Advances in Consumer Research Volume 8, 1981 Pages 116-121
A MEANS-END MODEL FOR FACILITATING ANALYSES OF PRODUCT MARKETS BASED ON CONSUMER JUDGEMENT
Jonathan Gutman, University of Southern California
Day, Shocker, and Srivastava (1979) have discussed a number of customer-oriented approaches for identifying product markets. This paper presents a means-end chain model for facilitating analyses of product markets using consumer judgements. The model is based on categorization processes which consumers use to create sets of substitutable products.
Day, Shocker, and Srivastava (1979) have discussed a number of customer-oriented approaches for identifying product markets. The approaches include decision sequence analysis, perceptual mapping, and several variations for obtaining direct consumer judgements of substitutability (free response data, direct grouping into categories, products by uses analysis, and substitution-in-use analysis).
Although these approaches differ, they share a number of related aspects. These aspects can all be embodied in a model of consumer categorization processes that relates "ends" in terms of consumer values to "means," which often are the products consumers consume. Such a model can provide a basis for increasing our understanding of how consumers cognitively define competitive sets of products.
This paper will briefly outline and point out the critical aspect each of these approaches. Then the model will be presented. Lastly, some examples of relevant research will be presented that suggest how the model facilitates research on product markets.
APPROACHES FOR OBTAINING CONSUMER JUDGEMENT OF SUBSTITUTABILITY
Decision Sequence Analysis
Bettman (1971) and Haines (1974) have discussed the use of protocols of consumer decision making which indicate the sequence in which various criteria are employed in reaching a final choice. The data are typically gathered by asking shoppers to verbalize their train of thought as they make actual shopping decisions. The critical aspect of this approach is the specification of attributes used in the choice process and the sequence in which they come into play.
Perceptual maps are created by a variety of techniques that create geometric representations of consumer's perceptions of product markets. Such maps require that the relevant product market be already defined before the start of the study. However, it is possible to create maps for "different levels of product competition to explore competitive relations at the level of product types, variants, or brands" (Day, Shocker, and Srivastava 1979 p.14).
Consumer Judgements of Substitutability
The Free Response approach (Green, Wind, and Jain 1973) allows the respondent to decide how similar brands must be before they can be considered as substitutes. Consumers are presented with target brands and asked to free associate the names of similar or substitute brands. Frequency and order of mention are used as indications of similarity.
Direct grouping into categories (see Gutman 1979, Gutman 4 Reynolds 1979, and Bourgeois, Haines, and Sommers 1979) involves asking consumers to directly assign products to categories based on their judgments as to their degree of substitutability. Products-by-uses analysis (Stefflre 1979, Myers 4 Tauber 1977) involves ascertaining which usage situations are similar by virtue of their evoking the same benefits. Competitive products are determined by their being appropriate in equivalent usage situations. Substitutability-in-use analysis (Srivastava, Shocker, and Day 1977) is a refinement of the above approach brought about by a three-stage approach to refine the set of usage situations to be used.
Each of the above approaches requires that we know something different about consumers' cognitive structure. Decision sequence analysis requires that we understand the hierarchical ordering of elements in consumers' cognitive structure. The use of perceptual mapping requires that we be capable of representing the levels of inclusiveness at which products may be represented within a consumer's cognitive structure. The free-response approach emphasizes the individual nature of consumers' categorical systems. The direct grouping approaches emphasize the need to understand equivalence ranges (when are products substitutable or how similar do they have to be to be substitutable). And the products-by-uses analysis suggests that all of the above aspects vary by circumstance or situation.
DEVELOPMENT OF A MEANS-END CHAIN MODEL
It is possible to develop a model of consumers' cognitive structure that will provide for representation of all of these aspects. Such a model offers a theoretical and conceptual structure for connecting consumers' values to their behavior (Rokeach 1968, Howard 1977, Vinson, Scott, and Lamont 1977). Models of this type can be subsumed under the rubric of means-end chains. Means are objects (viz., products) or activities that people engage in (viz., running, reading, etc.). Ends are valued states of being, such as happiness, security, accomplishment. A means-end chain is defined as consisting of an interconnected set of cognitive elements that allows a person to select objects or activities that enable him to achieve his desired end states.
Thus, the means-end chain permits us to focus on the basic aims consumers have in life while not losing sight of how these aims influence choices in specific situations. The process of categorization is proposed as the mechanism by which consumers organize thinking about specific product alternatives so as to create arrays of products which will be instrumental in helping them achieve their values. If this connection can be made, the bases consumers have for creating sets of substitutable products will be understood more thoroughly.
