SAS – Business Intelligence – Churn and Campaign Management Solution For Telecom Industry

Introduction

In the modern Telecommunication with the competition mounting up between the service providers, customer acquisition and retention is a considerable challenge. For the new entrants, acquiring the new customers is the highest priority, whereas for the incumbents, retaining the revenue earning customers is essential.

The telecom companies can increase profitability by creating a predictive modeling for identifying potential churn candidates and non-revenue earning customers; and can increase revenue and profitability by targeted campaigning and promotional offers which will not only retain these customers but also convert the non-revenue earning customers to profitable revenue earning customers.

This article highlights the necessity of churn and campaign management and the usage of SAS – Telecommunication Intelligence software (TIS) for the purpose. It also includes various implementation challenges for SAS – TIS in the real time scenario.

Churn Management

Customer acquisition and retention is a significant challenge in all industries. In the Telecom industry it affects profitability of the company if a customer churns before the company can earn back the investment it incurred in acquiring the customer. Therefore, it is very critical to identify the profitable customers and retain them.

With the telecom market becoming more competitive, determining the reasons of the customer leaving the service of the company is increasingly difficult. In this circumstance, it is even more difficult to predict the probability of the customer to leave in near future. It is increasingly challenging to devise a cost-effect incentive to target the right customer to convince him to stay with the company.

Predictive modeling of churn analysis and management aims at generating scores depicting the probability of the customers to churn out in future. This takes into consideration different aspects of customer’s susceptibility to churn, including the history of people those who have churned in the past and build a data model that generates an easy-to-understand reference numbers (scores) assigned to each customers. These customers are then targeted with incentives to deter their cancellation. In other words, Churn analysis determines the probable reasons for a future cancellation depending on the past records which will help the companies to customize their offer. For example: if analysis reveals that many customers have churned from a particular area last month and further investigation has identified that there are frequent call drops (disruptions in service) in that exchange (or BTS area). It can be concluded that due to the technical inadequacy of that particular exchange, frequent call drops are experienced which has contributed to the customer dissatisfaction and their moving out of the company. So further technical solution for that exchange can prevent future potential churns.

Business Definition of Churn Management

Defining churn is the first and foremost activity in Churn Management designing. Different companies define churn according to their business experiences.

Churn definition differs from a Pre-paid to Post-paid scenario.

In pre-paid scenario, a customer can be considered as churned in the following cases:

a) If the customer goes out of network (deactivated)

b) If the customer is an active non user (ANU)

A customer can be considered as ANU when:

i. the customer has no outgoing or incoming usage for last (X) rolling days

ii. the customer has only incoming usage but no out-going usage for last (X) rolling days iii. If the customer’s usage is below a pre-determined (business decided) amount for last (X) rolling days.

In post-paid scenario, a customer pays a rental on monthly basis. So in case of non-usage or lower-usage, the company earns fixed revenue from every post-paid customer. Therefore, the customer is considered as churned only when he/she goes out of network (Deactivated).

Churn Parameters for business analysis

After defining churn, next activity is identifying the correct parameters for the contribution of churn. The churn probability or churn scores for individual customers can be generated on the basis of following categorical details:

1. Customer demographics Customer demographics related data are used for segmenting the entire customer base depending on:

a) Age

b) Sex

c) Income

d) Customer Account Information

e) Subscription life cycle

2. Billing and Usage:

Billing and usage related information which is obtained from switch (Call Data Records) is mainly used for detection of churn probability. The following details are used:

a. Price plan

b. Monthly usage summary (Charged call count, Charged data volume, Free call & Data amount)

c. Monthly profit contribution

d. Bounced payment

e. Managing channel information

f. Recharge channel information

g. Network Product information ( Voice, Messaging, Data)

3. Technical Quality:

Quality of service is a potential churn driver as call drops or inferior service quality increases the customer dissatisfaction and therefore churn probability. In case of CDMA, as the customer is tightly coupled with the handset equipment, the aging of handset impacts the probability of the customer churn.

The following details are used:

a. Dropped call counts

b. Service quality

c. Equipment age (Handset age in case of CDMA)

4. Contract Details: At the end of the contract period or grace period, the probability of the customer leaving the connection is high, therefore it has a high impact in determination of churn. The following details are used:

a. Commitment period

b. Count of contract renewal

c. Current contract and end date

5. Event related:

Loyalty scheme or loyalty benefits are key drivers for retention. The Loyalty scheme related data is used for churn scoring.

