SAS - Business Intelligence - Churn & Campaign Management Solution For Telecom Industry

Introduction				to revenue earning.
In the modern Telecommunication with the				Higher Off-net usage: The higher score on
competition mounting up between the service				"off-net usage" signifies that the particular
providers, customer acquisition and retention is a				customer has called very frequently to other
considerable challenge. For the new entrants,				networks. A targeted campaign can be
acquiring the new customers is the highest				performed with the price plan beneficial to call
priority, whereas for the incumbents, retaining the				other networks. A further analysis of the called
revenue earning customers is essential.				off-net numbers can result in identifying frequently
The telecom companies can increase profitability				called off-net numbers which can be targeted by
by creating a predictive modeling for identifying				campaigns as a candidate of acquisition.
potential churn candidates and non-revenue				Handset Features: The handset used by the
earning customers; and can increase revenue and				customer can be old and be lacking the modern
profitability by targeted campaigning and				features. In this case, the probability of the
promotional offers which will not only retain these				customer to change to a newer handset is high
customers but also convert the non-revenue				and there is a considerable susceptibility of that
earning customers to profitable revenue earning				customer to move to another service provider
customers.				having bundled handset offer. A retention
This article highlights the necessity of churn and				campaign can be targeted (to this group of
campaign management and the usage of SAS -				customers having high Handset churn score) with
Telecommunication Intelligence software (TIS) for				new service offer bundled with handset.
the purpose. It also includes various				Customer Service/Complaints: The higher score in
implementation challenges for SAS - TIS in the				Customer service/Complaints signifies that the
real time scenario.				customer has called the customer care frequently
Churn Management				and probability of that customer dissatisfied with
Customer acquisition and retention is a significant				the service is higher. Further investigation to the
challenge in all industries. In the Telecom industry it				customer call interaction details can reveal the
affects profitability of the company if a customer				cause of frequently calling to customer service.
churns before the company can earn back the				After the execution of campaigns on the basis of
investment it incurred in acquiring the customer.				the churn score and churn drivers, the campaign
Therefore, it is very critical to identify the				response needs to be captured and fed into the
profitable customers and retain them.				database for analysis of successfulness of
With the telecom market becoming more				campaigns.
competitive, determining the reasons of the				Implementing Churn Management Solution
customer leaving the service of the company is				Implementation Steps
increasingly difficult. In this circumstance, it is even				The following phases are involved in Churn
more difficult to predict the probability of the				Management solution implementation:
customer to leave in near future. It is increasingly				1. Requirement Analysis: In this phase, the
challenging to devise a cost-effect incentive to				business requirements are gathered and analyzed
target the right customer to convince him to stay				and business definitions for churn are decided
with the company.				2. Solution Assessment: In this phase, the business
Predictive modeling of churn analysis and				intelligence solutions are assessed with the high
management aims at generating scores depicting				level requirement of the implementing company.
the probability of the customers to churn out in				The feasibility test is done depending on the high
future. This takes into consideration different				level business requirement and data availability.
aspects of customer's susceptibility to churn,				3. Detailed Analysis/Detailed design: In this stage,
including the history of people those who have				the business requirements for the Churn
churned in the past and build a data model that				Management project are analyzed in depth for
generates an easy-to-understand reference				design, development and enhancement of the
numbers (scores) assigned to each customers.				project. An exercise is performed to understand
These customers are then targeted with				the availability/unavailability of information required
incentives to deter their cancellation. In other				to fulfill the business requirements and data
words, Churn analysis determines the probable				mapping from source system.
reasons for a future cancellation depending on the				4. Data Analysis - ETL: In this stage, the data is
past records which will help the companies to				extracted from the source system, transformed
customize their offer. For example: if analysis				(cleaned/modified for missing fields and data
reveals that many customers have churned from				quality is analyzed) and then loaded into Data
a particular area last month and further				Warehouse of the business intelligence tool.
investigation has identified that there are frequent				5. Data Modeling: In this stage, the analytical data
call drops (disruptions in service) in that exchange				models are created by statistical methods (eg:
(or BTS area). It can be concluded that due to				Logistic regression method) on historical data for
the technical inadequacy of that particular				churn score prediction and Analytical Base tables
exchange, frequent call drops are experienced				are populated by data.
which has contributed to the customer				6. Reporting: The churn score (0-1: 0 - means less
dissatisfaction and their moving out of the				probability of churn, 1 - Maximum probability of
company. So further technical solution for that				churn) is generated at each customer/account
exchange can prevent future potential churns.				subscription level and corresponding report is
Business Definition of Churn Management				generated.