Categorization processes, then, represent ways people segment their environments into meaningful groups by creating equivalences among nonidentical stimuli. Thus, for example, the categories of paper towels, toothpaste, or coffee have meaning to people because the objects in these categories all have something in common.
This process of categorization takes place at many levels of inclusiveness, which refers to the degree of similarity among products in a category. Brands of paper towels grouped together would represent a category of low inclusiveness. If consumers distinguish between plain paper towels and those with a pattern, two product categories of lower inclusiveness would be created. If paper towels were grouped with paper napkins, cups, and tissues, a more inclusive category of "paper products" would be created.
Categorization as a Basis for Means-End Chain
Groups or categories of products (product classes, for example) have to be related systematically to the higher-level ends if the "chain" is to serve its instrumental purpose of enabling the person to achieve his or her values. This means that values have to be translated from their context at the more inclusive levels of the chain to the less inclusive levels where products are categorized into product classes, thus establishing sets of products which will be competitive (alternative satisfiers of the same ends).
The basis of this clustering is mutually determined by the object properties and value-determined criteria for goal instrumentality. This categorization process takes place at each level of the means-end chain as categories of greater inclusiveness are formed at higher and higher-levels of the means-end chain. For example, some consumers may treat the entire category of soft drinks as a single category of sugary substances that ought not to be consumed, while others regard them as tasty and delicious. Others may make a distinction between fruit drinks and carbonated drinks. Still others may make a distinction between drinks that are "fun to drink" (soft drinks) and those that are "good for you." Most consumers will be in agreement as to the identity of products--what is a soft drink and what isn't. Differences in consumer categorization are less at low levels of inclusiveness (where assignment to categories is based on product attributes) than at high levels of inclusiveness (where categories are based on function or value instrumentality). Consumers group products in different categories, depending on which features they emphasize and which features they ignore.
These categories are created by the application of a cognitive mechanism called a "distinction." Distinctions are dichotomies that represent the end points of dimensions along which products may be compared. If two products can be compared using a hot/cold distinction, this indicates that they may be compared by temperature although they may differ with respect to temperature. An array of products can be created by applying this distinction to any of a limitless number of products. However, many products may not be comparable on this distinction. If a consumer doesn't think of a chain saw, a dry cleaner, or a box of paperclips as being either hot or cold, then none of these objects would be part of the array. Other products, such as orange juice, coffee, tea, milk, water, corn flakes, oatmeal, etc., could be compared on this basis.
The resulting array can contain as many categories as products or just one category (if all products were thought to be the same temperature). The essential point is that the distinction being dichotomous has nothing to do with the nature of the array of products created by its application. A distinction is not an array of products--it is a dichotomous cognitive element in which the two parts of the dichotomy have some meaningful relation to one another. What is meaningful is a personal decision for each consumer to make.
Description of Means-End Chain Model
Two major events occur in the development of a means-end chain. One is the crystallization of values to the point where they become capable of articulation into goals or valued states. This event occurs at the top of the chain. At the bottom of the chain, individual products have to be arrayed into product classes to limit the complexity of the choice situation to acceptable levels. Products have to be given meaning in terms of the internal coding system established by the individual for making sense of the world (Olson and Muderrisoglu 1978).
These two issues define the basis for the model shown in Figure 1. The figure is divided into three parts: the means-ends chain, matrices, and inputs and outputs. The figure shows that the means-end chain is comprised of three levels of distinctions--values, benefits, and grouping. The focus of distinctions at the values and benefits levels is on what products can do for the consumer. Distinctions at the grouping level focus on properties of the product or evaluations of properties of products. Inputs to the system are products and situations; outputs are products chosen for final consideration. Matrices are intersections between inputs and means-end chain distinctions or between outputs from such matrices.
MEANS-END CHAIN MODEL
Situations encountered by consumers are assessed in terms of values to be sought in each. This occurs in the situations-by-values matrix and results in a set of relevant values emerging that will guide behavior in a given situation. These relevant values are arrayed in a matrix with potential benefits. Analysis of the instrumentality of the benefits for achieving relevant values results in a set of relevant benefits which guide the choice of product alternatives.
At the bottom of the figure, grouping level distinctions form a hierarchy wherein the identity and evaluation of product alternatives are determined at different levels of inclusiveness.