Identifying the source systems:

After deciding the Churn parameters, next step is to identify the source systems from where the respective data will be extracted.

For example:

Cusomer details from CRM system

Usage & Billing related details from Billing system

Technical Quality from Exchange & CellSite

Activation details from Provisioning system

Data Management

Data management is the foundation for a business analysis. Correct data should be present in correct place.

Data Management has three parts:

Extraction: Involves extracting of data from source system and loading to data interchange layer

Transformation: Involves validation of the extracted data (eg: Validation for unique keys), creation of joining conditions among the tables, cleaning of invalid data etc.

Load: Involves loading the data in the Business Intelligence Data Warehouse

Data Modeling and Churn Score generation

Once the authenticated data is available in the data warehouse, the data modeling is performed. It is an iterative process. The quality of the model is accessed and the model which returns the best business value is considered. This model provides results in the form of churn score of individual customers which can be used for determining campaign targets.

Using the churn scores for Retention Campaigns

The data model generates individual customer’s churn score which ranges from 0 to 1.

0 – Signifies least probability of the customer to churn

1 – Signifies highest probability of the customer to churn.

These scores are weighted components of various parameters, such as

Usage information

Balance information

Recharge information

Decrement (Promotional and Core) information

Handset feature

Network coverage

Quality of service

Customer service/complaints

Price plan sensitivity

Business decision needs to be taken to determine an upper threshold of the churn score. The customers above this threshold need to be analyzed further (eg: customers with score 0.7 and above). The top two parameters contributing to the churn score to be generated on individual customer level (for customers having churn scores greater than the threshold). Depending on these parameters retention campaign can be carried out. The parameters can be as follows:

Usage statistics: The usage behavior can be derived from the combination of decrement (promo and core), balance and recharge information. The customer who has higher score in “lesser usage” can be targeted with promotional price plan offers to enhance his/her usage and convert that customer from non-revenue earning to revenue earning.

Higher Off-net usage: The higher score on “off-net usage” signifies that the particular customer has called very frequently to other networks. A targeted campaign can be performed with the price plan beneficial to call other networks. A further analysis of the called off-net numbers can result in identifying frequently called off-net numbers which can be targeted by campaigns as a candidate of acquisition.

Handset Features: The handset used by the customer can be old and be lacking the modern features. In this case, the probability of the customer to change to a newer handset is high and there is a considerable susceptibility of that customer to move to another service provider having bundled handset offer. A retention campaign can be targeted (to this group of customers having high Handset churn score) with new service offer bundled with handset.

Customer Service/Complaints: The higher score in Customer service/Complaints signifies that the customer has called the customer care frequently and probability of that customer dissatisfied with the service is higher. Further investigation to the customer call interaction details can reveal the cause of frequently calling to customer service. After the execution of campaigns on the basis of the churn score and churn drivers, the campaign response needs to be captured and fed into the database for analysis of successfulness of campaigns.

Implementing Churn Management Solution Implementation Steps

The following phases are involved in Churn Management solution implementation:

1. Requirement Analysis: In this phase, the business requirements are gathered and analyzed and business definitions for churn are decided

2. Solution Assessment: In this phase, the business intelligence solutions are assessed with the high level requirement of the implementing company. The feasibility test is done depending on the high level business requirement and data availability.

3. Detailed Analysis/Detailed design: In this stage, the business requirements for the Churn Management project are analyzed in depth for design, development and enhancement of the project. An exercise is performed to understand the availability/unavailability of information required to fulfill the business requirements and data mapping from source system.

4. Data Analysis – ETL: In this stage, the data is extracted from the source system, transformed (cleaned/modified for missing fields and data quality is analyzed) and then loaded into Data Warehouse of the business intelligence tool.

5. Data Modeling: In this stage, the analytical data models are created by statistical methods (eg: Logistic regression method) on historical data for churn score prediction and Analytical Base tables are populated by data.

6. Reporting: The churn score (0-1: 0 – means less probability of churn, 1 – Maximum probability of churn) is generated at each customer/account/subscription level and corresponding report is generated.

