The Concept and Measurement of a Hierarchy of Product Characteristics

Loren V. Geistfeld, Purdue University
George B. Sproles, Purdue University
Suzanne B. Badenhop, Purdue University
ABSTRACT - The concept of product characteristics has generally been approached on an ad hoc basis, with only limited efforts directed toward conceptualizing levels and types of product characteristics. This paper presents a preliminary unifying theory of product characteristics, termed the hierarchy of product characteristics, which is a first step in eliminating this deficiency in theoretical development. The hierarchy identifies three levels of product characteristics, each of which is functionally related to the others. An exploratory analysis then examines one application of the hierarchical concept in identifying product characteristics which have the greatest potential significance to consumer decision-making. The example illustrates that characteristics at all three levels may be needed in empirical analysis of consumer preferences and choices based on product characteristics. Implications for future applications of the hierarchical concept are also discussed.
[ to cite ]:
Loren V. Geistfeld, George B. Sproles, and Suzanne B. Badenhop (1977) ,"The Concept and Measurement of a Hierarchy of Product Characteristics", in NA - Advances in Consumer Research Volume 04, eds. William D. Perreault, Jr., Atlanta, GA : Association for Consumer Research, Pages: 302-307.

Advances in Consumer Research Volume 4, 1977   Pages 302-307


Loren V. Geistfeld, Purdue University

George B. Sproles, Purdue University

Suzanne B. Badenhop, Purdue University

[The authors wish to acknowledge financial support for this investigation from the Institute for Consumer and Family Studies, Purdue University.]


The concept of product characteristics has generally been approached on an ad hoc basis, with only limited efforts directed toward conceptualizing levels and types of product characteristics. This paper presents a preliminary unifying theory of product characteristics, termed the hierarchy of product characteristics, which is a first step in eliminating this deficiency in theoretical development. The hierarchy identifies three levels of product characteristics, each of which is functionally related to the others. An exploratory analysis then examines one application of the hierarchical concept in identifying product characteristics which have the greatest potential significance to consumer decision-making. The example illustrates that characteristics at all three levels may be needed in empirical analysis of consumer preferences and choices based on product characteristics. Implications for future applications of the hierarchical concept are also discussed.


Historically, the concept of product characteristics (often synonymously termed product attributes) has been approached intuitively, without any precise definition as to what a product characteristic is and what it is not. However, there are at least two important and sharply contrasting streams of thought which have emerged among marketing researchers and economists. For example, marketing researchers have implicitly defined product characteristics in terms of consumers' subjective judgments (perceptions, attitudes) directed toward specific features possessed by a product (Wilkie and Pessemier, 1973). This perspective has been extensively applied in recent research on consumers' preferences using multi-attribute attitude models. Typically, such research has involved measuring the importance or "salience" of product characteristics as perceived by a consumer, and the consumer's belief that a particular brand would possess some desirable magnitude (high to low) of each characteristic. "Characteristics" (attributes) have ranged from such specific purchasing criteria as price and brand name, to more abstract and subjective perceptions of features such as "convenience," "durability," "sex appeal," etc. From this perspective, nearly anything that a consumer perceives about a product may qualify as a product characteristic.

A differing stream of thought has recently emerged among economists, as exemplified by the works of Lancaster (1971), Cowling and Cubbin (1971), and Maynes (1976). [For a complete discussion of the use of product characteristic in the economics literature, see Geistfeld (1974).] Lancaster, in what has been termed a new economic theory of consumer behavior (Ratchford, 1975), has explicitly defined product characteristics as those properties of a product which are relevant to consumer choice, quantitative, objectively measurable, and universal (Lancaster, 1971, pp. 6, 15, 18). To exemplify this perspective, we might identify several product characteristics which would be possessed by an orange: diameter, weight, skin thickness, ratio of juice weight to weight of solid matter, sugar content, and vitamin C content. Each of these is objectively measurable (quantitative), universal to oranges, and relevant to consumer choices; therefore, each would qualify as a product characteristic. However, the taste of the resulting juice would not be a product characteristic by Lancaster's definition, since it would not be objectively measurable.