At the lowest levels of inclusiveness, products are grouped together in accordance with Howard's notion of semantic structure or Lunn's (1972) organization of ways of satisfying consumer needs by product types, product variants, and brands. Product types satisfy different needs beyond generic needs (tooth paste vs. tooth power; hot vs. cold cereal). Product variants are variations within types (with or without fluoride, presweetened vs. unsweetened). And brands represent minor differences within variants. Thus, brands as specific alternatives are at the bottom of the hierarchy; those are subsumed by product variants, and product variants are grouped by product type.
At the upper levels of the hierarchy, evaluative categories subsume identity categories. Thus, for cereal, nutritious vs. unnutritious or tastes good vs. doesn't taste good might subsume distinctions based more on physical properties of cereals. The distinctions at the top of the grouping hierarchy are implied by, or predicted from, the distinctions at the bottom of the hierarchy. Relevant benefit distinctions for the situation are applied to the hierarchy of grouping distinctions. This determines the level in the hierarchy at which a set of products would be selected for further processing--perhaps the level at which the relevant product market would be defined for that consumer. Products at this level of grouping would themselves be compared in the "relevant benefit distinctions by products" matrix. This comparison would result in a product to be chosen for that situation or a set of products from which a choice would have to be made on a nonvalues-related basis.
Values, Situations, and Value-Situation Matrices
Consumers obviously encounter many potential product-use situations. There are many ways of defining and describing such situational variables (Balk 1975). But for the level of analysis referred to in this context we will follow Fennell (1978), who defines a situational unit as corresponding "to the activities and conditions for which products are crested and marketed, such as doing the laundry, feeding the dog, having a headache."
Consumers have valued ends they are trying to achieve. And consumption situations provide them with an opportunity to achieve these values. Each consumer learns over time which choices in a given situation are instrumental to achieving values and which are not. Thus, consumers learn to perceive certain situations as opportunities for achieving certain of their values. The situations by values matrix is a means by which the accumulated experience of the consumer can be represented. It illustrates the process by which a particular set of values comes to guide the selection of benefits sought in a given consumption situation. All of a consumer's values can't be achieved in one situation, and, frequently, achieving some values means not achieving other values. This point of view is consistent with the conclusion that situational influence is a pervasive factor in consumer behavior (Belk 1975, p. 161).
The grouping level is the level at which meaning is ascribed to products by grouping them together with other similar products and giving them a group label. Physical and surface properties of products are the bases for grouping products at the grouping level. As the most immediately perceived aspects of products, they serve to identify them in their most fundamental sense. Cars are to drive; refrigerators refrigerate; stoves cook; peas, green beans, carrots, etc., are all vegetables. It is possible to discriminate products more finely within each of these categories. There are many kinds of vegetables (root crops, green leafy, squash, etc.), many' ways of classifying cars, and many types of refrigerators.
Finer and finer discriminations within categories and broader and broader discriminations of subsuming categories (cars are a means of locomotion; vegetables are food; refrigerators are appliances) create product hierarchies. The number of levels in the hierarchy is a function of the knowledge of the consumer and his or her interest in maintaining a cognitive structure of varying amounts of complexity. Someone who doesn't like coffee would have a much flatter hierarchy of types of coffee than a coffee connoisseur would. Obviously, then, product markets would be defined differently for different types of consumers.
Distinctions at this level often have the form of logical negation, which means that one side of the dichotomy is "not" ("~" in Boolean algebra) what the other side "is." Soft drink vs. not a soft drink in effect creates one group of products--the other "group" representing all other objects not in the first group (e.g., fruit juice, water, elephants, etc.). Obviously, such a distinction could be applied to all objects, whereas "hot vs. cold" implies that objects be comparable along a temperature dimension. Distinctions in the form of logical negation create the simple identity groupings that are characteristic of the lower levels of the grouping hierarchy.
At low levels in the grouping hierarchy, distinctions serve to identify products; at higher levels within the grouping hierarchy, evaluative distinctions subsume those identity distinctions. At the lower levels, where identity is the focus, level of inclusiveness accounts for the level of a distinction within the hierarchy. Thus, in Figure 2, which shows one person's grouping hierarchy for breakfast beverages, soft drink subsumes colas, lemon-lime, and other flavors. At higher levels in the grouping hierarchy, the level of a distinction is determined by its ability to separate products into preferred and nonpreferred sets (see Tversky's (1972) theory of sequential choice). If "not easy to prepare" meant a product was not acceptable for use regardless of any other properties, it would be more functional to have it subsume "good for you"--"not good for you." However, if a consumer would be more likely not to consider a product if it were "not good for you," even if it were easier to prepare than "good for you," "not good for you" would be at a higher level in the hierarchy. The goal of categorizing products at the grouping level is to identify products and array them into sets based on their possession of attributes that imply the ability to provide desired benefits for the consumer.
