Chapter 9

Cluster Analysis

Learning Objectives
After reading this chapter you should understand:
The basic concepts of cluster analysis.
How basic cluster algorithms work.
How to compute simple clustering results manually.
The different types of clustering procedures.
The SPSS clustering outputs.

Keywords Agglomerative and divisive clustering
Chebychev distance

City-block distance
Clustering variables
Dendrogram
Distance matrix

Euclidean distance
Hierarchical and partitioning methods
Icicle diagram

k-means
Matching coefficients
Profiling clusters
Two-step clustering

Are there any market segments where Web-enabled mobile telephony is taking
off in different ways? To answer this question, Okazaki (2006) applies a two-
step cluster analysis by identifying segments of Internet adopters in Japan. The
findings suggest that there are four clusters exhibiting distinct attitudes towards
Web-enabled mobile telephony adoption. Interestingly, freelance, and highly
educated professionals had the most negative perception of mobile Internet
adoption, whereas clerical office workers had the most positive perception.
Furthermore, housewives and company executives also exhibited a positive
attitude toward mobile Internet usage. Marketing managers can now use these
results to better target specific customer segments via mobile Internet services.

Introduction

Grouping similar customers and products is a fundamental marketing activity. It is

used, prominently, in market segmentation. As companies cannot connect with all
their customers, they have to divide markets into groups of consumers, customers,
or clients (called segments) with similar needs and wants. Firms can then target
each of these segments by positioning themselves in a unique segment (such as
Ferrari in the high-end sports car market). While market researchers often form

E. Mooi and M. Sarstedt, A Concise Guide to Market Research, 237

DOI 10.1007/978-3-642-12541-6_9, # Springer-Verlag Berlin Heidelberg 2011
238 9 Cluster Analysis

market segments based on practical grounds, industry practice and wisdom, cluster
analysis allows segments to be formed that are based on data that are less dependent
on subjectivity.
The segmentation of customers is a standard application of cluster analysis, but it
can also be used in different, sometimes rather exotic, contexts such as evaluating
typical supermarket shopping paths (Larson et al. 2005) or deriving employers
branding strategies (Moroko and Uncles 2009).

Understanding Cluster Analysis

Cluster analysis is a convenient method for identifying homogenous groups of

objects called clusters. Objects (or cases, observations) in a specific cluster share
many characteristics, but are very dissimilar to objects not belonging to that cluster.
Lets try to gain a basic understanding of the cluster analysis procedure by
looking at a simple example. Imagine that you are interested in segmenting your
customer base in order to better target them through, for example, pricing strategies.
The first step is to decide on the characteristics that you will use to segment your
customers. In other words, you have to decide which clustering variables will be
included in the analysis. For example, you may want to segment a market based on
customers price consciousness (x) and brand loyalty (y). These two variables can
be measured on a 7-point scale with higher values denoting a higher degree of price
consciousness and brand loyalty. The values of seven respondents are shown in
Table 9.1 and the scatter plot in Fig. 9.1.
The objective of cluster analysis is to identify groups of objects (in this case,
customers) that are very similar with regard to their price consciousness and
brand loyalty and assign them into clusters. After having decided on the clustering
variables (brand loyalty and price consciousness), we need to decide on the
clustering procedure to form our groups of objects. This step is crucial for the
analysis, as different procedures require different decisions prior to analysis. There
is an abundance of different approaches and little guidance on which one to use in
practice. We are going to discuss the most popular approaches in market research,
as they can be easily computed using SPSS. These approaches are: hierarchical
methods, partitioning methods (more precisely, k-means), and two-step clustering,
which is largely a combination of the first two methods. Each of these procedures
follows a different approach to grouping the most similar objects into a cluster and
to determining each objects cluster membership. In other words, whereas an object
in a certain cluster should be as similar as possible to all the other objects in the

Table 9.1 Data

Customer A B C D E F G
x 3 6 5 3 6 4 1
y 7 7 6 5 5 3 2
Understanding Cluster Analysis 239

A B
7

C
6

D E
Brand loyalty (y)

3 F

G
2

0
0 1 2 3 4 5 6 7
Price consciousness (x)

Fig. 9.1 Scatter plot

same cluster, it should likewise be as distinct as possible from objects in different

clusters.
But how do we measure similarity? Some approaches most notably hierarchi-
cal methods require us to specify how similar or different objects are in
order to identify different clusters. Most software packages calculate a measure
of (dis)similarity by estimating the distance between pairs of objects. Objects with
smaller distances between one another are more similar, whereas objects with larger
distances are more dissimilar.
An important problem in the application of cluster analysis is the decision
regarding how many clusters should be derived from the data. This question is
explored in the next step of the analysis. Sometimes, however, we already know the
number of segments that have to be derived from the data. For example, if we were
asked to ascertain what characteristics distinguish frequent shoppers from infre-
quent ones, we need to find two different clusters. However, we do not usually
know the exact number of clusters and then we face a trade-off. On the one hand,
you want as few clusters as possible to make them easy to understand and action-
able. On the other hand, having many clusters allows you to identify more segments
and more subtle differences between segments. In an extreme case, you can address
each individual separately (called one-to-one marketing) to meet consumers vary-
ing needs in the best possible way. Examples of such a micro-marketing strategy
are Pumas Mongolian Shoe BBQ (www.mongolianshoebbq.puma.com) and Nike
ID (http://nikeid.nike.com), in which customers can fully customize a pair of shoes
in a hands-on, tactile, and interactive shoe-making experience. On the other hand,
the costs associated with such a strategy may be prohibitively high in many
240 9 Cluster Analysis

Decide on the clustering variables

Decide on the clustering procedure

Hierarchical methods Partitioning methods Two-step clustering

Select a measure of Select a measure of

similarity or similarity or
dissimilarity dissimilarity

Choose a clustering
algorithm

Decide on the number of clusters

Validate and interpret the cluster solution

Fig. 9.2 Steps in a cluster analysis

business contexts. Thus, we have to ensure that the segments are large enough
to make the targeted marketing programs profitable. Consequently, we have to
cope with a certain degree of within-cluster heterogeneity, which makes targeted
marketing programs less effective.
In the final step, we need to interpret the solution by defining and labeling the
obtained clusters. This can be done by examining the clustering variables mean
values or by identifying explanatory variables to profile the clusters. Ultimately,
managers should be able to identify customers in each segment on the basis of
easily measurable variables. This final step also requires us to assess the clustering
solutions stability and validity. Figure 9.2 illustrates the steps associated with a
cluster analysis; we will discuss these in more detail in the following sections.

Conducting a Cluster Analysis

Decide on the Clustering Variables

At the beginning of the clustering process, we have to select appropriate variables

for clustering. Even though this choice is of utmost importance, it is rarely treated
as such and, instead, a mixture of intuition and data availability guide most analyses
in marketing practice. However, faulty assumptions may lead to improper market
Conducting a Cluster Analysis 241

segments and, consequently, to deficient marketing strategies. Thus, great care should
be taken when selecting the clustering variables.
There are several types of clustering variables and these can be classified into
general (independent of products, services or circumstances) and specific (related
to both the customer and the product, service and/or particular circumstance),
on the one hand, and observable (i.e., measured directly) and unobservable
(i.e., inferred) on the other. Table 9.2 provides several types and examples of
clustering variables.

Table 9.2 Types and examples of clustering variables

General Specific
Observable (directly Cultural, geographic, demographic, User status, usage frequency,
measurable) socio-economic store and brand loyalty
Unobservable Psychographics, values, personality, Benefits, perceptions, attitudes,
(inferred) lifestyle intentions, preferences
Adapted from Wedel and Kamakura (2000)

The types of variables used for cluster analysis provide different segments and,
thereby, influence segment-targeting strategies. Over the last decades, attention has
shifted from more traditional general clustering variables towards product-specific
unobservable variables. The latter generally provide better guidance for decisions
on marketing instruments effective specification. It is generally acknowledged that
segments identified by means of specific unobservable variables are usually more
homogenous and their consumers respond consistently to marketing actions (see
Wedel and Kamakura 2000). However, consumers in these segments are also
frequently hard to identify from variables that are easily measured, such as demo-
graphics. Conversely, segments determined by means of generally observable
variables usually stand out due to their identifiability but often lack a unique
response structure.1 Consequently, researchers often combine different variables
(e.g., multiple lifestyle characteristics combined with demographic variables),
benefiting from each ones strengths.
In some cases, the choice of clustering variables is apparent from the nature of
the task at hand. For example, a managerial problem regarding corporate communi-
cations will have a fairly well defined set of clustering variables, including con-
tenders such as awareness, attitudes, perceptions, and media habits. However, this
is not always the case and researchers have to choose from a set of candidate
variables.
Whichever clustering variables are chosen, it is important to select those that
provide a clear-cut differentiation between the segments regarding a specific
managerial objective.2 More precisely, criterion validity is of special interest; that
is, the extent to which the independent clustering variables are associated with

1
See Wedel and Kamakura (2000).
2
Tonks (2009) provides a discussion of segment design and the choice of clustering variables in
consumer markets.
242 9 Cluster Analysis

one or more dependent variables not included in the analysis. Given this relation-
ship, there should be significant differences between the dependent variable(s)
across the clusters. These associations may or may not be causal, but it is essential
that the clustering variables distinguish the dependent variable(s) significantly.
Criterion variables usually relate to some aspect of behavior, such as purchase
intention or usage frequency.
Generally, you should avoid using an abundance of clustering variables, as this
increases the odds that the variables are no longer dissimilar. If there is a high
degree of collinearity between the variables, they are not sufficiently unique to
identify distinct market segments. If highly correlated variables are used for cluster
analysis, specific aspects covered by these variables will be overrepresented in the
clustering solution. In this regard, absolute correlations above 0.90 are always
problematic. For example, if we were to add another variable called brand pre-
ference to our analysis, it would virtually cover the same aspect as brand loyalty.
Thus, the concept of being attached to a brand would be overrepresented in the
analysis because the clustering procedure does not differentiate between the clus-
tering variables in a conceptual sense. Researchers frequently handle this issue
by applying cluster analysis to the observations factor scores derived from a
previously carried out factor analysis. However, according to Dolnicar and Grun
(2009), this factor-cluster segmentation approach can lead to several problems:
1. The data are pre-processed and the clusters are identified on the basis of trans-
formed values, not on the original information, which leads to different results.
2. In factor analysis, the factor solution does not explain a certain amount of
variance; thus, information is discarded before segments have been identified
or constructed.
3. Eliminating variables with low loadings on all the extracted factors means that,
potentially, the most important pieces of information for the identification of
niche segments are discarded, making it impossible to ever identify such groups.
4. The interpretations of clusters based on the original variables become question-
able given that the segments have been constructed using factor scores.
Several studies have shown that the factor-cluster segmentation significantly
reduces the success of segment recovery.3 Consequently, you should rather reduce
the number of items in the questionnaires pre-testing phase, retaining a reasonable
number of relevant, non-redundant questions that you believe differentiate the
segments well. However, if you have your doubts about the data structure, factor-
clustering segmentation may still be a better option than discarding items that may
conceptually be necessary.
Furthermore, we should keep the sample size in mind. First and foremost,
this relates to issues of managerial relevance as segments sizes need to be substan-
tial to ensure that targeted marketing programs are profitable. From a statistical
perspective, every additional variable requires an over-proportional increase in