7. User Acceptance Test and Roll-out: On completion of successful UAT, the software is rolled out for the business users.

Implementation Challenges

There are several challenges when a business intelligence solution is implemented in a huge scale of millions of customers.

The major time of the implementation is consumed by data management. Data management utilizes 75% of the total implementation time. Data Management includes:

Identification of source systems from where data needs to be extracted:

Due to the involvement of multiple source systems (CRM, Provisioning system, Billing, Mediation systems etc.), it becomes increasingly difficult to identify the correct source system for various data fields. Identification of the correct data source and mapping to DIL fields consumes majority of the implementation time. If the data source mapping is wrong, then the subsequent steps of implementation (modeling, analysis) will also be erroneous. Therefore, special care needs to be taken during the data gathering exercise.

Data Quality: Data obtained from the source systems need to be of high quality and error free. The major challenge in implementing a business analytics solution is obtaining a high quality data. Cleaning up of data and filling the missing fields consume considerable amount of implementation time.

Change management: With the implementation of a BI solution, the users need to change the way they used to conduct churn prediction and campaign management. Therefore, user adaptability and user awareness needs to be built up through proper training sessions

To make the Business Intelligence system operational: After the implementation, specific organizational structure for handling the BI operations needs to be planned and the resources need to be trained in the required areas.

SAS in business analytics

SAS is a leading business analytics software and service provider in the business intelligence domain. It has delivered proven solutions to access relevant, reliable, consistent information throughout the organizations assisting them to make the right decisions and achieve sustainable performance improvement as well as mitigate risks.

SAS has an extended capability of handling data of large scale (with the help of SAS-SPDS – scalable performance data server). This combined with strong programming language and enriched graphical interface has differentiated it from the other analytical tools available in the market. This makes SAS perfectly suitable for enterprise usage where it demands handling of huge data stores.

SAS – Telecommunication Intelligence Solution (TIS)

SAS has several industy specific solutions. SAS has packaged their business analytics knowledge in the form of models, processes, business logic, queries, reports and analytics.

TIS is the telecom industry specific business analytic solution which has been built specific to telecom industry needs. This solution assists the telecom service providers with specific modules, for example:

SAS Campaign Management for Telecommunication

SAS Customer segmentation for Telecommunication

SAS Customer retention for Telecommunication

SAS Strategic Performance Management for Telecommunication

SAS Cross sell and Up sell for Telecommunication

SAS Payment risk for Telecommunication

SAS churn management and campaign management solution includes Segmenting the entire customer base

Detecting the causes of churn

Scoring the individual customer on the basis of their churn probability

This churn score is further used as an input for campaign management.

SAS Data flow (Architecture)

The data needs to be collected from various source systems.

CRM system: Customer/Account/Subscription related data

Provisioning system: Activation date, equipment (Handset) age Billing System: Billing data

Mediation System: Call record details

The data is collected in the Data Interchange Layer (DIL). The data is then extracted, transformed and loaded into Detailed Data Store (DDS).

The data is used for:

1. Dimensional Data Modeling: This is used for query, reporting and OLAP (Online Analytical Processing)

2. ABT (Analytical Base Table): This is the solution specific model developed which can be used for a particular analysis. For example: The ABT for churn model.

3. Campaign Data Mart: This data is used for targeting specific customer segments for targeted campaign.

Conclusion

Therefore, it is imperative that churn management is an essential challenge in the modern day Indian telecommunication industry. Detecting the proper reason of churn and predicting churn in advance can save the company from substantial revenue loss.

Business Intelligence tools help the telecom service providers to perform data analysis and to predict churn probability of a particular customer. Apart from churn predictive analysis, the tools can be used for various other analysis to assist the business decisions.

SAS has a potential to handle huge volume of data. As a business intelligence tool, SAS empowers the business to efficiently handle enormous volume of data and perform analysis on the available information for millions of customers. Moreover, SAS with its telecommunication specific solution (TIS – Telecom Intelligence Solution) assists in building the data warehouse to hold the required parameters for further analysis.

Therefore, SAS-TIS can be an efficient tool for business intelligence activities in the telecom industry.

Link: SAS company details: http://www.sas.com/

Link: Arindam’s Profile: http://in.linkedin.com/in/arinmukh