Documente Academic
Documente Profesional
Documente Cultură
Services
Professor Rajiv Khosla
La Trobe Business School
La Trobe University
Bundoora, Melbourne, Vic. 3088
Australia
email: r.khosla@latrobe.edu.au
Our ability to collect data today far outstrips our ability to explore, analyse,
and understand it.
Improvement in financial services in regional Australia has been followed by
more rhetoric than any objective study of the needs of the people in the
region.
Effectively understand community needs by factoring in customer
demographics and customer focus .
Assist Org X in meeting their social objective in terms of developing a
better understanding the banking and finance needs of the their customers ,
providing training in modelling customer data using intelligent data mining
techniques to local personnel (e.g., Org X personnel),
Facilitate reduction in Org Xs CRM (Customer Relationship Management)
costs by better packaging of their financial products and more focussed
targeting of customers .It will help them plan their branch network and
internet services more effectively to meet the needs of the community in the
region.
Project Resources
Employed (Paid)
Project Plan
Project Plan
Channel Analysis
Frequency-Recency Clustering
S +L
S + I (Term)
4956
226(5%)
4064
1716(35% 659(13%)
)
130 (3%) 732(18%) 762(19%)
1798
53(3%)
250(14%) 328(18%)
16
548
11(2%)
64(12%)
32
225
4(2%)
20(9%)
42(19%)
64
120
1(1%)
6(5%)
17 (14)
128
27
2(7%)
4(15%)
5(19%)
102(19%)
Low Balance (< 60,000), age distribution skewed towards 18-34 years,
low investment opportunity
Balance Value (up to max of 120,000), value peaks in the range of
40,000m - 100, 000 normal age distribution (mean ~ 42 yrs)
Number of
Customer
400
300
200
100
88
80
72
64
56
48
40
32
24
16
-26
0
By Age
93
84
75
66
57
48
39
30
21
11
250
200
150
100
50
0
-26
Number of
Customer
By Age
Number of
Custom er
200
150
100
50
82
74
66
58
50
42
34
26
18
10
0
By Age
Observations
Cluster 18
Med (100-400) face-to-face and Med (50- 80) Internet Banking, Low (<
50) in other channels
Cluster 13
High (700-1500) face-to-face and High (60-160) Internet Banking, Low
(< 50) in other channels
Cluster 4
Cluster 12
Cluster 8
Observations
Cluster 10
Cluster 6
10
11
Num_of_Clients
I-I
1816
2.8%
867
I-L
56
0.07%
1128
I-S
1648
2.6%
807
L-I
24
0.04%
1240
L-L
257
0.4%
822
L-S
652
1.01%
366
S-I
5038
7.8%
1820
S-L
2206
3.4%
1555
S-S
12385
19.2%
1434
12
16500
16000
15500
15000
14500
14000
13500
13000
12500
NOV 00
DEC 00
JAN 01
FEB 01
MAR 01
APR 01
MAY 01
JUN 01
By Month
18000
17500
17000
16500
16000
15500
15000
14500
14000
JUL 01
AUG 01
SEP 01
OCT 01
NOV 01
DEC 01
By Month
13
19%
8%
No ATM/EFTPOS
No telephone
40%
2%
Miscellaneous
Summary
Association Analysis
14
Summary
Loan Account analysis
Summary
Internet Banking Analysis
Channel analysis
Future Work
Customer Retention
Customer Share
Customer Demographics Analysis
Loan Default Behaviour
Term Deposit Analysis
Cross Selling and Product Packaging
Internet Banking Analysis
Channel Usage Trends
15
Future Work
SERVER
Financial Services
Manager
Customer Loan
Payment Behaviour
Rival Bank
Information
Sales/ Relationship
Manager
Product Packaging
Rival Bank
Information
Fraud Section
Savings
Account
Database
Product
Packaging
(Cross Selling)
Customer Loan
Payment
Behaviour
Loan
Database
News
Wires
Customer
Frequency &
Recency
Rival Bank
Information
Fraudulent
Transactions
Customer
Credit
Card
Database
Via the
Internet
FIRE
WALL
Fraudulent Transactions
Frequency & Recency
Customer
Demograp
hics
Database
On Line(Real-Time) Customer
Recommendation Tool
Predictive
Models (e.g.,
cross selling)
Customer
Demographics
Transactional
Data
Real Time
Processing
Org X
Representative
Business Rules
Customer
Database
Y/N
1:1
Offer
16