3
See the studies by Arabie and Hubert (1994), Sheppard (1996), or Dolnicar and Gr
un (2009).
Conducting a Cluster Analysis 243

observations to ensure valid results. Unfortunately, there is no generally accepted

rule of thumb regarding minimum sample sizes or the relationship between the
objects and the number of clustering variables used.
In a related methodological context, Formann (1984) recommends a sample size
of at least 2m, where m equals the number of clustering variables. This can only
provide rough guidance; nevertheless, we should pay attention to the relationship
between the objects and clustering variables. It does not, for example, appear
logical to cluster ten objects using ten variables. Keep in mind that no matter how
many variables are used and no matter how small the sample size, cluster analysis
will always render a result!
Ultimately, the choice of clustering variables always depends on contextual
influences such as data availability or resources to acquire additional data. Market-
ing researchers often overlook the fact that the choice of clustering variables is
closely connected to data quality. Only those variables that ensure that high quality
data can be used should be included in the analysis. This is very important if a
segmentation solution has to be managerially useful. Furthermore, data are of high
quality if the questions asked have a strong theoretical basis, are not contaminated
by respondent fatigue or response styles, are recent, and thus reflect the current
market situation (Dolnicar and Lazarevski 2009). Lastly, the requirements of other
managerial functions within the organization often play a major role. Sales and
distribution may as well have a major influence on the design of market segments.
Consequently, we have to be aware that subjectivity and common sense agreement
will (and should) always impact the choice of clustering variables.

Decide on the Clustering Procedure

By choosing a specific clustering procedure, we determine how clusters are to be

formed. This always involves optimizing some kind of criterion, such as minimiz-
ing the within-cluster variance (i.e., the clustering variables overall variance of
objects in a specific cluster), or maximizing the distance between the objects or
clusters. The procedure could also address the question of how to determine the
(dis)similarity between objects in a newly formed cluster and the remaining objects
in the dataset.
There are many different clustering procedures and also many ways of classify-
ing these (e.g., overlapping versus non-overlapping, unimodal versus multimodal,
exhaustive versus non-exhaustive).4 A practical distinction is the differentiation
between hierarchical and partitioning methods (most notably the k-means procedure),
which we are going to discuss in the next sections. We also introduce two-step
clustering, which combines the principles of hierarchical and partitioning methods
and which has recently gained increasing attention from market research practice.

4
See Wedel and Kamakura (2000), Dolnicar (2003), and Kaufman and Rousseeuw (2005) for a
review of clustering techniques.
244 9 Cluster Analysis

Hierarchical Methods

Hierarchical clustering procedures are characterized by the tree-like structure

established in the course of the analysis. Most hierarchical techniques fall into a
category called agglomerative clustering. In this category, clusters are consecu-
tively formed from objects. Initially, this type of procedure starts with each object
representing an individual cluster. These clusters are then sequentially merged
according to their similarity. First, the two most similar clusters (i.e., those with
the smallest distance between them) are merged to form a new cluster at the bottom
of the hierarchy. In the next step, another pair of clusters is merged and linked to a
higher level of the hierarchy, and so on. This allows a hierarchy of clusters to be
established from the bottom up. In Fig. 9.3 (left-hand side), we show how agglome-
rative clustering assigns additional objects to clusters as the cluster size increases.
Step 5

Step 1
A, B, C, D, E
Agglomerative clustering

Step 4

Step 2

Divisive clustering
A, B C, D, E
Step 3

Step 3

A, B C, D E
Step 2

Step 4

A, B C D E
Step 1

Step 5

A B C D E

Fig. 9.3 Agglomerative and divisive clustering

A cluster hierarchy can also be generated top-down. In this divisive clustering,

all objects are initially merged into a single cluster, which is then gradually split up.
Figure 9.3 illustrates this concept (right-hand side). As we can see, in both agglom-
erative and divisive clustering, a cluster on a higher level of the hierarchy always
encompasses all clusters from a lower level. This means that if an object is assigned
to a certain cluster, there is no possibility of reassigning this object to another
cluster. This is an important distinction between these types of clustering and
partitioning methods such as k-means, which we will explore in the next section.
Divisive procedures are quite rarely used in market research. We therefore
concentrate on the agglomerative clustering procedures. There are various types
Conducting a Cluster Analysis 245

of agglomerative procedures. However, before we discuss these, we need to define

how similarities or dissimilarities are measured between pairs of objects.

Select a Measure of Similarity or Dissimilarity

There are various measures to express (dis)similarity between pairs of objects.

A straightforward way to assess two objects proximity is by drawing a straight line
between them. For example, when we look at the scatter plot in Fig. 9.1, we can
easily see that the length of the line connecting observations B and C is much
shorter than the line connecting B and G. This type of distance is also referred to as
Euclidean distance (or straight-line distance) and is the most commonly used type
when it comes to analyzing ratio or interval-scaled data.5 In our example, we have
ordinal data, but market researchers usually treat ordinal data as metric data to
calculate distance metrics by assuming that the scale steps are equidistant (very
much like in factor analysis, which we discussed in Chap. 8). To use a hierarchical
clustering procedure, we need to express these distances mathematically. By taking
the data in Table 9.1 into consideration, we can easily compute the Euclidean
distance between customer B and customer C (generally referred to as d(B,C))
with regard to the two variables x and y by using the following formula:
q
dEuclidean B; C xB  xC 2 yB  yC 2

The Euclidean distance is the square root of the sum of the squared differences in
the variables values. Using the data from Table 9.1, we obtain the following:
q p
dEuclidean B; C 6  52 7  62 2 1:414

This distance corresponds to the length of the line that connects objects B and C.
In this case, we only used two variables but we can easily add more under the root
sign in the formula. However, each additional variable will add a dimension to our
research problem (e.g., with six clustering variables, we have to deal with six
dimensions), making it impossible to represent the solution graphically. Similarly,
we can compute the distance between customer B and G, which yields the following:
q p
dEuclidean B; G 6  12 7  22 50 7:071

Likewise, we can compute the distance between all other pairs of objects. All
these distances are usually expressed by means of a distance matrix. In this distance
matrix, the non-diagonal elements express the distances between pairs of objects

5
Note that researchers also often use the squared Euclidean distance.
246 9 Cluster Analysis

and zeros on the diagonal (the distance from each object to itself is, of course, 0). In
our example, the distance matrix is an 8  8 table with the lines and rows
representing the objects (i.e., customers) under consideration (see Table 9.3). As
the distance between objects B and C (in this case 1.414 units) is the same as
between C and B, the distance matrix is symmetrical. Furthermore, since the
distance between an object and itself is zero, one need only look at either the
lower or upper non-diagonal elements.
Table 9.3 Euclidean distance matrix
Objects A B C D E F G
A 0
B 3 0
C 2.236 1.414 0
D 2 3.606 2.236 0
E 3.606 2 1.414 3 0
F 4.123 4.472 3.162 2.236 2.828 0
G 5.385 7.071 5.657 3.606 5.831 3.162 0

There are also alternative distance measures: The city-block distance uses the
sum of the variables absolute differences. This is often called the Manhattan metric
as it is akin to the walking distance between two points in a city like New Yorks
Manhattan district, where the distance equals the number of blocks in the directions
North-South and East-West. Using the city-block distance to compute the distance
between customers B and C (or C and B) yields the following:

dCityblock B; C jxB  xC j jyB  yC j j6  5j j7  6j 2

The resulting distance matrix is in Table 9.4.

Table 9.4 City-block distance matrix

Objects A B C D E F G
A 0
B 3 0
C 3 2 0
D 2 5 3 0
E 5 2 2 3 0
F 5 6 4 3 4 0
G 7 10 8 5 8 4 0

Lastly, when working with metric (or ordinal) data, researchers frequently use
the Chebychev distance, which is the maximum of the absolute difference in the
clustering variables values. In respect of customers B and C, this result is:

dChebychec B; C maxjxB  xC j; jyB  yC j maxj6  5j; j7  6j 1

Figure 9.4 illustrates the interrelation between these three distance measures
regarding two objects, C and G, from our example.
Conducting a Cluster Analysis 247

C
Brand loyalty (y)

Euclidean distance City-block distance

Chebychev distance

Price consciousness (x)

Fig. 9.4 Distance measures

There are other distance measures such as the Angular, Canberra or Mahalanobis
distance. In many situations, the latter is desirable as it compensates for collinearity
between the clustering variables. However, it is (unfortunately) not menu-accessible
in SPSS.
In many analysis tasks, the variables under consideration are measured on
different scales or levels. This would be the case if we extended our set of clustering
variables by adding another ordinal variable representing the customers income
measured by means of, for example, 15 categories. Since the absolute variation
of the income variable would be much greater than the variation of the remaining
two variables (remember, that x and y are measured on 7-point scales), this would
clearly distort our analysis results. We can resolve this problem by standardizing
the data prior to the analysis.
Different standardization methods are available, such as the simple z standardi-
zation, which rescales each variable to have a mean of 0 and a standard deviation of
1 (see Chap. 5). In most situations, however, standardization by range (e.g., to a
range of 0 to 1 or 1 to 1) performs better.6 We recommend standardizing the data
in general, even though this procedure can reduce or inflate the variables influence
on the clustering solution.