Defining churn is the first and foremost activity in				7. User Acceptance Test and Roll-out: On
Churn Management designing. Different companies				completion of successful UAT, the software is
define churn according to their business				rolled out for the business users.
experiences.				Implementation Challenges
Churn definition differs from a Pre-paid to				There are several challenges when a business
Post-paid scenario.				intelligence solution is implemented in a huge scale
In pre-paid scenario, a customer can be				of millions of customers.
considered as churned in the following cases:a) If				The major time of the implementation is
the customer goes out of network				consumed by data management. Data
(deactivated)b) If the customer is an active non				management utilizes 75% of the total
user (ANU)				implementation time. Data Management includes:
A customer can be considered as ANU when:i. the				Identification of source systems from where data
customer has no outgoing or incoming usage for				needs to be extracted:
last (X) rolling daysii. the customer has only				Due to the involvement of multiple source
incoming usage but no out-going usage for last (X)				systems (CRM, Provisioning system, Billing,
rolling days iii. If the customer's usage is below a				Mediation systems etc.), it becomes increasingly
pre-determined (business decided) amount for last				difficult to identify the correct source system for
(X) rolling days.				various data fields. Identification of the correct
In post-paid scenario, a customer pays a rental on				data source and mapping to DIL fields consumes
monthly basis. So in case of non-usage or				majority of the implementation time. If the data
lower-usage, the company earns fixed revenue				source mapping is wrong, then the subsequent
from every post-paid customer. Therefore, the				steps of implementation (modeling, analysis) will
customer is considered as churned only when he				also be erroneous. Therefore, special care needs
she goes out of network (Deactivated).				to be taken during the data gathering exercise.
Churn Parameters for business analysis				Data Quality: Data obtained from the source
After defining churn, next activity is identifying the				systems need to be of high quality and error free.
correct parameters for the contribution of churn.				The major challenge in implementing a business
The churn probability or churn scores for individual				analytics solution is obtaining a high quality data.
customers can be generated on the basis of				Cleaning up of data and filling the missing fields
following categorical details:				consume considerable amount of implementation
1. Customer demographics Customer				time.
demographics related data are used for				Change management: With the implementation of
segmenting the entire customer base depending				a BI solution, the users need to change the way
on:a) Ageb) Sexc) Incomed) Customer Account				they used to conduct churn prediction and
Informatione) Subscription life cycle				campaign management. Therefore, user
2. Billing and Usage:				adaptability and user awareness needs to be built
Billing and usage related information which is				up through proper training sessions
obtained from switch (Call Data Records) is mainly				To make the Business Intelligence system
used for detection of churn probability. The				operational: After the implementation, specific
following details are used:a. Price planb. Monthly				organizational structure for handling the BI
usage summary (Charged call count, Charged data				operations needs to be planned and the resources
volume, Free call & Data amount)c. Monthly				need to be trained in the required areas.
profit contributiond. Bounced paymente. Managing				SAS in business analytics
channel informationf. Recharge channel				SAS is a leading business analytics software and
informationg. Network Product information (				service provider in the business intelligence domain.
Voice, Messaging, Data)				It has delivered proven solutions to access
3. Technical Quality:				relevant, reliable, consistent information throughout
Quality of service is a potential churn driver as call				the organizations assisting them to make the right
drops or inferior service quality increases the				decisions and achieve sustainable performance
customer dissatisfaction and therefore churn				improvement as well as mitigate risks.
probability. In case of CDMA, as the customer is				SAS has an extended capability of handling data
tightly coupled with the handset equipment, the				of large scale (with the help of SAS-SPDS -
aging of handset impacts the probability of the				scalable performance data server). This combined
customer churn.				with strong programming language and enriched
The following details are used:a. Dropped call				graphical interface has differentiated it from the
countsb. Service qualityc. Equipment age (Handset				other analytical tools available in the market. This
age in case of CDMA)				makes SAS perfectly suitable for enterprise
4. Contract Details: At the end of the contract				usage where it demands handling of huge data
period or grace period, the probability of the				stores.
