The Mining Data Mart - Data Mining Techniques in CRM: Inside Customer Segmentation - page 142

Table 4.3 The segment and group membership table.

Field name Field type Field description

Year_Month Int The number that indicates the year and month of

data

Customer_ID Int The customer identification number

Banking_Code Int The basic banking code that categorizes customers

as retail, corporate, private banking, and so on

VIP_Flag Bit A flag that indicates if a customer is considered as a VIP

by the bank

Contact_Flag Bit A flag that indicates if a customer agrees to be

contacted by the bank

Person_Flag Bit A flag that indicates if a customer is a person or a legal

entity

Dormant_Flag Bit A flag that indicates, in accordance with the

bank's relevant definition, a dormant customer,
for instance a customer with no transactions for
the last 12 months

Active_Flag Bit A flag that separates active from inactive customers,

according to the bank's definition. For instance, a
customer with null or very low total balances could
be considered as inactive

Payroll_Flag Bit A flag that indicates if a customer has a payroll

account at the bank

Segment_Code Int A code that allocates customers to specific segments

such as affluent (customers with the highest assets,
namely savings and investment balances), mass
(customers with lower assets), staff (belonging to the
organization's staff), and so on

Year_Month and Customer_ID are the primary keys (PKs) of this table.

types of transactions (deposits, withdrawals, etc.). The proposed table of transac-
tions (''M_Transactions''), Table 4.5, captures all information related to customer
transactional behavior. It contains monthly aggregations which summarize the
number and amount of transactions per transaction type and channel, at a cus-
tomer and product account level. The data recorded enable the identification of
transactional patterns and the transactional profiling of the customers. Two lookup
tables, to be presented below, group transaction channels and transaction codes
into broader categories and facilitate further aggregation and summarization of the
tables' records.