6
See Milligan and Cooper (1988).
248 9 Cluster Analysis

Another way of (implicitly) standardizing the data is by using the correlation

between the objects instead of distance measures. For example, suppose a respon-
dent rated price consciousness 2 and brand loyalty 3. Now suppose a second
respondent indicated 5 and 6, whereas a third rated these variables 3 and 3. Eucli-
dean, city-block, and Chebychev distances would indicate that the first respondent is
more similar to the third than to the second. Nevertheless, one could convincingly
argue that the first respondents ratings are more similar to the seconds, as both rate
brand loyalty higher than price consciousness. This can be accounted for by com-
puting the correlation between two vectors of values as a measure of similarity (i.e.,
high correlation coefficients indicate a high degree of similarity). Consequently,
similarity is no longer defined by means of the difference between the answer
categories but by means of the similarity of the answering profiles. Using correlation
is also a way of standardizing the data implicitly.
Whether you use correlation or one of the distance measures depends on whether
you think the relative magnitude of the variables within an object (which favors
correlation) matters more than the relative magnitude of each variable across
objects (which favors distance). However, it is generally recommended that one
uses correlations when applying clustering procedures that are susceptible to out-
liers, such as complete linkage, average linkage or centroid (see next section).
Whereas the distance measures presented thus far can be used for metrically and
in general ordinally scaled data, applying them to nominal or binary data is
meaningless. In this type of analysis, you should rather select a similarity measure
expressing the degree to which variables values share the same category. These so-
called matching coefficients can take different forms but rely on the same allocation
scheme shown in Table 9.5.

Table 9.5 Allocation scheme for matching coefficients

Object 1
Number of variables Number of variables
with category 1 with category 2
Object 2 Number of variables a b
with category 1
Number of variables c d
with category 2

Based on the allocation scheme in Table 9.5, we can compute different matching
coefficients, such as the simple matching coefficient (SM):

ad
SM
abcd

This coefficient is useful when both positive and negative values carry an equal
degree of information. For example, gender is a symmetrical attribute because the
number of males and females provides an equal degree of information.
Conducting a Cluster Analysis 249

Lets take a look at an example by assuming that we have a dataset with three
binary variables: gender (male 1, female 2), customer (customer 1, non-
customer 2), and disposable income (low 1, high 2). The first object is a male
non-customer with a high disposable income, whereas the second object is a female
non-customer with a high disposable income. According to the scheme in Table 9.4,
a b 0, c 1 and d 2, with the simple matching coefficient taking a value
of 0.667.
Two other types of matching coefficients, which do not equate the joint absence
of a characteristic with similarity and may, therefore, be of more value in segmen-
tation studies, are the Jaccard (JC) and the Russel and Rao (RR) coefficients. They
are defined as follows:
a
JC
abc
a
RR
abcd

These matching coefficients are just like the distance measures used to
determine a cluster solution. There are many other matching coefficients such as
Yules Q, Kulczynski or Ochiai, but since most applications of cluster analysis rely
on metric or ordinal data, we will not discuss these in greater detail.7
For nominal variables with more than two categories, you should always convert
the categorical variable into a set of binary variables in order to use matching
coefficients. When you have ordinal data, you should always use distance measures
such as Euclidean distance. Even though using matching coefficients would be
feasible and from a strictly statistical standpoint even more appropriate, you
would disregard variable information in the sequence of the categories. In the end, a
respondent who indicates that he or she is very loyal to a brand is going to be closer
to someone who is somewhat loyal than a respondent who is not loyal at all.
Furthermore, distance measures best represent the concept of proximity, which is
fundamental to cluster analysis.
Most datasets contain variables that are measured on multiple scales. For
example, a market research questionnaire may ask about the respondents income,
product ratings, and last brand purchased. Thus, we have to consider variables
measured on a ratio, ordinal, and nominal scale. How can we simultaneously
incorporate these variables into one analysis? Unfortunately, this problem cannot
be easily resolved and, in fact, many market researchers simply ignore the scale
level. Instead, they use one of the distance measures discussed in the context
of metric (and ordinal) data. Even though this approach may slightly change
the results when compared to those using matching coefficients, it should not be
rejected. Cluster analysis is mostly an exploratory technique whose results provide
a rough guidance for managerial decisions. Despite this, there are several proce-
dures that allow a simultaneous integration of these variables into one analysis.

7
See Wedel and Kamakura (2000) for more information on alternative matching coefficients.
250 9 Cluster Analysis

First, we could compute distinct distance matrices for each group of variables; that
is, one distance matrix based on, for example, ordinally scaled variables and another
based on nominal variables. Afterwards, we can simply compute the weighted
arithmetic mean of the distances and use this average distance matrix as the input
for the cluster analysis. However, the weights have to be determined a priori
and improper weights may result in a biased treatment of different variable types.
Furthermore, the computation and handling of distance matrices are not trivial.
Using the SPSS syntax, one has to manually add the MATRIX subcommand, which
exports the initial distance matrix into a new data file. Go to the 8 Web Appendix
(! Chap. 5) to learn how to modify the SPSS syntax accordingly.
Second, we could dichotomize all variables and apply the matching coefficients
discussed above. In the case of metric variables, this would involve specifying
categories (e.g., low, medium, and high income) and converting these into sets of
binary variables. In most cases, however, the specification of categories would be
rather arbitrary and, as mentioned earlier, this procedure could lead to a severe loss
of information.
In the light of these issues, you should avoid combining metric and nominal
variables in a single cluster analysis, but if this is not feasible, the two-step clustering
procedure provides a valuable alternative, which we will discuss later. Lastly, the
choice of the (dis)similarity measure is not extremely critical to recovering the
underlying cluster structure. In this regard, the choice of the clustering algorithm
is far more important. We therefore deal with this aspect in the following section.

Select a Clustering Algorithm

After having chosen the distance or similarity measure, we need to decide which
clustering algorithm to apply. There are several agglomerative procedures and they
can be distinguished by the way they define the distance from a newly formed
cluster to a certain object, or to other clusters in the solution. The most popular
agglomerative clustering procedures include the following:
l Single linkage (nearest neighbor): The distance between two clusters corre-
sponds to the shortest distance between any two members in the two clusters.
l Complete linkage (furthest neighbor): The oppositional approach to single
linkage assumes that the distance between two clusters is based on the longest
distance between any two members in the two clusters.
l Average linkage: The distance between two clusters is defined as the average
distance between all pairs of the two clusters members.
l Centroid: In this approach, the geometric center (centroid) of each cluster is
computed first. The distance between the two clusters equals the distance bet-
ween the two centroids.
Figures 9.59.8 illustrate these linkage procedures for two randomly framed
clusters.
Conducting a Cluster Analysis 251

Fig. 9.5 Single linkage

Fig. 9.6 Complete linkage

Fig. 9.7 Average linkage

Fig. 9.8 Centroid

252 9 Cluster Analysis

Each of these linkage algorithms can yield totally different results when used
on the same dataset, as each has its specific properties. As the single linkage
algorithm is based on minimum distances, it tends to form one large cluster with
the other clusters containing only one or few objects each. We can make use of
this chaining effect to detect outliers, as these will be merged with the remain-
ing objects usually at very large distances in the last steps of the analysis.
Generally, single linkage is considered the most versatile algorithm. Conversely,
the complete linkage method is strongly affected by outliers, as it is based on
maximum distances. Clusters produced by this method are likely to be rather
compact and tightly clustered. The average linkage and centroid algorithms tend
to produce clusters with rather low within-cluster variance and similar sizes.
However, both procedures are affected by outliers, though not as much as
complete linkage.
Another commonly used approach in hierarchical clustering is Wards method.
This approach does not combine the two most similar objects successively. Instead,
those objects whose merger increases the overall within-cluster variance to the
smallest possible degree, are combined. If you expect somewhat equally sized
clusters and the dataset does not include outliers, you should always use Wards
method.
To better understand how a clustering algorithm works, lets manually examine
some of the single linkage procedures calculation steps. We start off by looking at
the initial (Euclidean) distance matrix in Table 9.3. In the very first step, the two
objects exhibiting the smallest distance in the matrix are merged. Note that we
always merge those objects with the smallest distance, regardless of the clustering
procedure (e.g., single or complete linkage). As we can see, this happens to two
pairs of objects, namely B and C (d(B, C) 1.414), as well as C and E (d(C, E)
1.414). In the next step, we will see that it does not make any difference whether we
first merge the one or the other, so lets proceed by forming a new cluster, using
objects B and C.
Having made this decision, we then form a new distance matrix by considering
the single linkage decision rule as discussed above. According to this rule, the
distance from, for example, object A to the newly formed cluster is the minimum of
d(A, B) and d(A, C). As d(A, C) is smaller than d(A, B), the distance from A to the
newly formed cluster is equal to d(A, C); that is, 2.236. We also compute the
distances from cluster [B,C] (clusters are indicated by means of squared brackets)
to all other objects (i.e. D, E, F, G) and simply copy the remaining distances such
as d(E, F) that the previous clustering has not affected. This yields the distance
matrix shown in Table 9.6.
Continuing the clustering procedure, we simply repeat the last step by merging
the objects in the new distance matrix that exhibit the smallest distance (in this case,
the newly formed cluster [B, C] and object E) and calculate the distance from this
cluster to all other objects. The result of this step is described in Table 9.7.
Try to calculate the remaining steps yourself and compare your solution with the
distance matrices in the following Tables 9.89.10.
Conducting a Cluster Analysis 253

Table 9.6 Distance matrix after first clustering step (single linkage)
Objects A B, C D E F G
A 0
B, C 2.236 0
D 2 2.236 0
E 3.606 1.414 3 0
F 4.123 3.162 2.236 2.828 0
G 5.385 5.657 3.606 5.831 3.162 0

Table 9.7 Distance matrix after second clustering step (single linkage)
Objects A B, C, E D F G
A 0
B, C, E 2.236 0
D 2 2.236 0
F 4.123 2.828 2.236 0
G 5.385 5.657 3.606 3.162 0

Table 9.8 Distance matrix after third clustering step (single linkage)
Objects A, D B, C, E F G
A, D 0
B, C, E 2.236 0
F 2.236 2.828 0
G 3.606 5.657 3.162 0

Table 9.9 Distance matrix after fourth clustering step (single linkage)
Objects A, B, C, D, E F G
A, B, C, D, E 0
F 2.236 0
G 3.606 3.162 0

Table 9.10 Distance matrix after fifth clustering step (single linkage)
Objects A, B, C, D, E, F G
A, B, C, D, E, F 0
G 3.162 0

By following the single linkage procedure, the last steps involve the merger
of cluster [A,B,C,D,E,F] and object G at a distance of 3.162. Do you get the same
results? As you can see, conducting a basic cluster analysis manually is not that
hard at all not if there are only a few objects in the dataset.
A common way to visualize the cluster analysiss progress is by drawing a
dendrogram, which displays the distance level at which there was a combination
of objects and clusters (Fig. 9.9).
We read the dendrogram from left to right to see at which distance objects
have been combined. For example, according to our calculations above, objects
B, C, and E are combined at a distance level of 1.414.
254 9 Cluster Analysis

