Segmentation Applications in Banking - Data Mining Techniques in CRM: Inside Customer Segmentation - page 282

Table 6.20 Clusters and main customer attributes.

Other patterns

Clusters Age Assets ratio Bucket max Payroll Alternative

(as % of total in previous flag (%) channel

balances) year usage (%)

Cluster 1 55 88 0 11 8

Cluster 2 39 75 0 77 28

Cluster 3 45 10 3 12 51

Cluster 4 45 48 1 11 49

Total 44 53 1 36 38

account for their everyday expenses. Members of cluster 3 are the biggest credit
card spenders. their average credit card spending (purchases and cash advances)
being almost double the overall average of all ''Mass'' customers. Finally, cluster 4
customers make many and large credit transactions. They have the highest average
number and the second highest average amount of credit transactions.
The last profiling table, Table 6.20, examines the age distribution of the
clusters and their differentiation on some main behavioral KPIs.
Cluster 1 contains the most mature part of the customer base with an average
age of 55 years. The assets orientation of this cluster is underlined by one more
indicator that expresses the assets as a percentage of the total balances. On
average, assets account for about 88% of the total balances of each customer in this
cluster, the highest ratio observed. Additionally, these customers seem to prefer
the traditional channels for their transactions, namely the teller in the branch
office and ATMs. Cluster 2 on the other hand contains younger customers with an
average age below 40 years. Another distinctive characteristic of this cluster is its
raised percentage of payroll customers. Almost 80% of them have a salary account
at the bank.
Both clusters 3 and 4 are very close to the average age and also make much use
of alternative channels for their transactions, including e-banking, SMS banking,
and web banking. Almost half of their members have used alternative channels
during the time period examined. However, the similarities end there. Cluster 4
has an average assets ratio as opposed to cluster 3, which has the smallest assets
ratio and total balances dominated by liabilities (lending balances). Cluster 3,
mainly composed of borrowers, also has the largest maximum bucket (maximum
number of months in arrears during the last 12 months for all credit cards and
loans), suggesting that some members of this cluster fail to pay their due amount
on time.