Cluster Analysis

7 Types of
Starbucks
Customers
7 Types of
Starbucks
Customers
7 Types of
Starbucks
Customers
7 Types of
Starbucks
Customers
7 Types of
Starbucks
Customers
7 Types of
Starbucks
Customers
7 Types of
Starbucks
Customers
Segmentation at Sticks Kebob
 Consumer
Median Median Major customer
Location Population spend per
age income profiles
household
Blue Blood Estates;
Brite Lites, Li’l City;
A 29,321 39.1 $92,700 $62,404 Executive Suites;
Upward Bound;
Winner’s Circle
City Startups; Family
Thrifts; Hometown
B 34,183 32.5 $31,900 $36,720 Retired; New
Beginnings; Sunset
City Blues
Brite Lites, Li’l City;
Family Thrifts; Up
C 42,913 32.5 $55,700 $46,828 and-Comers; Upward
Bound; White Picket
Fences
Brite Lites, Li’l City;
Country Squires; Up
D 57,509 34.8 $75,500 $57,880 and-Comers; Upward
Bound; White Picket
Fences
https://claritas360.claritas.com/mybestsegments/#zipLookup
Multi-Attribute Segmentation
• The PRIZM® Methodology: Neighborhood Lifestyle
Segmentation
 “Birds of a feather flock together”
 “People with similar cultural backgrounds, needs, and
perspectives naturally… chose to live in neighborhoods that
offer affordable advantages and compatible lifestyles.”
 “These neighborhoods can be grouped into ‘clusters’ that
exhibit similar demographic and behavioral characteristics.
These neighborhood clusters can then be used to identify and
locate marketing targets.”
• http://tiny.cc/ClaritasPrizm
PRIZM® by Claritas
High T1 5.8mm HHs C1 5.2mm HHs S1 7.8mm HHs U1 5.4mm HHs

Landed Gentry
2nd City Elite Urban
Society Suburbs Uptown
Socioeconomic Status

R1 5.5mm HHs T2 5.5mm HHs

Country
S2 7.3mm HHs
Medium

Families C2 6.9mm HHs The Affluentials U2

Exurban 6.2mm HHs

Blues 2nd City Urban

R2 3.4mm HHs S3 5.8mm HHs
Centers Midscale
Heartlanders Inner
Suburbs
R3 8.8mm HHs T3 6.3mm HHs
Low

Rustic Working C3 6.8mm HHs U3 5.2mm HHs

Living Towns 2nd City Blues Urban Cores

Rural Town 2nd City Metro Suburb Metro Urban

Urbanization
Identifying Segments: Possible Bases for Segmenting Consumers
and Customers

Purchase Behavior
and Product Use

Demographics

Psychographics

Desired
Benefits/Needs
Cluster Analysis

• Deriving Clusters and Assessing Overall Fit

 Decide on Clustering Algorithm
− Many algorithms are available
− More coming all the time
• Main goal: maximize differences between
clusters relative to variation within clusters
K-Means Clustering
10 10
10
9 9
9
8 8
8
7 7
7
6 6
6
5 5
5
4 4
4
Assign 3 Update the 3

cluster
3
2
2 each 2

1
objects to
1
means 1

0 0
0
0 1 2 3 4 5 6 7 8 9 10 most 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10

similar
center Reassign Reassign
10 10

K=2 9 9

8 8

Arbitrarily choose K object 7 7

as initial cluster center 6 6

5 5

4 Update the 4

2
cluster 3

1 means 1

0 0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Criteria for Effective Segmentation

 Identifiable?
 Sustainable?
 Accessible?
 Actionable?
 Responsive?
When Segmentation Fails
• Using same segmentation scheme for different
business objectives
 Different advertising or different products?

• Too much focus on techniques

• Using only psychographic or demographic variables
• Not focusing on differences in customer needs
• Static segmentation schemes
• Lack of senior management buy-in
 Steps to Implement

• Identify Demographic, Psychographic Variables

• Choose the variables to cluster
• Describe behavior using clusters
• Merge with Demographic variables to operationalize