Page 509 - Handbook of Modern Telecommunications

4-40 CRC Handbook of Modern Telecommunications, Second Edition

BI efforts are taken up in some cases at the enterprise level, and in other cases at the department or
function level. However, when building a BI solution, a common framework is usually adopted. The
typical features of this framework are:

A common data repository with simplified structures to facilitate its consumption, periodic acqui-
sition, and refining of data from many sources, loading this data into the repository, and extraction
of data from these sources with a set of reporting tools.

BI solutions built using this approach have grown exponentially in size and scope, forcing the cre-
ation of more manageable subsets of data to suit the business requirements of different departments.
Even as these smaller subsets of data became de facto sources, seeds for their uncontrolled proliferation
were sown and, today, telecommunications companies have thousands of these smaller subsets of data
that business users depend on for their day-to-day analysis and reporting. With this as the context, let’s
examine some of the challenges telecommunications companies face today with respect to their busi-
ness intelligence infrastructure.

4.3.5.1 I Do Not Get the Full Picture
Telecommunications companies find it extremely difficult to acquire a cross-functional or holistic view
of data. For example, how can a risk manager combine the profitability view of customers with their
risk view, and analyze dependencies between profits and risky behavior, a common business need for
managing the risk of a portfolio?
Cause: Risk data and profitability data reside in silos. The basic definitions, data structures, and gran-
ularity of representation are different. It is a systems integration nightmare to combine the two views
of data.
4.3.5.2 I Do Not Trust the Data
The same question often elicits different responses from two different departments of a telecommunica-
tions company. Which is the data that the senior management should trust and why?
Cause: Answers come from different silos although both silos may have obtained their data from
the same underlying data repository. Different definitions: What is the process to ensure that the silos
are leveraging the same definition or calculations? No traceability: How were the estimates and figures
arrived at? What were the bases? Can these figures be traced back to the source data? What were the
transformations along the way? Invariably, a telecommunications company would draw a blank on all
these questions.
4.3.5.3 I Am Not Empowered
Can a user gather simple business intelligence on, say, the total number of new customers or the cus-
tomer attrition figures as of a specified date? Very often, this may be an involved exercise. Again, to
gather this information, does a user have to be familiar with multiple query languages, tools, reporting
interfaces, and databases? The answer from most business users is usually in the affirmative.
Cause: Users need to know the silo or data source to go to. Most users depend on “techies” to find
them the answers. Most reporting tools require a basic understanding of syntax. Business users, again,
depend on technical resources to delve into databases for relevant, often critical information. There is
no common framework for leveraging definitions, and calculation reporting tools do not provide a com-
mon and consistent language for interaction.

4.3.5.4 I Cannot Close the Loop
To act, managers and analysts need current information proactively. For example, up-to-date informa-
tion on high-end customers is critical to a customer relationship manager in determining customer sat-
isfaction and probability of attrition. However, this kind of information seldom reaches line managers