0 1 2 3
Distance

Fig. 9.9 Dendrogram

Decide on the Number of Clusters

An important question we havent yet addressed is how to decide on the number of

clusters to retain from the data. Unfortunately, hierarchical methods provide only
very limited guidance for making this decision. The only meaningful indicator
relates to the distances at which the objects are combined. Similar to factor
analysiss scree plot, we can seek a solution in which an additional combination
of clusters or objects would occur at a greatly increased distance. This raises the
issue of what a great distance is, of course.
One potential way to solve this problem is to plot the number of clusters on the
x-axis (starting with the one-cluster solution at the very left) against the distance at
which objects or clusters are combined on the y-axis. Using this plot, we then search
for the distinctive break (elbow). SPSS does not produce this plot automatically
you have to use the distances provided by SPSS to draw a line chart by using a
common spreadsheet program such as Microsoft Excel.
Alternatively, we can make use of the dendrogram which essentially carries the
same information. SPSS provides a dendrogram; however, this differs slightly from
the one presented in Fig. 9.9. Specifically, SPSS rescales the distances to a range
of 025; that is, the last merging step to a one-cluster solution takes place at a
(rescaled) distance of 25. The rescaling often lengthens the merging steps, thus
making breaks occurring at a greatly increased distance level more obvious.
Despite this, this distance-based decision rule does not work very well in all
cases. It is often difficult to identify where the break actually occurs. This is also
the case in our example above. By looking at the dendrogram, we could justify
a two-cluster solution ([A,B,C,D,E,F] and [G]), as well as a five-cluster solution
([B,C,E], [A], [D], [F], [G]).
Conducting a Cluster Analysis 255

Research has suggested several other procedures for determining the number of
clusters in a dataset. Most notably, the variance ratio criterion (VRC) by Calinski
and Harabasz (1974) has proven to work well in many situations.8 For a solution
with n objects and k segments, the criterion is given by:

VRCk SSB =k  1=SSW =n  k;

where SSB is the sum of the squares between the segments and SSW is the sum of the
squares within the segments. The criterion should seem familiar, as this is nothing
but the F-value of a one-way ANOVA, with k representing the factor levels.
Consequently, the VRC can easily be computed using SPSS, even though it is not
readily available in the clustering procedures outputs.
To finally determine the appropriate number of segments, we compute ok for
each segment solution as follows:

ok VRCk1  VRCk  VRCk  VRCk1 :

In the next step, we choose the number of segments k that minimizes the value in
ok. Owing to the term VRCk1, the minimum number of clusters that can be
selected is three, which is a clear disadvantage of the criterion, thus limiting its
application in practice.
Overall, the data can often only provide rough guidance regarding the number of
clusters you should select; consequently, you should rather revert to practical
considerations. Occasionally, you might have a priori knowledge, or a theory on
which you can base your choice. However, first and foremost, you should ensure
that your results are interpretable and meaningful. Not only must the number of
clusters be small enough to ensure manageability, but each segment should also be
large enough to warrant strategic attention.

Partitioning Methods: k-means

Another important group of clustering procedures are partitioning methods. As with

hierarchical clustering, there is a wide array of different algorithms; of these, the
k-means procedure is the most important one for market research.9 The k-means
algorithm follows an entirely different concept than the hierarchical methods
discussed before. This algorithm is not based on distance measures such as
Euclidean distance or city-block distance, but uses the within-cluster variation as a

8
Milligan and Cooper (1985) compare various criteria.
9
Note that the k-means algorithm is one of the simplest non-hierarchical clustering methods.
Several extensions, such as k-medoids (Kaufman and Rousseeuw 2005) have been proposed to
handle problematic aspects of the procedure. More advanced methods include finite mixture
models (McLachlan and Peel 2000), neural networks (Bishop 2006), and self-organizing maps
(Kohonen 1982). Andrews and Currim (2003) discuss the validity of some of these approaches.
256 9 Cluster Analysis

measure to form homogenous clusters. Specifically, the procedure aims at segmenting

the data in such a way that the within-cluster variation is minimized. Consequently, we
do not need to decide on a distance measure in the first step of the analysis.
The clustering process starts by randomly assigning objects to a number of
clusters.10 The objects are then successively reassigned to other clusters to mini-
mize the within-cluster variation, which is basically the (squared) distance from
each observation to the center of the associated cluster. If the reallocation of an
object to another cluster decreases the within-cluster variation, this object is reas-
signed to that cluster.
With the hierarchical methods, an object remains in a cluster once it is assigned
to it, but with k-means, cluster affiliations can change in the course of the clustering
process. Consequently, k-means does not build a hierarchy as described before
(Fig. 9.3), which is why the approach is also frequently labeled as non-hierarchical.
For a better understanding of the approach, lets take a look at how it works in
practice. Figs. 9.109.13 illustrate the k-means clustering process.
Prior to analysis, we have to decide on the number of clusters. Our client could,
for example, tell us how many segments are needed, or we may know from previous
research what to look for. Based on this information, the algorithm randomly selects
a center for each cluster (step 1). In our example, two cluster centers are randomly
initiated, which CC1 (first cluster) and CC2 (second cluster) in Fig. 9.10

A B
CC1

D E
Brand loyalty (y)

CC2

Price consciousness (x)

Fig. 9.10 k-means procedure (step 1)

10
Note this holds for the algorithms original design. SPSS does not choose centers randomly.
Conducting a Cluster Analysis 257

A B
CC1

D E
Brand loyalty (y)

CC2

Price consciousness (x)

Fig. 9.11 k-means procedure (step 2)

A B
CC1
CC1
C

D E
Brand loyalty (y)

CC2

CC2
F

Price consciousness (x)

Fig. 9.12 k-means procedure (step 3)

258 9 Cluster Analysis

A B
CC1

C
Brand loyalty (y)

D E

CC2

Price consciousness (x)

Fig. 9.13 k-means procedure (step 4)

represent.11 After this (step 2), Euclidean distances are computed from the cluster
centers to every single object. Each object is then assigned to the cluster center with
the shortest distance to it. In our example (Fig. 9.11), objects A, B, and C are
assigned to the first cluster, whereas objects D, E, F, and G are assigned to the
second. We now have our initial partitioning of the objects into two clusters.
Based on this initial partition, each clusters geometric center (i.e., its centroid)
is computed (third step). This is done by computing the mean values of the objects
contained in the cluster (e.g., A, B, C in the first cluster) regarding each of the variables
(price consciousness and brand loyalty). As we can see in Fig. 9.12, both clusters
centers now shift into new positions (CC1 for the first and CC2 for the second cluster).
In the fourth step, the distances from each object to the newly located cluster
centers are computed and objects are again assigned to a certain cluster on the basis
of their minimum distance to other cluster centers (CC1 and CC2). Since the
cluster centers position changed with respect to the initial situation in the first step,
this could lead to a different cluster solution. This is also true of our example, as
object E is now unlike in the initial partition closer to the first cluster center
(CC1) than to the second (CC2). Consequently, this object is now assigned to the
first cluster (Fig. 9.13). The k-means procedure now repeats the third step and
re-computes the cluster centers of the newly formed clusters, and so on. In other

11
Conversely, SPSS always sets one observation as the cluster center instead of picking some
random point in the dataset.
Conducting a Cluster Analysis 259

words, steps 3 and 4 are repeated until a predetermined number of iterations are
reached, or convergence is achieved (i.e., there is no change in the cluster affiliations).
Generally, k-means is superior to hierarchical methods as it is less affected by
outliers and the presence of irrelevant clustering variables. Furthermore, k-means
can be applied to very large datasets, as the procedure is less computationally
demanding than hierarchical methods. In fact, we suggest definitely using k-means
for sample sizes above 500, especially if many clustering variables are used. From
a strictly statistical viewpoint, k-means should only be used on interval or ratio-
scaled data as the procedure relies on Euclidean distances. However, the procedure is
routinely used on ordinal data as well, even though there might be some distortions.
One problem associated with the application of k-means relates to the fact that
the researcher has to pre-specify the number of clusters to retain from the data. This
makes k-means less attractive to some and still hinders its routine application in
practice. However, the VRC discussed above can likewise be used for k-means
clustering (an application of this index can be found in the 8 Web Appendix !
Chap. 9). Another workaround that many market researchers routinely use is to
apply a hierarchical procedure to determine the number of clusters and k-means
afterwards.12 This also enables the user to find starting values for the initial cluster
centers to handle a second problem, which relates to the procedures sensitivity to
the initial classification (we will follow this approach in the example application).

Two-Step Clustering

We have already discussed the issue of analyzing mixed variables measured on

different scale levels in this chapter. The two-step cluster analysis developed
by Chiu et al. (2001) has been specifically designed to handle this problem. Like
k-means, the procedure can also effectively cope with very large datasets.
The name two-step clustering is already an indication that the algorithm is based
on a two-stage approach: In the first stage, the algorithm undertakes a procedure
that is very similar to the k-means algorithm. Based on these results, the two-step
procedure conducts a modified hierarchical agglomerative clustering procedure that
combines the objects sequentially to form homogenous clusters. This is done by
building a so-called cluster feature tree whose leaves represent distinct objects in
the dataset. The procedure can handle categorical and continuous variables simul-
taneously and offers the user the flexibility to specify the cluster numbers as well as
the maximum number of clusters, or to allow the technique to automatically choose
the number of clusters on the basis of statistical evaluation criteria. Likewise, the
procedure guides the decision of how many clusters to retain from the data by
calculating measures-of-fit such as Akaikes Information Criterion (AIC) or Bayes

12
See Punji and Stewart (1983) for additional information on this sequential approach.
260 9 Cluster Analysis

Information Criterion (BIC). Furthermore, the procedure indicates each variables

importance for the construction of a specific cluster. These desirable features make
the somewhat less popular two-step clustering a viable alternative to the traditional
methods. You can find a more detailed discussion of the two-step clustering
procedure in the 8 Web Appendix (! Chap. 9), but we will also apply this method
in the subsequent example.

