What paths are my customers taking before canceling their service?

You may, for example, uncover that a high percentage of customers who cancel have also had bill disputes. This may be a new pattern that was previously unknown and using Path uncovered.

As a result, you might decide to take steps to retain customers who are visibly on this path and have not canceled their service. Using Path, you can create a new pattern where:
  • Event A = Bill Dispute
  • Event B = Any Event

The results expose that Bill Dispute events leads to Review Contract and then Cancel Contract. You can target that path and save the customer list, which can then be used in a customized marketing offer to retain at-risk customers.

You might want to take these findings further and share the results with the data science team for inclusion in their churn prediction model. The information you could share is that bill disputes have been shown to lead to churn, and it might be valuable to add bill disputes to the model.

Assume the data science team adds bill dispute to the customer profile table used by a model, such as Model. Running the Decision Forest model shows that the bill dispute field is important to the model.

These findings are then used to create a new target table using the top 20% of high-churn at-risk customers. These results are then added to the list of customers for the customized marketing campaign.

Using Path and Model, you can also overlay customer value scores onto the list so that they can expedite reaching out to the high-value customers who are potentially at risk of canceling their service contracts.