Cowling and Cubbin (1971), using the hedonic approach to product quality, [The hedonic approach to product quality is an empirical technique based on multiple regression. The independent variables are product characteristics, the dependent variable is price, and the estimated coefficients are interpreted as implicit characteristic prices. For more information see Griliches (1971).] argue that there is a functional relationship between those services of a product which a consumer demands and the characteristics of a product. Their definition of a product characteristic is not explicit, but from context it appears they mean such product features as brakes horsepower, passenger area, fuel consumption, length, power-assisted brakes, and four-forward gears. These specific features would then "produce" what the consumer wants--speed, convenience, capacity and comfort (Cowling and Cubbin, 1971, pp. 383-384). This also implies at least two levels of product characteristics: 1) a set of basic objectively measurable product characteristics, and 2) an abstraction of those basic characteristics to a higher level of "performance" or service characteristics.

Maynes (1976a, pp. 52-53) has used the concept of product characteristic to reflect Cowling and Cubbins's "product services which a consumer demands." Maynes defines what he calls a service characteristic to be that "basic factor which gives rise to utility." In this context product features such as durability, beauty and safety are considered product characteristics while the product features which give rise to these elements are not considered product characteristics.

Even though there is variation as to how different disciplinarians view product characteristics, there does appear to be a common thread which runs through all of these perspectives. It is this thread which is now explored, focusing on the concept of a hierarchy of product characteristics.


The problem of conceptualizing levels of characteristics has been alluded to by Maynes (1976a). He notes that "/-a-/ vexatious problem in the identification of characteristics is the determination of the optimal level of abstraction...consider the performance of a soprano...should we consider 'beauty' as a characteristic? Or, assuming that beauty in a vocal performance has as its components such things as 'color' and 'range', should we consider each of these components of beauty as separate characteristics?" (Maynes, 1976a, p. 53)

This clearly points to a problem of dimensionality (level of "abstraction") among product characteristics: in Maynes' example "beauty" is a major dimension (perhaps existing as a subjective characteristic), but beauty also has more objectively measurable sub-dimensions (sub-characteristics) of color and range. However, even these sub-dimensions can he further subdivided, suggesting the existence of yet another level of product characteristics. [The issue of dimensionality has also been raised in the use of multi-attribute attitude models. However, in that model the issue centers on avoidance of "double counting" of interdependent characteristics included in the model rather than the more theoretical issue of hierarchically-ordered product characteristics.]

Based on the various definitions of product characteristic and the concept of dimensionality (interrelationships) among product characteristics, we may define levels of characteristics which are functionally related to one another. This, in essence, results in the hierarchy of product characteristics.

The Concept of the Hierarchy

The concept of a hierarchy of product characteristics depends on three fundamental ideas: definition of a product characteristic, dimensionality, and measurability. We define a product characteristic as any feature of a product which is intrinsic to the product and which, directly or indirectly, influences a consumer's evaluation of a specific product variety. [A product variety is a product/brand/model combination. For example, washing machine/Maytag/ A606 is a product variety.] The intrinsic aspect of the definition precludes brand, price, guarantees/warranties, quality marks, instruction books, etc. from being considered product characteristics since they are extrinsic to the product variety. This is not to say that these elements are not important to the purchase decision, but that they do not belong in the characteristic domain of the purchase decision; however, they may be identified as a separate set of extrinsic purchasing criteria. But, the true product characteristic is one based on an identifiable physical (intrinsic) feature (i.e. material, type of construction), or an abstraction of features to the service performed for the consumer (i.e. durability).

The second concept is that of dimensionality. Any characteristic is said to be multi-dimensional if it is functionally related to other product characteristics, which themselves may be either multidimensional or uni-dimensional. Measurability, the third concept, is the extent to which a standard exists or can be developed for quantitatively measuring how much of a specific characteristic is possessed by a specific product variety.

With these considerations in mind, a hierarchy of product characteristics including three fundamental levels, "A", "B" and "C", may be formulated. "A" level characteristics are abstract, multi-dimensional characteristics which are difficult to empirically measure (as implied by Maynes). They are multidimensional in that they are dependent upon lower level characteristics ("B" and "C" levels) and are difficult to measure since they are abstract and tend to reflect the "overall character" of a product. For example, the safety of a product might be considered an "A" level characteristic. The multi-dimensionality of safety is obvious, and there is no single unit of measure even though consumers may perceive a specific product variety as being "safe" or "unsafe."