Validate and Interpret the Cluster Solution

Before interpreting the cluster solution, we have to assess the solutions stability
and validity. Stability is evaluated by using different clustering procedures on the
same data and testing whether these yield the same results. In hierarchical cluster-
ing, you can likewise use different distance measures. However, please note that it
is common for results to change even when your solution is adequate. How much
variation you should allow before questioning the stability of your solution is a
matter of taste. Another common approach is to split the dataset into two halves and
to thereafter analyze the two subsets separately using the same parameter settings.
You then compare the two solutions cluster centroids. If these do not differ
significantly, you can presume that the overall solution has a high degree of
stability. When using hierarchical clustering, it is also worthwhile changing the
order of the objects in your dataset and re-running the analysis to check the results
stability. The results should not, of course, depend on the order of the dataset. If
they do, you should try to ascertain if any obvious outliers may influence the results
of the change in order.
Assessing the solutions reliability is closely related to the above, as reliability
refers to the degree to which the solution is stable over time. If segments quickly
change their composition, or its members their behavior, targeting strategies are
likely not to succeed. Therefore, a certain degree of stability is necessary to ensure
that marketing strategies can be implemented and produce adequate results. This
can be evaluated by critically revisiting and replicating the clustering results at
a later point in time.
To validate the clustering solution, we need to assess its criterion validity.
In research, we could focus on criterion variables that have a theoretically based
relationship with the clustering variables, but were not included in the analysis.
In market research, criterion variables usually relate to managerial outcomes
such as the sales per person, or satisfaction. If these criterion variables differ signifi-
cantly, we can conclude that the clusters are distinct groups with criterion validity.
To judge validity, you should also assess face validity and, if possible, expert
validity. While we primarily consider criterion validity when choosing clustering
variables, as well as in this final step of the analysis procedure, the assessment of face
validity is a process rather than a single event. The key to successful segmentation is
to critically revisit the results of different cluster analysis set-ups (e.g., by using
Conducting a Cluster Analysis 261

different algorithms on the same data) in terms of managerial relevance. This under-
lines the exploratory character of the method. The following criteria will help you
make an evaluation choice for a clustering solution (Dibb 1999; Tonks 2009; Kotler
and Keller 2009).
l Substantial: The segments are large and profitable enough to serve.
l Accessible: The segments can be effectively reached and served, which requires
them to be characterized by means of observable variables.
l Differentiable: The segments can be distinguished conceptually and respond
differently to different marketing-mix elements and programs.
l Actionable: Effective programs can be formulated to attract and serve the
segments.
l Stable: Only segments that are stable over time can provide the necessary
grounds for a successful marketing strategy.
l Parsimonious: To be managerially meaningful, only a small set of substantial
clusters should be identified.
l Familiar: To ensure management acceptance, the segments composition should
be comprehensible.
l Relevant: Segments should be relevant in respect of the companys competencies
and objectives.
l Compactness: Segments exhibit a high degree of within-segment homogeneity
and between-segment heterogeneity.
l Compatibility: Segmentation results meet other managerial functions require-
ments.
The final step of any cluster analysis is the interpretation of the clusters.
Interpreting clusters always involves examining the cluster centroids, which are
the clustering variables average values of all objects in a certain cluster. This step
is of the utmost importance, as the analysis sheds light on whether the segments are
conceptually distinguishable. Only if certain clusters exhibit significantly different
means in these variables are they distinguishable from a data perspective, at least.
This can easily be ascertained by comparing the clusters with independent t-tests
samples or ANOVA (see Chap. 6).
By using this information, we can also try to come up with a meaningful name or
label for each cluster; that is, one which adequately reflects the objects in the
cluster. This is usually a very challenging task. Furthermore, clustering variables
are frequently unobservable, which poses another problem. How can we decide to
which segment a new object should be assigned if its unobservable characteristics,
such as personality traits, personal values or lifestyles, are unknown? We could
obviously try to survey these attributes and make a decision based on the clustering
variables. However, this will not be feasible in most situations and researchers
therefore try to identify observable variables that best mirror the partition of the
objects. If it is possible to identify, for example, demographic variables leading to a
very similar partition as that obtained through the segmentation, then it is easy to
assign a new object to a certain segment on the basis of these demographic
262 9 Cluster Analysis

characteristics. These variables can then also be used to characterize specific

segments, an action commonly called profiling.
For example, imagine that we used a set of items to assess the respondents
values and learned that a certain segment comprises respondents who appreciate
self-fulfillment, enjoyment of life, and a sense of accomplishment, whereas this is
not the case in another segment. If we were able to identify explanatory variables
such as gender or age, which adequately distinguish these segments, then we could
partition a new person based on the modalities of these observable variables whose
traits may still be unknown.
Table 9.11 summarizes the steps involved in a hierarchical and k-means clustering.

While companies often develop their own market segments, they frequently use
standardized segments, which are based on established buying trends, habits, and
customers needs and have been specifically designed for use by many products in
mature markets. One of the most popular approaches is the PRIZM lifestyle
segmentation system developed by Claritas Inc., a leading market research com-
pany. PRIZM defines every US household in terms of 66 demographically and
behaviorally distinct segments to help marketers discern those consumers likes,
dislikes, lifestyles, and purchase behaviors.
Visit the Claritas website and flip through the various segment profiles. By
entering a 5-digit US ZIP code, you can also find a specific neighborhoods top
five lifestyle groups.
One example of a segment is Gray Power, containing middle-class, home-
owning suburbanites who are aging in place rather than moving to retirement
communities. Gray Power reflects this trend, a segment of older, midscale
singles and couples who live in quiet comfort.

http://www.claritas.com/MyBestSegments/Default.jsp

We also introduce steps related to two-step clustering which we will further

introduce in the subsequent example.
Conducting a Cluster Analysis 263

Table 9.11 Steps involved in carrying out a factor analysis in SPSS

Theory Action
Research problem
Identification of homogenous groups
of objects in a population
Select clustering variables that should be Select relevant variables that potentially exhibit
used to form segments high degrees of criterion validity with regard
to a specific managerial objective.
Requirements
Sufficient sample size Make sure that the relationship between objects
and clustering variables is reasonable (rough
guideline: number of observations should be at
least 2m, where m is the number of clustering
variables). Ensure that the sample size is large
enough to guarantee substantial segments.
Low levels of collinearity among the variables Analyze Correlate Bivariate
Eliminate or replace highly correlated variables
(correlation coefficients > 0.90).
Specification
Choose the clustering procedure If there is a limited number of objects in your
dataset or you do not know the number of
clusters:
Analyze Classify Hierarchical Cluster
If there are many observations (> 500) in your
dataset and you have a priori knowledge
regarding the number of clusters:
Analyze Classify K-Means Cluster
If there are many observations in your dataset and
the clustering variables are measured on
different scale levels:
Analyze Classify Two-Step Cluster
Select a measure of similarity or dissimilarity Hierarchical methods:
(only hierarchical and two-step clustering) Analyze Classify Hierarchical Cluster
Method Measure
Depending on the scale level, select the measure;
convert variables with multiple categories into
a set of binary variables and use matching
coefficients; standardize variables if necessary
(on a range of 0 to 1 or 1 to 1).
Two-step clustering:
Analyze Classify Two-Step Cluster
Distance Measure
Use Euclidean distances when all variables
are continuous; for mixed variables, use
log-likelihood.
Choose clustering algorithm Analyze Classify Hierarchical Cluster
(only hierarchical clustering) Method Cluster Method
Use Wards method if equally sized clusters are
expected and no outliers are present.
Preferably use single linkage, also to detect
outliers.
Decide on the number of clusters Hierarchical clustering:
Examine the dendrogram:
Analyze Classify Hierarchical Cluster
Plots Dendrogram
(continued)
264 9 Cluster Analysis

Table 9.11 (continued)

Theory Action
Draw a scree plot (e.g., using Microsoft Excel)
based on the coefficients in the agglomeration
schedule.
Compute the VRC using the ANOVA procedure:
Analyze Compare Means One-Way
ANOVA
Move the cluster membership variable in the
Factor box and the clustering variables in the
Dependent List box.
Compute VRC for each segment solution and
compare values.
k-means:
Run a hierarchical cluster analysis and decide
on the number of segments based on a
dendrogram or scree plot; use this information
to run k-means with k clusters.
Compute the VRC using the ANOVA procedure:
Analyze Classify K-Means Cluster
Options ANOVA table; Compute VRC for
each segment solution and compare values.
Two-step clustering:
Specify the maximum number of clusters:
Analyze Classify Two-Step Cluster
Number of Clusters
Run separate analyses using AIC and,
alternatively, BIC as clustering criterion:
Analyze Classify Two-Step Cluster
Clustering Criterion
Examine the auto-clustering output.
Validate and interpret the cluster solution
Assess the solutions stability Re-run the analysis using different clustering
procedures, algorithms or distance measures.
Split the datasets into two halves and compute the
clustering variables centroids; compare
centroids for significant differences (e.g.,
independent-samples t-test or one-way
ANOVA).
Change the ordering of objects in the dataset
(hierarchical clustering only).
Assess the solutions reliability Replicate the analysis using a separate, newly
collected dataset.
Assess the solutions validity Criterion validity:
Evaluate whether there are significant differences
between the segments with regard to one or
more criterion variables.
Face and expert validity:
Segments should be substantial, accessible,
differentiable, actionable, stable,
parsimonious, familiar and relevant. Segments
should exhibit high degrees of within-segment
homogeneity and between-segment
heterogeneity. The segmentation results
should meet the requirements of other
managerial functions.
(continued)
Example 265

Table 9.11 (continued)

Theory Action
Interpret the cluster solution Examine cluster centroids and assess whether
these differ significantly from each other (e.g.,
by means of t-tests or ANOVA; see Chap. 6).
Identify names or labels for each cluster and
characterize each cluster by means of
observable variables, if necessary (cluster
profiling).

Example

Silicon-Valley-based Tesla Motors Inc. (http://www.teslamotors.com) is an auto-

mobile startup company focusing on the production of high performance electrical
vehicles with exceptional design, technology, performance, and efficiency. Having
reported the 500th delivery of its roadster in June 2009, the company decided to
focus more strongly on the European market. However, as the company has only
limited experience in the European market, which has different characteristics than
that of the US, it asked a market research firm to provide a segmentation concept.
Consequently, the market research firm gathered data from major car manufacturers
on the following car characteristics, all of which have been measured on a ratio
scale (variable names in parentheses):
l Engine displacement (displacement)
l Turning moment in Nm (moment)
l Horsepower (horsepower)
l Length in mm (length)
l Width in mm (width)
l Net weight in kg (weight)
l Trunk volume in liters (trunk)
l Maximum speed in km/h (speed)
l Acceleration 0100 km/h in seconds (acceleration)

http://www.teslamotors.com
266 9 Cluster Analysis

The pretest sample of 15, randomly taken, cars is shown in Fig. 9.14. In practice,
clustering is done on much larger samples but we use a small sample size to
illustrate the clustering process. Keep in mind that in this example, the ratio
between the objects and clustering variables is much too small. The dataset used
is cars.sav (8 Web Appendix ! Chap. 9).

