Download Cluster Analysis

MKT 700
Business Intelligence and
Decision Models
Week 6:
Segmentation and Cluster Analysis
What have we seen so far?
Overview:Analytical CRM
Overview:CRISP-DM Methodology
Data Preparation
Legacy Approach: RFM
Customer Value
Where are we going from
now?
Classification: Clustering
Classification +: Profiling
Predictive Modeling: Response Probability
Outline for Today

Clustering:
Clustering and Segmentation
 B2C and B2B
 Clustering theory


Lab
Typical Classroom
Segmentation (Beer)
Local
Office
Worker
Maple
Leaf
Fans
Blue
Collar
Light
Strong
Occasional
Selective
Foodies
Import
Clusters and Segments

Differences between clusters and
segments
Learning segmentation
 Dynamic segmentation

Consumer Segmentation
Taxonomy





Product usage/loyalty
Buying behaviour
Preferred channel
Family life cycle (stage in life)
Lifestyle (personal values)
Status Levels and Segments
(needs + treatment)
Status Levels
Marketing Segments
Business
Customers
Gold
Affluent
Retired
Silver
Young
Singles
Families
With Kids
Customer
Bronze
Bargain
Shoppers
Occasional
Buyers
Marketing
Staff
Data Sources for
Segmentation

Internal



Transactions
Surveys & Customer Service
External (Data overlays)




Lists
Census
Taxfiler
Geocoding
Geo-Segmentation in CDA
Birds of a feather f___k together…

Environics (Prizm)


Generation5 (Mosaic)


http://www.generation5.ca
Manifold:


http://www.environicsanalytics.ca/prizm-c2-clusterlookup
http://www.manifolddatamining.com/html/lifestyle/lifes
tyle171.htm
Pitney-Bowes (Mapinfo)

http://www.utahbluemedia.com/pbbi/psyte/psyteCanad
a.html
B2B Segmentation
Taxonomy
Firm size (employees, sales)
 Industry (SIC, NAICS)
 Buying process
 Value within finished product
 Usage (Production/Maintenance)
 Order size and Frequency
 Expectations

Clustering

Measuring distances (differences) or
proximities (similarities) between
subjects
BI Modeling Techniques

No Target (No dependent variable,
unsupervised learning)
• RFM
• Cluster Analysis

Target (Dependent variable,
supervised learning)
• Regression Analysis
• Decision Trees
• Neural Net Analysis
Measuring distances
(two dimensions, x and y)
A
B
C
Pythagoras
17
Measuring distances
(two dimensions)
D(b,a)
A
B
D(a,c)
D(b,c)
C
18
Measuring distances
(two dimensions)
dac2 = (dx2 + dy2)
A
B
C
Euclid
dac2 = (di)2
dac = [(di)2]1/2
19
Distances between US cities
ATL
CHI
DEN
HOU
LA
MIA
NY
SF
SEA
DC
0
587
1212
701
1936
604
748
2139
2182
543
Chicago
587
0
920
940
1745
1188
713
1858
1737
597
Denver
1212
920
0
879
831
1726
1631
949
1021
1494
701
940
879
0
1374
968
1420
1645
1891
1220
1936
1745
831
1374
0
2339
2451
347
959
2300
Miami
604
1188
1726
968
2339
0
1092
2594
2734
923
New_York
748
713
1631
1420
2451
1092
0
2571
2408
205
2139
2182
543
1858
1737
597
949
1021
1494
1645
1891
1220
347
959
2300
2594
2734
923
2571
2408
205
0
678
2442
678
0
2329
2442
2329
0
Atlanta
Houston
Los_Angeles
San_Francisco
Seattle
Washington_DC
Cluster Analysis Techniques

Hierarchical Clustering

Metric, small datasets
SPSS Hierarchical Clusters
Dendogram
SPSS Multidimensional Scaling
(Euclidean Distance)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Atlanta
Chicago
Denver
Houston
Los_Angeles
Miami
New_York
San_Francisco
Seattle
Washington
1
2
.9575
.5090
-.6416
.2151
-1.6036
1.5101
1.4284
-1.8925
-1.7875
1.3051
-.1905
.4541
.0337
-.7631
-.5197
-.7752
.6914
-.1500
.7723
.4469
Euclidean distance mapping
Cluster Analysis Techniques

Hierarchical Clustering


K-mean Clustering


Metric variables, small datasets
Metric, large datasets
Two-Step Clustering

Metric/non-metric, large datasets,
optimal clustering
Cluster Analysis Techniques
See Chapter 23, SPSS Base Statistics for description of methods
Two-Step Cluster Tutorials

SPSS, Direct Marketing, Chapter 3 and 9
 Help  Case Studies  Direct Marketing 
Cluster Analysis
 File to be used: dmdata.sav

SPSS, Base Statistics, Chapter 24
 Analyze  Classifiy  Two-Step Cluster
 File to be used: Car_Sales.sav
 Help: “Show me”