"B" and "C" level product characteristics, the second and third levels respectively, are closely aligned with Lancaster's conception of a product characteristic. "B" level characteristics are often multidimensional characteristics which can be empirically measured. Generally, they are arguments in the functional relationships determining "A" level characteristics. "B" level characteristics are the specific properties and services which are determined by "C" level characteristics and affect the overall desirability of a product variety. An example of a "B" level characteristic might be warmth of a blanket which could, in principle, be measured using any technique appropriate to determining insulating ability; however, warmth depends on such things as weave and fiber.

"C" level characteristics are often uni-dimensional and measurable. These characteristics are functionally related to "B" level characteristics, and generally include features related to composition and construction of the product. Examples of "C" level characteristics for a blanket would be weave and fiber content. Both of these are measurable in a nominal sense rather than along a continuum as are "A" and "B" level characteristics.

An Illustration of the Hierarchy

To facilitate a clearer understanding of what types of product characteristics are appropriate to each level, a detailed example is useful. In Tables 1-3, a listing of product characteristics for blankets and slow cookers are categorized according to the appropriate level. Even though the product characteristic lists may not be exhaustive, they do convey the flavor of the hierarchy. The lists also point to the multiplicity of product characteristics which might conceivably influence a consumer's choice. Table 4 shows how the three levels of characteristics are functionally related to one another for specific characteristics of a blanket. [Maynes (1976b, pp. 552-553) has developed a taxonomy of automobile service characteristics which is very similar to our Table 4. In this taxonomy "A" level characteristics and the appropriate "B" level characteristics associated with each "A" level characteristic are set forth.] For example, the safety of a blanket (A level) may be determined by its flammability and non-allergenic nature (B level), which in turn are determined by specific compositional features (C level - fiber content, weave, etc.). In principle, this exercise could be done for any consumer product with similar results.










The concept of the hierarchy suggests that product characteristics at all three levels are important determinants of consumer choice. However, there still remains the problem of validating and operationalizing this concept. The following discussion and research illustration presents one preliminary approach to this issue.

The first problem which must be faced involves the identification of product characteristics appropriate to a given product. Tables 1-3 present an example of characteristics for two products, blankets and slow cookers, arrayed by levels. As these tables indicate, the arrays of potentially relevant product characteristics, even for these relatively non-complex products, can be massive.

Generation of comprehensive lists of product characteristics may be accomplished by several methods: advice from product designers, engineers, or other "experts", open-ended interviews with consumers, focus group interviews with consumers, a researcher's professional judgment, etc. The characteristics included in Tables 1-3 were generated by a combination of advice from "experts", an open-ended questionnaire administered to student subjects, and the researcher's technical knowledge and a priori judgments of significant considerations in the design of these products. Use of these approaches will generally lead to identification of the most relevant (though perhaps not exhaustive) listing of product characteristics.

The analytical problem then becomes one of narrowing the listing of characteristics to a more manageable size. One approach to this involves measuring consumers' judgments of the importance, "salience", or significance of each characteristic. A simple method for accomplishing this is now discussed.

An exploratory survey of product characteristics was conducted using the characteristic lists generated for blankets and slow cookers. Subjects for the investigation were 151 upper-division undergraduate students enrolled in consumer-oriented courses of the School of Consumer and Family Sciences, Purdue University. Each subject completed a self-administered

6This is one reason why Maynes (1976b, p. 548) advocates consideration of only "A" level characteristics. He implicitly argues that the number of "A" level characteristics is much smaller than the factors ("B" and "C" level characteristics) which give rise to an "A" level characteristic survey which asked the following question: [For purposes of brevity, only the question and results for blankets are reported. Similar results were obtained in the analysis of slow cookers.] "Here is a list of features which a consumer might consider in purchasing a blanket. For each feature please circle the number most representing your opinion of the importance of each feature in choosing a blanket." The students then rated each feature on a scale on which zero indicated they would not consider the characteristic at all and the integer values between one and seven indicated not important to very important. The percentage of students indicating either a six or a seven (a rating of high importance for the characteristic) for each blanket characteristic is reported in Table 5.