Fig. 9.14 Data

In the next step, we will run several different clustering procedures on the basis
of these nine variables. We first apply a hierarchical cluster analysis based on
Euclidean distances, using the single linkage method. This will help us determine
a suitable number of segments, which we will use as input for a subsequent k-means
clustering. Finally, we will run a two-step cluster analysis using SPSS.
Before we start with the clustering process, we have to examine the variables for
substantial collinearity. Just by looking at the variable set, we suspect that there are
some highly correlated variables in our dataset. For example, we expect rather high
correlations between speed and acceleration. To determine this, we run a bivariate
correlation analysis by clicking Analyze Correlate Bivariate, which will open
a dialog box similar to that in Fig. 9.15. Enter all variables into the Variables box
and select the box Pearson (under Correlation Coefficients) because these are
continuous variables.
The correlation matrix in Table 9.12 supports our expectations there are
several variables that have high correlations. Displacement exhibits high (absolute)
correlation coefficients with horsepower, speed, and acceleration, with values well
above 0.90, indicating possible collinearity issues. Similarly, horsepower is highly
correlated with speed and acceleration. Likewise, length shows a high degree of
correlation with width, weight, and trunk.
A potential solution to this problem would be to run a factor analysis and
perform a cluster analysis on the resulting factor scores. Since the factors obtained
are, by definition, independent, this would allow for an effective handling of the
collinearity issue. However, as this approach is associated with several problems
(see discussion above) and as there are only several variables in our data set,
we should reduce the variables, for example, by omitting displacement, horsepower,
Example 267

Fig. 9.15 Bivariate correlations dialog box

and length from the subsequent analyses. The remaining variables still provide a
sound basis for carrying out cluster analysis.
To run the hierarchical clustering procedure, click on Analyze Classify
Hierarchical Cluster, which opens a dialog box similar to Fig. 9.16.
Move the variables moment, width, weight, trunk, speed, and acceleration into
the Variable(s) box and specify name as the labeling variable (box Label Cases
by). The Statistics option gives us the opportunity to request the distance matrix
(labeled proximity matrix in this case) and the agglomeration schedule, which
provides information on the objects being combined at each stage of the clustering
process. Furthermore, we can specify the number or range of clusters to retain from
the data. As we do not yet know how many clusters to retain, just check the box
Agglomeration schedule and continue.
Under Plots, we choose to display a dendrogram, which graphically displays the
distances at which objects and clusters are joined. Also ensure you select the icicle
diagram (for all clusters), which is yet another graph for displaying clustering
solutions.
The option Method allows us to specify the cluster method (e.g., single linkage
or Wards method), the distance measure (e.g., Chebychev distance or the Jaccard
coefficient), and the type of standardization of values. In this example, we use the
single linkage method (Nearest neighbor) based on Euclidean distances. Since
the variables are measured on different levels (e.g., speed versus weight), make sure
to standardize the variables, using, for example, the Range 1 to 1 (by variable) in
the Transform Values drop-down list.
Table 9.12 Correlation matrix
268

Correlations
displacement moment horsepower length width weight trunk speed acceleration

displacement Pearson Correlation 1 .875** .983** .657** .764** .768** .470 .967** .969**
Sig. (2-tailed) .000 .000 .008 .001 .001 .077 .000 .000
N 15 15 15 15 15 15 15 15 15
moment Pearson Correlation .875** 1 .847** .767** .766** .862** .691** .859** .861**
Sig. (2-tailed) .000 .000 .001 .001 .000 .004 .000 .000
N 15 15 15 15 15 15 15 15 15
horsepower Pearson Correlation .983** .847** 1 .608* .732** .714** .408 .968** .961**
Sig. (2-tailed) .000 .000 .016 .002 .003 .131 .000 .000
N 15 15 15 15 15 15 15 15 15
length Pearson Correlation .657** .767** .608* 1 .912** .921** .934** .741** .714**
Sig. (2-tailed) .008 .001 .016 .000 .000 .000 .002 .003
N 15 15 15 15 15 15 15 15 15
width Pearson Correlation .764** .766** .732** .912** 1 .884** .783** .819** .818**
Sig. (2-tailed) .001 .001 .002 .000 .000 .001 .000 .000
N 15 15 15 15 15 15 15 15 15
weight Pearson Correlation .768** .862** .714** .921** .884** 1 .785** .778** .763**
Sig. (2-tailed) .001 .000 .003 .000 .000 .001 .001 .001
N 15 15 15 15 15 15 15 15 15
trunk Pearson Correlation .470 .691** .408 .934** .783** .785** 1 .579* .552*
Sig. (2-tailed) .077 .004 .131 .000 .001 .001 .024 .033
N 15 15 15 15 15 15 15 15 15
speed Pearson Correlation .967** .859** .968** .741** .819** .778** .579* 1 .971**
Sig. (2-tailed) .000 .000 .000 .002 .000 .001 .024 .000
N 15 15 15 15 15 15 15 15 15
acceleration Pearson Correlation .969** .861** .961** .714** .818** .763** .552* .971** 1
Sig. (2-tailed) .000 .000 .000 .003 .000 .001 .033 .000
N 15 15 15 15 15 15 15 15 15
**. Correlation is significant at the 0.01 level (2-tailed).
9 Cluster Analysis

*. Correlation is significant at the 0.05 level (2-tailed).

Example 269

Fig. 9.16 Hierarchical cluster analysis dialog box

Lastly, the Save option enables us to save cluster memberships for a single
solution or a range of solutions. Saved variables can then be used in subsequent
analyses to explore differences between groups. As a start, we will skip this option,
so continue and click on OK in the main menu.

Table 9.13 Agglomeration schedule

Agglomeration Schedule

Stage Cluster Combined Stage Cluster First Appears

Cluster 1 Cluster 2 Coefficients Cluster 1 Cluster 2 Next Stage

1 5 6 .149 0 0 2
2 5 7 .184 1 0 3
3 4 5 .201 0 2 5
4 14 15 .213 0 0 6
5 3 4 .220 0 3 8

6 13 14 .267 0 4 11

7 11 12 .321 0 0 9

8 2 3 .353 0 5 10

9 10 11 .357 0 7 11

10 1 2 .389 0 8 14
11 10 13 .484 9 6 13

12 8 9 .575 0 0 13

13 8 10 .618 12 11 14

14 1 8 .910 10 13 0
270 9 Cluster Analysis

First, we take a closer look at the agglomeration schedule (Table 9.13), which
displays the objects or clusters combined at each stage (second and third column)
and the distances at which this merger takes place. For example, in the first stage,
objects 5 and 6 are merged at a distance of 0.149. From here onward, the resulting
cluster is labeled as indicated by the first object involved in this merger, which is
object 5. The last column on the very right tells you in which stage of the algorithm
this cluster will appear next. In this case, this happens in the second step, where it is
merged with object 7 at a distance of 0.184. The resulting cluster is still labeled 5,
and so on. Similar information is provided by the icicle diagram shown in
Fig. 9.17. Its name stems from the analogy to rows of icicles hanging from the eaves
of a house. The diagram is read from the bottom to the top, therefore the columns
correspond to the objects being clustered, and the rows represent the number of
clusters.

Fig. 9.17 Icicle diagram

As described earlier, we can use the agglomeration schedule to determine the

number of segments to retain from the data. By plotting the distances (Coefficients
column in Table 9.13) against the number of clusters, using a spreadsheet pro-
gram, we can generate a scree plot. The distinct break (elbow) generally indicates
the solution regarding where an additional combination of two objects or clusters
would occur at a greatly increased distance. Thus, the number of clusters prior to
this merger is the most probable solution. The scree plot - which we made
Example 271

Scree Plot
1
0.9

0.8
Coefficient (distance)

0.7
0.6
0.5
0.4

0.3

0.2
0.1
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Number of clusters

Fig. 9.18 Scree plot

separately using Microsoft Excel - (Fig. 9.18) does not show such a distinct break.
Note that unlike in the factor analysis we do not pick the solution with one
cluster less than indicated by the elbow. The sharp increase in distance when
switching from a one to a two-cluster solution occurs in almost all analyses and
must not be viewed as a reliable indicator for the decision regarding the number of
segments.
The scree plot in Fig. 9.18 shows that there is no clear elbow indicating a suitable
number of clusters to retain. Based on the results, one could argue for a five-segment
or six-segment solution. However, considering that there are merely 15 objects in the
dataset, this seems too many, as we then have very small (and, most probably,
meaningless) clusters. Consequently, a two, three or four-segment solution is
deemed more appropriate.
Lets take a look at the dendrogram shown in Fig. 9.19. We read the dendrogram
from left to right. Vertical lines are objects and clusters joined together their position
indicates the distance at which this merger takes place. When creating a dendrogram,
SPSS rescales the distances to a range of 025; that is, the last merging step to a one-
cluster solution takes place at a (rescaled) distance of 25. Note that this differs from our
manual calculation shown in Fig. 9.9, where we did not do any rescaling! Again, the
analysis only provides a rough guidance regarding the number of segments to retain.
The change in distances between the mergers indicates that (besides a two-segment
solution) both a three and four-segment solution are appropriate.
To clarify this issue, lets re-run the analysis, but this time we pre-specify dif-
ferent segment numbers to compare these with regard to content validity. To do
so, just re-run the analysis using hierarchical clustering. Now switch to the Save
option, specify a range of solutions from 2 to 4 and run the analysis. SPSS generates
272 9 Cluster Analysis

Fig. 9.19 Dendrogram

the same output but also adds three additional variables to your dataset (CLU4_1,
CLU3_1, and CLU2_1), which reflect each objects cluster membership for the
respective analysis. SPSS automatically places CLU in front, followed by the number
of clusters (4, 3, or 2), to identify each objects cluster membership. The results are
Table 9.14 Cluster memberships
Name Four clusters, Three clusters, Two clusters,
observation member observation member observation member
of cluster of cluster of cluster
Kia Picanto 1.1 Start 1 1 1
Suzuki Splash 1.0 1 1 1
Renault Clio 1.2 1 1 1
Dacia Sandero 1.6 1 1 1
Fiat Grande Punto 1.4 1 1 1
Peugot 207 1.4 1 1 1
Renault Clio 1.6 1 1 1
Porsche Cayman 2 2 2
Nissan 350Z 3 2 2
Mercedes C 200 CDI 4 3 2
VW Passat Variant 2.0 4 3 2
Skoda Octavia 2.0 4 3 2
Mercedes E 280 4 3 2
Audi A6 2.4 4 3 2
BMW 525i 4 3 2
Example 273

illustrated in Table 9.14. SPSS does not produce this table for us, so we need to
enter these cluster memberships ourselves in a table or spreadsheet.
When we view the results, a three-segment solution appears promising. In this
solution, the first segment comprises compact cars, whereas the second segment
contains sports cars, and the third limousines. Increasing the solution by one
segment would further split up the sports cars segment into two sub-segments.
This does not appear to be very helpful, as now two of the four segments comprise
only one object. This underlines the single linkage methods tendency to identify
outlier objects in this case the Nissan 350Z and Porsche Cayman. In this specific
example, the Nissan 350Z and Porsche Cayman should not be regarded as outliers
in a classical sense but rather as those cars which may be Teslas main competitors
in the sports car market.
In contrast, the two-segment solution appears to be rather imprecise considering
the vast differences in the mix of sports and middle-sized cars in this solution.
To get a better overview of the results, lets examine the cluster centroids; that is, the
mean values of the objects contained in the cluster on selected variables. To do so, we
split up the dataset using the Split File command ( Data Split File) (see Chap. 5).
This enables us to analyze the data on the basis of a grouping variables values. In this
case, we choose CLU3_1 as the grouping variable and select the option Compare
groups. Subsequently, we calculate descriptive statistics ( Analyze Descriptive
Statistics Descriptives, also see Chap. 5) and calculate the mean, minimum
and maximum values, as well as the standard deviations of the clustering variables.
Table 9.15 shows the results for the variables weight, speed, and acceleration.

