Segmentation Applications in Banking - Data Mining Techniques in CRM: Inside Customer Segmentation - page 267

according to business rules that summarize their general behavior. The specific
definitions for assigning segments were determined after an examination of the
characteristics of the entire population.
3. Further segmentation of the emerging core segments into refined sub-segments
of differentiated behavior.
SEGMENTING CUSTOMERS ACCORDING TO THEIR VALUE:
THE VITAL FEW CUSTOMERS
Value-based segmentation is one of the key mining tasks. It differentiates customers
according to their profit contribution and can be used to prioritize customer
handling. High-value customers, after being identified, should always receive
top-level service in every interaction with the bank. Since they account for a
significant share of the organization's profit, the bank has to build their loyalty
by offering benefits and incentives that will make clear that it recognizes their
significance andwill discourage them from turning to competitors. These customers
should be thoroughly monitored over time and any signs indicating a possible
decrease in their relationship with the bank should trigger appropriate retention
activities.
The marketers of the bank decided to partition their customer base according
to marginal revenue, a value index calculated previously and readily available in the
bank's mining data mart. At first, all customers were ordered by rank in respect of
their marginal revenue and subsequently binned into groups of equal size, referred
to as quantiles or tiles. Customers with zero and negative revenue contribution
were distinguished and assigned to two distinct groups. Customers who cause a
loss to the bank formed tile 10 and customers with no profit contribution formed
tile 9. The rest of the customers were grouped into eight value segments of about
10% each.
As shown in Table 6.12, the relative revenue contribution of each derived
segment does not correspond to its size. On the contrary, respective value steeply
decreases from the top to the bottom segments, despite the fact that each segment
has approximately the same number of customers. Thus the top 10% of the most
valuable customers (tile 1) accounts for about 60% of the total bank's revenue.
In a similar manner, the second most valuable segment (tile 2) also accounts for
a disproportionately large part of revenue. The Pareto principle holds true in
the bank's case presented here since, jointly, tiles 1 and 2 provide about 80% of
the overall bank's marginal revenue. These are the vital few customers that the
bank should distinguish and focus on. The bank should try to further develop the
relationship with them and prevent their churn, since their possible attrition will
drastically decrease the bank's profit.