customer leaving the connection is high, therefore				SAS - Telecommunication Intelligence Solution
it has a high impact in determination of churn. The				(TIS)
following details are used:a. Commitment periodb.				SAS has several industy specific solutions. SAS
Count of contract renewalc. Current contract and				has packaged their business analytics knowledge in
end date				the form of models, processes, business logic,
5. Event related:				queries, reports and analytics.
Loyalty scheme or loyalty benefits are key				TIS is the telecom industry specific business
drivers for retention. The Loyalty scheme related				analytic solution which has been built specific to
data is used for churn scoring.				telecom industry needs. This solution assists the
Identifying the source systems:				telecom service providers with specific modules,
After deciding the Churn parameters, next step is				for example:
to identify the source systems from where the				SAS Campaign Management for
respective data will be extracted.				Telecommunication
For example:				SAS Customer segmentation for
Cusomer details from CRM system				Telecommunication
Usage & Billing related details from Billing				SAS Customer retention for Telecommunication
system				SAS Strategic Performance Management for
Technical Quality from Exchange & CellSite				Telecommunication
Activation details from Provisioning system				SAS Cross sell and Up sell for Telecommunication
Data Management				SAS Payment risk for Telecommunication
Data management is the foundation for a				SAS churn management and campaign
business analysis. Correct data should be present				management solution includes Segmenting the
in correct place.				entire customer base
Data Management has three parts:				Detecting the causes of churn
Extraction: Involves extracting of data from				Scoring the individual customer on the basis of
source system and loading to data interchange				their churn probability
layer				This churn score is further used as an input for
Transformation: Involves validation of the				campaign management.
extracted data (eg: Validation for unique keys),				SAS Data flow (Architecture)
creation of joining conditions among the tables,				The data needs to be collected from various
cleaning of invalid data etc.				source systems.
Load: Involves loading the data in the Business				CRM system: Customer/Account/Subscription
Intelligence Data Warehouse				related data
Data Modeling and Churn Score generation				Provisioning system: Activation date, equipment
Once the authenticated data is available in the				(Handset) age Billing System: Billing data
data warehouse, the data modeling is performed.				Mediation System: Call record details
It is an iterative process. The quality of the model				The data is collected in the Data Interchange
is accessed and the model which returns the best				Layer (DIL). The data is then extracted,
business value is considered. This model provides				transformed and loaded into Detailed Data Store
results in the form of churn score of individual				(DDS).
customers which can be used for determining				The data is used for:
campaign targets.				1. Dimensional Data Modeling: This is used for
Using the churn scores for Retention Campaigns				query, reporting and OLAP (Online Analytical
The data model generates individual customer's				Processing)
churn score which ranges from 0 to 1.				2. ABT (Analytical Base Table): This is the solution
0 - Signifies least probability of the customer to				specific model developed which can be used for a
churn				particular analysis. For example: The ABT for
1 - Signifies highest probability of the customer to				churn model.
churn.				3. Campaign Data Mart: This data is used for
These scores are weighted components of				targeting specific customer segments for
various parameters, such as				targeted campaign.
Usage information				Conclusion
Balance information				Therefore, it is imperative that churn
Recharge information				management is an essential challenge in the
Decrement (Promotional and Core) information				modern day Indian telecommunication industry.
Handset feature				Detecting the proper reason of churn and
Network coverage				predicting churn in advance can save the
Quality of service				company from substantial revenue loss.
Customer service/complaints				Business Intelligence tools help the telecom service
Price plan sensitivity				providers to perform data analysis and to predict
Business decision needs to be taken to determine				churn probability of a particular customer. Apart
an upper threshold of the churn score. The				from churn predictive analysis, the tools can be
customers above this threshold need to be				used for various other analysis to assist the
analyzed further (eg: customers with score 0.7				business decisions.
and above). The top two parameters contributing				SAS has a potential to handle huge volume of
to the churn score to be generated on individual				data. As a business intelligence tool, SAS
customer level (for customers having churn				empowers the business to efficiently handle
scores greater than the threshold). Depending on				enormous volume of data and perform analysis
these parameters retention campaign can be				on the available information for millions of
carried out. The parameters can be as follows:				customers. Moreover, SAS with its
Usage statistics: The usage behavior can be				telecommunication specific solution (TIS - Telecom
derived from the combination of decrement				Intelligence Solution) assists in building the data
(promo and core), balance and recharge				warehouse to hold the required parameters for
information. The customer who has higher score				further analysis.
in "lesser usage" can be targeted with promotional				Therefore, SAS-TIS can be an efficient tool for
price plan offers to enhance his/her usage and				business intelligence activities in the telecom
convert that customer from non-revenue earning				industry.

TeleClean

SAS - Business Intelligence - Churn & Campaign Management Solution For Telecom Industry

Articles