• 
  Robinson (1938):-Market Segmentation involves viewing a heterogeneous market as a number of smaller homogeneous markets, in response to differing preferences, attributable to the desires of consumers for more precise satisfaction of their varying needs. He also emphasized that market segments arise from managers conceptualization of a structured and partitioned market, rather than empirical partitioning of the market on the basis of collected data on consumer characteristics. However, even if market can be partitioned into homogeneous segments, market segmentation will be useful only if effectiveness, efficiency and manageability of marketing activity are influenced substantially by discerning separate homogeneous groups of customers.
  
  6 criteria for segmentation concept:
  
  1. Identifiability –is the extent to which managers can recognize distinct groups and able to identify the customers in each segment on the basis of variables that can be easily measured.
  2. Substantiality- is satisfied if the targeted segments represent a large enough portion of the market to ensure the profitability of targeted marketing programs and closely connected to the marketing’s goals and cost structure.
  3. Acessibility –is the degree to which managers are able to reach the targeted segments through promotional or distributional effects.
  4.Responsiveness is critical for the effectiveness of any marketing segmentation strategy- because differentiated marketing mixtures will be effective only if each segment is homogeneous and unique in its response to them.
  5. Stability
  6. Actionability
  7. Measurability -the marketing firm should be able to identify and quantify the potential of each segment
  
  Segmentation – is essentially grouping task, for which a large variety of methods are available and have been used.
  
  The method employed in segmentation research can be classified in two ways:
  
  1. A-priori and post-hoc approaches.
  
  Segmentation is called a-priori –then the type and number of segments are determined in advance by researcher and the post-hoc –then the type and the number of segments are determined on the basis of the results of data analyses.
  For example, a manager may decide to segment the market for the fast-food chain by usage situation, into the breakfast, lunch and dinner sub-markets. Often, multiple segmentation bases are used to form the segments, and the segments obtained from each of those criteria are assessed by looking at the associations between grouping arising from the alternative bases. In the example the manager, might in addition segment the market by usage situation and location.

2. The second way of classifying segmentation approaches – is according to whether descriptive or predictive statistical methods are used.

• Descriptive methods analyze the association across a single set of segmentation bases, with no distinction between dependent and independent variables.
• Predictive methods analyze the association between two sets of variables, where one set consists of dependent variables to be explained/predicted by the set of independent variables

	a-priori	post-hoc
Descriptive
	Contingency tables	Clustering methods:
	Log-linear models	nonoverllapping
		overlapping
		fuzzy techniques
		ANN
		mixture models
Predictive
	Cross-tabulation	AID,CART
	Regression	Clusterwise regression
	Logit	ANN
	Discriminant analysis	mixture models

The most important segmentation methods:

• Cluster analysis
• Mixture regression
• Scaling

  Cross-tabulation - appears to have been the popular technique for the evaluation of bases in the earlier years of segmentation research. A problem of this method is that higher order interactions are difficult to detect and interpret in the tables. Green and Carmone suggested to use Log-linear models for this purpose.

The main objective of cross-tabulation and log-linear analyses in such cases is to test segments arising from alternative bases, and to predict one segmentation base from other bases. For example to compare heavy and regular users of a brand by lifestyle (e.g VALS) . Although they are not very effective, they continue to be used, especially in hybrid segmentation procedures that combine a-priory and post-hoc methods. Often it is desirable to obtain segments for two separate strata in a population defined a-priori, such as a business and private users, users and nonusers, new customers versus current customers.

In this case a two-stage approach is taken .

1. A sample is partitioned a priori on the basis of variable in the question.
2. Within each of the strata that arises, a post hoc, mostly clustering-based procedure is used.  

  Post-Hoc descriptive methods :

In lifestyle segmentation, for example, consumers are first measured along several demographic and psychographic characteristic; a clustering procedure is then applied to the data, to identify groups of consumers that are similar in terms of their values, activities, interests and opinions.

Clustering methods:

• Non-overlapping – each subject belongs to a single segment only.
• Overlapping  - a subject may belong to multiple segments
• Fuzzy – the hard membership or nonmembership of a subject is replaced by the degree of membership in each segment. For example a subject may belong partly to segment A (0.6) and B (0.3) and C (0.1)

  Non-overlapping clustering – are the most used in marketing research.

  Two types of Non-overlapping clustering are distinguished :

• Hierarchical -start with single – subject clusters and link cluster in successive stages. Two consumers who are placed in the same group at an early stage of the process will remain in the same segment up to the final clustering solution. Several hierarchical methods can be distinguished : single linkage, compete linkage and minimum variance linkage (Ward’s method)
• Nonhierarchical –methods start from a random (initial) division of the subjects  into a predetermined number of clusters and reassign subjects to a clusters until a certain criterion is optimized. Two consumers who are placed in the same group at an early stage may end up in different segments. A large number of nonhierarchical methods is available; k-means is the best known and most widely used.

 Several extensions have been proposed for both hierarchical and nonhierarchical clustering methods.

 De Soete Desarbo and Clark extended the k –means algorithm in a similar way. The method also allows for the analyses of several groups of a-priory weighted variables.

  Benefits of marketing segmentation:

Marketing segmentation should result in benefits for both the marketinf firm and its customers.

If no such benefits accrue to either party, then the segmentation exercise is a meaningless waste of time.

Possible benefits:

1. Greater sales and profitability
2. Allow the producers to design products and marketing appeals that are more "finely turned " to the needs of the market.
3. Greater consumer satisfaction
4. Focus on sub-markets with the greatest potential
5. Allows greater product differintiation and variety as firms seek further market opportunities by developing new segments
6. Better competitive position for existing brands
   

• Why do we segment?
• When it is mostly important?

A Definition:

 Market Segmentation is concerned with individual or intergroup differences in response to marketing mix variables. The managerial presumption is that if these response differences exist, can be identified, are reasonably stable over time and the segments can be efficiently reached the firm may increase its sales and profits beyond those obtained by assuming market homogeneity.

Du-Pont’s Definition:

 “A group of customers anywhere along the distribution chain who have common needs and values - who will respond similarly to our offerings and who are large enough to be strategically important to our business.”

Cluster Analysis:

 C.A is a set of techniques which Classify, based on observed characteristics, an heterogeneous aggregate of people, objects or variables, into more homogeneous groups.C.A is useful to identify market segments, competitors in market structure analysis, matched cities in test market etc.

Cluster analysis is widely used in market research when working with multivariate data from surveys and test panels. Market researchers use cluster analysis to partition the general population of consumers into market segments and to better understand the relationships between different groups of consumers/potential customers.

• Segmenting the market and determining target markets
• Product positioning
• New product development
• Selecting test markets (see : experimental techniques)

Hierarchical clusters  

are nested tree-like structures, and usually reflect a development sequence. Each person, product or occasion is treated as a separate and distinct cluster to begin with. They are merged using an appropriate similarity measure until every object belongs to a large cluster. It may help for “seeing the market structure” in terms of brands. For a set of 100 persons the H.C.A will start with 100 clusters, each containing 1 object and finish with 1 cluster.

Non-hierarchical methods  

cluster a data set into a single classification of a number of clusters fewer than the number of objects. The number of the cluster may be specified a-priori or determined as part of the clustering method