GROUPING HIERARCHY FOR BREAKFAST BEVERAGES
The model presented above includes all the relevant aspects of the measures of consumer judgement that are used to measure product markets. It allows for the situational selection of relevant values that in turn select appropriate benefits to be sought in the consumption situation. Benefits, in turn, determine the level in the hierarchical structure at which products will be aggregated for further consideration. Thus, there are families of product markets at different levels of inclusiveness, each level of which is appropriate in different sets of circumstances. The criterion of substitutability of alternatives for defining product markets is consistent with the notion that values guide behavior and determine relevant objects for consideration in those situations.
The model also suggests that categorizing products at the values level leads to the development of a wider, more complete definition of relevant product markets than starling with the products themselves and trying to establish categories. Several studies the author has done bear on these issues. The first study (Gutman, 1978) involved eliciting distinctions from respondents and having them sort products into categories based on these distinctions. Twenty triads (sets of three cereals in this instance) were presented to each respondent, who was asked to indicate a way in which all three products could be compared. Then the respondent was to specify how two products are the same and different from the third. This allowed us to define the two poles or ends of the dimension defined by the distinction being made. Respondents were then asked to sort 23 brands of cereal into categories on the basis of their distinctions. They could use as many or few categories as desired.
Another group of respondents was asked to sort the cereals into categories without having been asked to elicit any distinctions they might have. After sorting the cereals, these respondents were asked to label the categories they had created.
The results indicated that it is difficult to get many distinctions from people. People don't have too many ways of thinking about products (at least for breakfast cereals which may not be too complex). The free sortings were different from the sortings based on the distinctions in that the former yielded sets of nominal categories rather than dimensional arrays. People tended to see products as organized wholes and didn't look for ways to link dissimilar products to each other.
This might suggest narrow product markets definitions where only highly similar products are considered as substitutes. If consumers are concrete and incapable of thinking abstractly about products, different types of products are not likely to be seen or thought of as alternatives.
It is possible to get consumers to operate at higher levels of inclusiveness, but they have to be "encouraged" to do so. Respondents can be asked to state their preferred poles for each distinction they elicit. Then they are asked why they prefer that pole (or that end of the dimension). The answer to the "why" question becomes the basis for asking another "why" question, and so on until the respondent can answer no further "why" questions. The results of such a procedure (referred to as laddering--see Bannister, (1970) are shown in Figure 3. The figure can be interpreted as providing a chain of implication between the distinctions at various levels in the hierarchy from means to ends.
HIGHER ORDER IMPLICATION CHAIN EXTENDING FROM "CRUNCHY"
Another study (Gutman and Reynolds 1979) focused on getting respondents to categorize products (drinks in this instance) at higher levels of inclusiveness. Here again, laddering was used to get distinctions at higher and higher levels of inclusiveness. A content analysis of distinctions at each level showed initial distinctions referred to the contents of the drinks or related to their immediately observable properties. Distinctions at higher levels referred to functions (what the drinks did to/for the person) or end states of being achieved through consumption of the drinks. An example of one respondent's hierarchy of responses to the laddering process is shown in Table l.
EXAMPLE OF LADDERING DERIVED FROM REPERTORY GRID PROCEDURE USING THE TRIAD: SOFT DRINK, FRUIT DRINK, FRUIT JUICE
The sorting data consisting of a "persons-by-drinks" matrix, with category membership nominally coded within served as the basis for generating the distance data for input to POLYCON, a multidimensional scaling program (see Rosenberg and Kim, 1975 for the procedure used). Three maps are shown in Figure 4. They show the product space in two dimensions for sorts based on distinctions at the first, third, and fifth levels (as obtained through the laddering procedure). The changes in groupings of drinks can be readily seen. The juices, fruit drinks, and soft drinks show movement across levels. In effect, these figures represent cognitive maps of product markets at different levels of inclusiveness. One has to get the consumer to consider the benefits derived and values satisfied to develop the proper context for arraying products into meaningful groups. Otherwise, the analogy is similar to that of the problem with SIC codes in that grouping is based on properties other than those related to consumption.
