The Mining Data Mart - Data Mining Techniques in CRM: Inside Customer Segmentation - page 138

The main information tables are supplemented by lookup tables (also referred
to as reference tables) which group products and transactions into a manageable
number of categories and facilitate further summarization of the information.
The presented data mart refers to retail customers and mainly focuses on
information from internal data sources, such as core banking platforms, other trans-
actional systems, call center platforms, and other operational systems. Although
most of the contained data are supposed to be extracted from the data warehouse,
the data mart should be designed and constructed as a supplement to and not as
a replacement for the data warehouse. Therefore, all data stored in the data mart
should also reside in the data warehouse to ensure the consistency and integrity of
the information.
CURRENT INFORMATION
Current data mart tables contain information on current and past customers and
only include the most recent update of the relevant information. The records in
this type of table are updated to reflect the latest information and are not deleted
when a relationship terminates. For example, a current table with demographic
information contains the latest update on such data for all customers, even those
who have terminated their relationship with the bank (ex-customers). This type of
table might also be used to track the status of products (openings/closings).
Customer Information
The first proposed table (''C_Customer''), Table 4.1, builds in the basic customer
information, including demographic and contact details data. This table is at
a customer level. Thus, each row is uniquely identified by the corresponding
customer IDfieldwhich constitutes the primary key (PK) of the table. All customers
that ever had a relationship with the bank are included and ex-customers are not
deleted. The table contains the most recent updates of the customer information
as well as the date of each record update.
Product Status
Table 4.2 records detailed information about the status of the products (product
accounts) that each customer currently has or used to have in the past. Therefore
it should contain all critical dates including opening, expiration, and closing dates.
The first product opening can be used to calculate the tenure of the customer. If
all product accounts of a customer are closed, a full churned customer is indicated.
The contained information can be used to calculate the tenure of each product and