Table 9.15 Cluster centroids

Descriptive Statistics

CLU3_1 N Minimum Maximum Mean Std. Deviation

1 weight 7 929 1215 1115.57 100.528

speed 7 154 180 170.00 9.950

acceleration 7 11.40 15.10 12.9571 1.50317

Valid N (listwise) 7
2 weight 2 1340 1610 1475.00 190.919
speed 2 250 275 262.50 17.678
acceleration 2 5.40 5.80 5.6000 .28284
Valid N (listwise) 2
3 weight 6 1425 1660 1560.17 81.081
speed 6 201 250 223.67 21.163

acceleration 6 7.30 10.80 9.1167 1.48649

Valid N (listwise) 6
274 9 Cluster Analysis

From the descriptive statistics, it seems that the first segment contains light-weight
compact cars (with a lower maximum speed and acceleration). In contrast, the second
segment comprises two sports cars with greater speed and acceleration, whereas the
third contains limousines with an increased weight and intermediate speed and
acceleration. Since the descriptives do not tell us if these differences are significant,
we could have used a one-way ANOVA (menu Analyze Compare Means One-
Way ANOVA) to calculate the cluster centroids and compare the differences formally.
In the next step, we want to use the k-means method on the data. We have previously
seen that we need to specify the number of segments when conducting k-means
clustering. SPSS then initiates cluster centers and assigns objects to the clusters
based on their minimum distance to these centers. Instead of letting SPSS choose the
centers, we can also save the centroids (cluster centers) from our previous analysis as
input for the k-means procedure. To do this, we need to do some data management in
SPSS, as the cluster centers have to be supplied in a specific format. Conse-
quently, we need to aggregate the data first (briefly introduced in Chap. 5).
By selecting Data Aggregate, a dialog box similar to Fig. 9.20 opens up.

Fig. 9.20 Aggregate data dialog box

Example 275

Note that we choose Display Variable Names instead of Display Variable Labels
by clicking the right mouse button on the left box showing the variables in the
dataset. Now we proceed by choosing the cluster membership variable (CLU3_1)
as a break variable and move the moment, width, weight, trunk, speed, and
acceleration variables into the Summaries of Variable(s) box. When using the
default settings, SPSS computes the variables mean values along the lines of the
break variable (indicated by the postfix _mean, which is added to each aggregate
variables name), which corresponds to the cluster centers that we need for the
k-means analysis. You can change each aggregate variables name from the original
one by removing the postfix _mean using the Name & Label option if you want to.
Lastly, we do not want to add the aggregated variables to the active dataset, but rather
need to create a new dataset comprising only the aggregated variables. You must
therefore check this under SAVE and specify a dataset label such as aggregate. When
clicking on OK, a new dataset labeled aggregate is created and opened automatically.
The new dataset is almost in the right format but we still need to change the
break variables name from CLU3_1 to cluster_ (SPSS will issue a warning but this
can be safely ignored). The final dataset should have the form shown in Fig. 9.21.
Now lets proceed by using k-means clustering. Make sure that you open the
original dataset and go to Analyze Classify K-Means Cluster, which brings up a
new dialog box (Fig. 9.22).

Fig. 9.21 Aggregated data file

As you did in the hierarchical clustering analysis, move the six clustering
variables to the Variables box and specify the case labels (variable name). To
use the cluster centers from our previous analysis, check the box Read initial and
click on Open dataset. You can now choose the dataset labeled aggregate. Specify
3, which corresponds to the result of the hierarchical clustering analysis, in the
Number of Clusters box. The Iterate option is of less interest to us. Instead, click
on Save and check the box Cluster Membership. This creates a new variable
indicating each objects final cluster membership. SPSS indicates whether each
observation is a member of cluster 1, 2, or 3. Under Options, you can request
several statistics and specify how missing values should be treated. Ensure that you
request the initial cluster centers as well as the ANOVA table and that you exclude
the missing values listwise (default). Now start the analysis.
The k-means procedure generates Tables 9.16 and 9.17, which show the initial
and final cluster centers. As you can see, these are identical (also compare
Fig. 9.21), which indicates that the initial partitioning of the objects in the first
step of the k-means procedure was retained during the analysis. This means that it
276 9 Cluster Analysis

Fig. 9.22 K-means cluster analysis dialog box

Table 9.16 Initial cluster centers

Initial Cluster Centers

Cluster

1 2 3
moment 117 347 282
width 1699 1808 1814
weight 1116 1475 1560
trunk 249 323 543
speed 170 263 224
acceleration 12.96 5.60 9.12

Input from FILE Subcommand

Example 277

Table 9.17 Final cluster centers

Final Cluster Centers

Cluster

1 2 3

moment 117 347 282

width 1699 1808 1814
weight 1116 1475 1560
trunk 249 323 543
speed 170 263 224
acceleration 12.96 5.60 9.12

was not possible to reduce the overall within-cluster variation by re-assigning

objects to different clusters.
Likewise, the output Iteration History shows that there is no change in the
cluster centers. Similarly, if you compare the partitioning of objects into the three
clusters by examining the newly generated variable QCL_1, you see that there is no
change in the clusters composition. At first sight, this does not look like a very
exciting result, but this in fact signals that the initial clustering solution is stable.
In other words, the fact that two different clustering methods yield the same
outcomes provides some evidence of the results stability.
In contrast to hierarchical clustering, the k-means outputs provide us with an
ANOVA of the cluster centers (Table 9.18). As you can see, all the clustering
variables means differ significantly across at least two of the three segments,
because the null hypothesis is rejected in every case (Sig.  0.05).

Table 9.18 ANOVA output

ANOVA

Cluster Error
Mean Square df Mean Square df F Sig.

moment 64318.455 2 784.224 12 82.015 .000

width 23904.771 2 1966.183 12 12.158 .001
weight 339920.393 2 10829.712 12 31.388 .000
trunk 142764.143 2 4311.754 12 33.110 .000
speed 8628.283 2 262.153 12 32.913 .000
acceleration 50.855 2 2.057 12 24.722 .000

Since we used the prior analysis results from hierarchical clustering as an input
for the k-means procedure, the problem of selecting the correct number of
segments is not problematic in this example. As discussed above, we could have
278 9 Cluster Analysis

also used the VRC to make that decision. In the 8 Web Appendix (! Chap. 9), we
present a VRC application to this example.
As a last step of the analysis, we conduct a two-step clustering approach. First,
go to Analyze Classify Two-Step Cluster. A new dialog box is opened, similar
to that shown in Fig. 9.23.

Fig. 9.23 Two-step cluster analysis dialog box

Move the variables we used in the previous analyses to the Continuous Vari-
ables box.
The Distance Measure box determines how the distance between two objects or
clusters is computed. While Log-likelihood can be used for categorical and contin-
uous variables, the Euclidean distance can only be applied when all of the variables
are continuous. Unless your dataset contains categorical variables (e.g., gender) you
should choose the Euclidean distance measure, as this generally provides better
results. If you use ordinal variables and therefore use the Log-likelihood procedure,
check that the answer categories are equidistant. In our dataset, all variables are
continuous, therefore select the second option, namely Euclidean.
Example 279

Under Number of Clusters, you can specify a fixed number or a maximum

number of segments to retain from the data. One of two-step clusterings major
advantages is that it allows the automatic selection of the number of clusters. To
make use of this advantage, you should specify a maximum number of clusters, for
example, 6. Next to this box, in which the number of clusters is specified, you can
choose between two criteria (also referred to as model selection or information
criteria) which SPSS can use to pick an appropriate number of segments, namely
Akaikes information criterion (AIC) and Bayes information criterion (BIC).
These are relative measures of goodness-of-fit and are used to compare different
solutions with different numbers of segments. Relative means that these criteria
are not scaled on a range of, for example, 0 to 1 but can generally take any value.
Compared to an alternative solution with a different number of segments, smaller
values in AIC or BIC indicate an increased fit. SPSS computes solutions for different
segment numbers (up to the maximum number of segments specified before) and
chooses the appropriate solution by looking for the smallest value in the chosen
criterion. However, which criterion should we choose? AIC is well-known for
overestimating the correct number of segments, while BIC has a slight tendency
to underestimate this number. Thus, it is worthwhile comparing the clustering
outcomes of both criteria and selecting a smaller number of segments than
actually indicated by AIC. Nevertheless, when running two separate analyses,
one based on AIC and the other based on BIC, SPSS usually renders the same
results. But what do we do if the two criteria indicate different numbers of
clusters? In such a situation, we should evaluate each solution on practical
grounds as well as in light of the solutions interpretability. Do not solely rely
on the automatic model selection, especially when there is a combination of
continuous and categorical variables, as this does not always work well. Examine
the results very carefully!
Under Options, you can specify issues related to outlier treatment, memory
allocation, and variable standardization. Variables that are already standardized
have to be assigned as such, but since this is not the case in our analysis, you can
simply proceed.
Lastly, under the option Output, we can specify additional variables for describ-
ing the obtained clusters. However, lets stick to the default option for now. Make
sure that you click the box Create cluster membership variable before clicking
Continue. Note that the menu options as well as the outputs in SPSS 18 are no
longer the same as in prior SPSS versions. Here, we discuss the menus and outputs
as provided in SPSS 18, but if you want to learn what the analysis looks like in
SPSS 17 (and prior versions), go to the 8 Web Appendix (! Chap. 9).
SPSS produces a very simple output, as shown in Fig. 9.24. The upper part of the
output describes the algorithm applied, the number of variables used (labeled input
features) and the final number of clusters retained from the data. In our case, the
number of clusters is chosen according to BIC, which indicates a two-segment
solution (the same holds when using AIC instead of BIC). Note that this result
differs from our previous analysis!
280 9 Cluster Analysis