The last study (Gutman 1980) shows some evidence of preferred ranges for equivalence in sorting products into categories. Consumers differ in terms of the limits of what they consider similar. Some people will accept a slight basis of comparability as enough to group two products together. Other people would only do so if the two products were highly similar. This tendency may be related to actual product market limits as well as to cognitive mappings of such product markets. If people are narrow and concrete in terms of what they consider equivalent, then their cognitive view of the world will restrict their breadth of their solutions in terms of searching for means to satisfy their ends.
The paper has presented a means-end chain model for facilitating the analysis of product markets based on consumer judgement. The levels represented in the model-extending from values to benefits to product attributes-provide for an integrated set of product groupings at different levels of inclusiveness. This family of product groupings should permit managers to relate levels of market aggregation to the criteria used by consumers at those levels. Furthermore, operating at the values level can provide the most comprehensive view of the market in terms of learning about competition from outside traditional market boundaries.
The model has the advantage of letting the consumer decide the criteria to be used in grouping products. It may exhibit the linkages between physical aspects of products (which may offer a misleading view of product markets from a consumer perspective), benefits provided by these products, and values consumers are striving for in their consumption of these products.
Bannister, Donald (1970), "Science Through the Looking Glass." in D. Bannister (Editor), Perspectives in Personal Construct Theory, London: Academic Press, 47-62.
Belk, Russell (1975), "Situational Variables and Consumer Behavior," Journal of Consumer Research, 2, 157-164.
Bettman, James R. (1971), "The Structure of Consumer Choice Processes," Journal of Marketing Research, 8, 465-471.
Bourgeois, Jacques, Haines, George R., and Summers, Montrose S. (1979), "Defining An Industry." Paper presented to the TIMS/ORSA Special Interest Conference on Market Measurement and Analysis, Stanford, CA., March 26.
Day, George S., Shocker, Allan O., and Srivastava, Rajendra, K. (1979), "Consumer-Oriented Approaches to Identifying Product-Markets," Journal of Marketing, 43, 8-19.
Fennell, Geraldine (1978), "Consumers Perceptions of the Product-Use Situation," Journal of Marketing, 42, 38-47.
Green, Paul E., Wind, Yoram, and Jain, Arun K. (1973), "Analyzing Free Response Data in Marketing Research," Journal of Marketing Research, 10, 45-52.
Gutman, Jonathan (1978), "Uncovering the Distinctions People Make Versus The Use of Multi-Attribute Models: Do a Number of Little Truths Make Wisdom." In Proceedings of the 23rd Annual Conference, New York: Advertising Research Foundation.
Gutman, Jonathan (1980), "Equivalence Range in Categorization," in Advances in Consumer Research, Vol. VII, Olson, Jerry C. (Editor). Ann Arbor: Association for Consumer Research, 411-416.
Gutman, Jonathan and Reynolds, Thomas (1979), "An Investigation of the Levels of Cognitive Abstraction Utilized by Consumers," in Research Under the Sun, Eighney, J. (Editor), Chicago: American Marketing Association.
Haines, George H. (1974), "Process Models of Consumer Decision Marketing," in Buyer/Consumer Information Processing, Hughes, G. D. and Ray, M. L. (Editors), Chapel Hill, NC.: University of North Carolina Press.
Howard, John A. (1977), Consumer Behavior: Application and Theory, New York: McGraw rim Book Company.
Myers, James H. and Tauber, Edward (1977), Market Structure Analysis, Chicago: American Marketing Association.
Rokeach, Milton J. (1968), Beliefs, Attitudes, and Values, San Francisco: Jossey Bass.
Rosenberg, Symour and Lin, M. P. (1975), "The Method of Sorting as a Data-Gathering Procedure in Multivariate Research," Multivariate Behavioral Research, 10 (October), 489-502.
Srivastava, Rajendra, K., Shocker, Allan O., and Day, George S. (1978), "An Exploratory Study of Situational Effects on Product Market Definition," in Advances in Consumer Research, Vol. V, Hunt, R. Keith, (Editor), Ann Arbor: Association for Consumer Research, 32-38.
Stefflre, Volney (1979), "New Products: Organizational and Technical Problems and Opportunities," in Analytic Approaches to Product and Marketing Planning, Shocker, A. D. (Editor), Cambridge, MA: Marketing Science Institute, April Report 79-104, 415-480.
Tversky, Amos (1972), "Elimination by Aspects: A Theory of Choice," Psychological Review, 281-299.
Vinson, Donald E., Scott, Jerome E., and Lamont, L. H. (1977), "The Role of Personal Values in Marketing and Consumer Behavior," Journal of Marketing, 41 (April) 44-50.