TMF Frameworx & Big Data

Posted by cfaurer on May 12th, 2014 filed in Frameworx Consulting, Frameworx Training

The membership at the TeleManagement Forum kicked off an effort to understand what the relationship is between TMF Frameworx & Big Data. The result of their efforts are represented in Technical Report 202 and summarized in this article.

An overview of the elements that represent the ‘Big Data’ challenge are shown in the figure below. From multiple sources the data is collected and integrated into the platform where formatting is done to get the data into a more uniform format. After Data Ingestion, any of the other modules can used used to further process the collected data.



Data Management provides functionality that can be applied on ingested datasets to transform, correlate, enrich and manipulate the data further. Ensuring data quality and security are also important consideration.

The Data Analysis functionality supports advanced data analytics including calculation of  metrics and data modeling.

Real-time Stream Processing can be viewed as an adjunct to Data Analysis in that it provides support for real-time processing of datasets.

The Big Data Repository provides storage and access to the the ingested ‘raw’ data as well as intermediate and processed forms of the data.

TR202 describes 23 use cases covering several data processing scenarios; offline batch processing, real-time stream processing and real-time/offline hybrid processing. These use cases have been mapped to the TMF Frameworx Business Process framework:



The list of use cases are:

Use Case ID Use Case
SMO1 Personalized Offers while Browsing
SMO2 Personalized Offers during Checkout
SMO3 Real-time Personalized Offers during a Live Interaction
SMO4 Real-time Personalized Offers Based on Customer Location
SMO5 Real-time Personalized Offers Based on Customer Usage
SMO6 Purchase Propensity Analytics for Enhanced Offer Targeting
SMO7 Product Introduction Analytics
SMO8 Product Performance Optimization
SRDM1 Value-based Network Planning
SRDM2 New Enterprise Order Impact Analysis
SRDM3 Policy-based Capacity Management
OCRM1 Personalization of Real-Time Interaction in Assisted Care
OCRM2 Personalization of Real-Time Interaction in Unassisted Care
OCRM3 Analytics of Customer Satisfaction in Care Channels
OCRM4 Right Outbound Care Channel and Time
OCRM5 Churn Risk Prediction for Customer Retention
OCRM6 Churn Motivation Prediction for Customer Retention
OCRM7 Personalized Retention Campaigns for Customer Retention
OCRM8 Retention Campaign Acceptance Propensity Analytics
ORMO1 Network Fault Prediction
ORMO2 Real-time Value-based Congestion Management
ORMO3 Real-Time Customer Offload Management
OSPRM1 Partner Value Optimization
OBRM1 Revenue Assurance
EEEM1 Business Process Optimization
EFAM1 Fraud Management

Here’s a look at one of the use cases – OCRM5: Churn Risk Prediction for Customer Retention

Name: Automatic generation of churn risk scores for customers based upon their behavior
Horizontal: Customer Relationship Management
  • Churn/Retention Manager
Business Drivers: The average communication provider loses 10-20% of its customers annually. Depending on geography and market dynamics, some even face 40% Churn. This has implications on revenues as estimations show acquiring new customers can cost 5 -7 times more than retaining current customers and 2% increase in customer retention has the same effect on profits as cutting costs by 10%.This use case provides the provider with visibility of which customers have the highest churn risk so that campaigns can be run to retain these customers.
PI’s and Business Metrics:
  • Improved awareness of customer state
  • Increased Effect of Campaigns on Customers
Story: The churn/retention manager needs to have clear visibility of the state of each of the provider’s customers in order to ensure that the right customers are targeted with retention offers. Big data analytics are employed to continually monitor customer interactions with the provider and provider services to identify behavior that is indicative of churn. For example, call-drops, complaints, care issues, or overage, along with demographic information that provides baseline probability of the customer churning.The analytics produces a comprehensive report of customer churn risk either on a scheduled basis or on demand. This report constitutes the customer’s unique identifier and associated churn risk score. The churn/retention manager can tune the output of this report by specifying a threshold churn risk score, such that only high risk customers above this score are returned in the report.Using this report the churn/retention manager can target the high risk customers with campaigns to attempt to retain these customers. This can either be performed manually or using the big data analytics use cases described in the following sections.
Required Data Sources:
  • CRM Data
  • Usage & Billing Information
  • Purchase History
  • Payment History
  • Network Quality
  • Call logs
  • Call transcripts from assisted channels
Optional Data Sources:
  • Complaints on social media channels

If you would like to learn more about Big Data and how it can be used by Communication Service Providers (CSP) to support business processes ranging from product offering to billing and how to increase their business value then visit the TeleManagement Forum website and get involved.

Leave a Comment

You must be logged in to post a comment.