Model Summary

Algorithm TwoStep

Input Features 6

Clusters 2

Cluster Quality

Poor Fair Good

1.0 0.5 0.0 0.5 1.0

Silhouette measure of cohesion and sepearation

Fig. 9.24 Two-step clustering output

The lower part of the output (Fig. 9.24) indicates the quality of the cluster
solution. The silhouette measure of cohesion and separation is a measure of the
clustering solutions overall goodness-of-fit. It is essentially based on the average
distances between the objects and can vary between 1 and 1. Specifically, a
silhouette measure of less than 0.20 indicates a poor solution quality, a measure
between 0.20 and 0.50 a fair solution, whereas values of more than 0.50 indicate a
good solution (this is also indicated under the horizontal bar in Fig. 9.24). In our
case, the measure indicates a satisfactory cluster quality. Consequently, you can
proceed with the analysis by double-clicking on the output. This will open up the
model viewer (Fig. 9.25), an evaluation tool that graphically presents the structure
of the revealed clusters.
The model viewer provides us with two windows: the main view, which initially
shows a model summary (left-hand side), and an auxiliary view, which initially
features the cluster sizes (right-hand side). At the bottom of each window, you can
request different information, such as an overview of the cluster structure and the
overall variable importance as shown in Fig. 9.25.
In the main view, we can see a description of the two clusters, including their
(relative) sizes. Furthermore, the output shows each clustering variables mean
values across the two clusters as well as their relative importance. Darker shades
(i.e., higher values in feature importance) denote the variables greater importance
for the clustering solution. Comparing the results, we can see that moment is the
most important variable for each of the clusters, followed by weight, speed, width,
acceleration, and trunk. Clicking on one of the boxes will show a graph with
the frequency distribution of each cluster. The auxiliary view shows an overview
Case Study 281

Fig. 9.25 Additional options in the model viewer

of the variables overall importance for the clustering solution, which provides the
same result as the cluster-specific analysis. The model viewer provides us with
additional options for visualizing the results or comparing clustering solutions. It is
worthwhile to simply play around with the different self-explanatory options. So go
ahead and explore the model viewers features yourself!

Case Study

Facing dramatically declining sales and decreased turnover, retailers such as Saks
Fifth Avenue and JCPenney are rethinking their pricing strategies, scaling back
inventories, and improving the fashion content. Mens accessories are one of the
bright spots and Saks Fifth Avenue has jumped on the trend with three recently
opened shops prominently featuring this category. The largest mens store opened
in Beverly Hills in the late 2008 and stocks top brands in jewelry, watches,
sunglasses, and leather goods. By providing a better showcase for mens acces-
sories, Saks aims at strengthening its position in a market that is often neglected in
the larger department store arena. This is because the mens accessories business
generally requires expertise in buying since this typically involves small, artisan
vendors an investment many department stores are not willing to make.
282 9 Cluster Analysis

The Beverly Hills store was chosen to spearhead the accessories program
because it is considered the companys West Coast flagship and the department
had not had a significant facelift since the store opened in 1995.13
Sakss strategy seemed to be successful if one considers that the newly opened
boutiques already exerted an impact on sales during their first holiday season.
However, before opening accessories shops in any other existing Saks stores, the
company wanted to gain further insights into their customers preferences. Con-
sequently, a survey was conducted among visitors of the Beverly Hills store to gain
a deeper understanding of their attitudes to buying and shopping. Overall, 180
respondents were interviewed using mall-intercept interviewing. The respondents
were asked to indicate the importance of the following factors when buying products
and services using a 5-point scale (1 not at all important, 5 very important):
l Saving time (x1)
l Getting bargains (x2)
l Getting products that arent on the high street (x3)
l Trying new things (x4)
l Being aware of what companies have to offer (x5)
The resulting dataset Buying Attitudes.sav (8 Web Appendix ! Chap. 9) also
includes each respondents gender and monthly disposable income.14
1. Given the levels of measurement, which clustering method would you prefer?
Carry out a cluster analysis using this procedure.
2. Interpret and profile the obtained clusters by examining cluster centroids.
Compare differences across clusters on observed variables using ANOVA and
post-hoc tests (see Chap. 6).
3. Use a different clustering method to test the stability of your results. If necessary,
omit or rescale certain variables.
4. Based on your evaluation of the dataset, make recommendations to the manage-
ment of Sakss Beverly Hills store.

Questions

1. In your own words, explain the objective and basic concept of cluster analysis.
2. What are the differences between hierarchical and partitioning methods? When
do we use hierarchical or partitioning methods?
3. Run the k-means analysis again from the example application (Cars.sav, 8 Web
Appendix ! Chap. 9). Compute a three-segment solution and compare the
results with those obtained by the initial hierarchical clustering.

13
For further information, see Palmieri JE (2008). Saks Adds Mens Accessories Shops,
Womens Wear Daily, 196 (128), 14.
14
Note that the data are artificial.
Further Readings 283

4. Run the k-means analysis again from the example application (Cars.sav, 8 Web
Appendix ! Chap. 9). Use a factor analysis considering all nine variables and
perform a cluster analysis on the resulting factor scores (factor-cluster segmen-
tation). Interpret the results and compare these with the initial analysis.
5. Repeat the manual calculations of the hierarchical clustering procedure from the
beginning of the chapter, but use complete or average linkage as clustering
method. Compare the results with those of the single linkage method.
6. Make a list of the market segments to which you belong! What clustering
variables did you take into consideration when you placed yourself in those
segments?

Further Readings

Bottomley P, Nairn A (2004) Blinded by science: The managerial consequences of

inadequately validated cluster analysis solutions. Int J Mark Res 46(2):171187
In this article, the authors investigate if managers could distinguish between
cluster analysis outputs derived from real-world and random data. They show that
some managers feel able to assign meaning to random data devoid of a meaningful
structure, and even feel confident formulating entire marketing strategies from
cluster analysis solutions generated from such data. As such, the authors provide
a reminder of the importance of validating clustering solutions with caution.
Everitt BS, Landau S, Leese M (2001) Cluster analysis, 4th edn. Arnold, London
This book is comprehensive yet relatively non-mathematical, focusing on the
practical aspects of cluster analysis. The authors discuss classical approaches as
well as more recent methods such as finite mixture modeling and neural networks.
Journal of Classification. New York, NY: Springer, available at:
http://www.springer.com/statistics/statisticaltheoryandmethods/journal/357
If you are interested in the most recent advances in clustering techniques and
have a strong background in statistics, you should check out this journal. Among
the disciplines represented are statistics, psychology, biology, anthropology, arche-
ology, astronomy, business, marketing, and linguistics.
Punj G, Stewart DW (1983) Cluster analysis in marketing research: review and
suggestions for application. J Mark Res 20(2):134148
In this seminal article, the authors discuss several issues in applications of
cluster analysis and provide further theoretical discussion of the concepts and
rules of thumb that we included in this chapter.
Romesburg C (2004) Cluster analysis for researchers. Lulu Press, Morrisville, NC
Charles Romesburg nicely illustrates the most frequently used methods of
hierarchical cluster analysis for readers with limited backgrounds in mathematics
and statistics.
Wedel M, Kamakura WA (2000) Market segmentation: conceptual and methodo-
logical foundations, 2nd edn. Kluwer Academic, Boston, NE
This book is a clear, readable, and interesting presentation of applied market
segmentation techniques. The authors explain the theoretical concepts of recent
284 9 Cluster Analysis

analysis techniques and provide sample applications. Probably the most comprehen-
sive text in the market.

References

Andrews RL, Currim IS (2003) Recovering and profiling the true segmentation structure in
markets: an empirical investigation. Int J Res Mark 20(2):177192
Arabie P, Hubert L (1994) Cluster analysis in marketing research. In: Bagozzi RP (ed) Advanced
methods in marketing research. Blackwell, Cambridge, pp 160189
Bishop CM (2006) Pattern recognition and machine learning. Springer, Berlin
Calinski, T, Harabasz J (1974) A dendrite method for cluster analysis. Commun Stat Theory
Methods 3(1):127
Chiu T, Fang D, Chen J, Wang Y, Jeris C (2001) A robust and scalable clustering algorithm for
mixed type attributes in large database environment. In: Proceedings of the 7th ACM SIGKDD
international conference in knowledge discovery and data mining, Association for Computing
Machinery, San Francisco, CA, pp 263268
Dibb S (1999) Criteria guiding segmentation implementation: reviewing the evidence. J Strateg
Mark 7(2):107129
Dolnicar S (2003) Using cluster analysis for market segmentation typical misconceptions,
established methodological weaknesses and some recommendations for improvement.
Australas J Mark Res 11(2):512
Dolnicar S, Grun B (2009) Challenging factor-cluster segmentation. J Travel Res 47(1):6371
Dolnicar S, Lazarevski K (2009) Methodological reasons for the theory/practice divide in market
segmentation. J Mark Manage 25(34):357373
Formann AK (1984) Die Latent-Class-Analyse: Einf uhrung in die Theorie und Anwendung. Beltz,
Weinheim
Kaufman L, Rousseeuw PJ (2005) Finding groups in data. An introduction to cluster analysis.
Wiley, Hoboken, NY
Kohonen T (1982) Self-organized formation of topologically correct feature maps. Biol Cybern 43
(1):5969
Kotler P, Keller KL (2009) Marketing management, 13th edn. Pearson Prentice Hall, Upper
Saddle River, NJ
Larson JS, Bradlow ET, Fader PS (2005) An exploratory look at supermarket shopping paths.
Int J Res Mark 22(4):395414
McLachlan GJ, Peel D (2000) Finite mixture models. Wiley, New York, NY
Milligan GW, Cooper M (1985) An examination of procedures for determining the number of
clusters in a data set. Psychometrika 50(2):159179
Milligan GW, Cooper M (1988) A study of variable standardization. J Classification 5(2):181204
Moroko L, Uncles MD (2009) Employer branding and market segmentation. J Brand Manage
17(3):181196
Okazaki S (2006) What do we know about Mobile Internet Adopters? A Cluster Analysis. Inf
Manage 43(2):127141
Punji G Stewart DW (1983) Cluster analysis in marketing research: review and suggestions for
application. J Mark Res 20(2):134148
Sheppard A (1996) The sequence of factor analysis and cluster analysis: differences in segmenta-
tion and dimensionality through the use of raw and factor scores, tourism analysis. Tourism
Anal 1(Inaugural Volume):4957
Tonks DG (2009) Validity and the design of market segments. J Mark Manage 25(34):341356
Wedel M, Kamakura WA (2000) Market segmentation: conceptual and methodological founda-
tions, 2nd edn. Kluwer, Boston, NE
http://www.springer.com/978-3-642-12540-9