Upon examination of Table 5, it is interesting to note the types of characteristics which are highly rated by the major proportion of respondents. Using 75% as the critical response rate, 12 characteristics (out of an original list of 54) are found to be highly important. Among these characteristics, all levels of the hierarchy are represented with two of the three "A" level characteristics, only one "C" level characteristic (fiber content), and nine "B" level characteristics being in the set.

The second point to note is the overall ordering of importance ratings for the complete array of characteristics. First, "A" level characteristics dominate the top of the importance ratings. Even though this may seem insignificant since only three "A" characteristics were included, it does support Maynes' idea that it is services or performance which are most relevant to consumer choice. Second, the middle of the list is dominated by the more specific "B" level characteristics. Finally, very specific "C" level characteristics are clustered at the lower end of the array. Tentatively, this would imply that a hierarchy of product characteristics identifies relative levels of each characteristic's importance to consumers in the aggregate; as one proceeds through the hierarchy from higher to lower level characteristics, consumer diversity (Lancaster, 1976) appears to become more important.


The hierarchical concept of product characteristics offers a wide range of possibilities for future investigations. Several of these are now briefly discussed.

The Issue of "Consumer Sophistication"

For an individual consumer to transform lower level characteristics into higher level characteristics requires some level of sophistication. For instance to correctly transform "C" level characteristics to "B" level characteristics often requires basic knowledge of physical or chemical properties of materials and combinations of materials. Similarly, to transform "B" level characteristics into "A" level characteristics is an identification problem, i.e. defining which "B" level characteristics are actually reflected in a given "A" level characteristic. The amount of knowledge implicit to these transformations can be staggering, and the greater the technical complexity of a product, the greater the requisite knowledge for the transformation. [Haynes (1976b, p. 548) has alluded to this problem when he suggests that one reason why one may wish to focus on "A" level characteristics is that the relationships between "A" level characteristics and lower level characteristics may be filled with nonlinearities and multiplicative relationships.]

Unsophisticated consumers, one might hypothesize, would use easily accessible and perhaps superficial measures to determine the extent to which a product variety possesses an "A" level characteristic. In comparison, a sophisticated consumer would look at more objective information to make an estimate of an "A" level characteristic based on What they know of lower level characteristics and an ability to transform them into higher level characteristics. For example, contrast two individuals looking for a new stereo console. Both consumers might want their purchase to have "good sound" (an "A" level characteristic); however, an unsophisticated consumer might just listen to a unit in a store to determine good sound, while a sophisticated consumer might consider product characteristics such as frequency range and distortion as well as "good sound."

In brief, this suggests that consumer sophistication or prior knowledge of a product will influence the level of characteristics ("A", "B", or "C") consumers use in the purchase decision. This leads to an hypothesis which can be explored. Given that sophisticated consumers will have a greater tendency to use lower level characteristics (they more readily recognize the relationship between all levels of characteristics), one would expect higher level characteristics to be considered important by more consumers than lower level characteristics for a given purchase decision. For example, one would expect nearly all consumers to look for durability in a blanket; however, only sophisticated consumers would appreciate the factors which give rise to durability such as resistance to abrasion, type of construction, fiber content, etc.

Consumer Information Programs

In the future, it is likely that consumer information programs will stress objective and/or performance characteristics which are most susceptible to standardization (Wilkie, no date, p. 7). However, if as Haynes (1976a) suggests, it is "A" level characteristics which consumers really desire, it is imperative that the relationship between the "A" level characteristics and lower level characteristics be studied to insure that appropriately defined informational contents are delivered to consumers.

Number of Characteristics

In the example presented above, an exploratory procedure was used to determine characteristics to be included in an analysis. An arbitrary decision was then made to consider the top Twelve characteristics as being the most relevant to consumer choice. Clearly, the selection of specific characteristics and number of characteristics remains an issue (Wilkie and Pessemier, 1973). However, since consumers may process information on between five and nine characteristics (Miller, 1956; Lancaster, 1976), the selection of a few "most important" characteristics for an investigation may be critical. However, this is dependent upon the nature of the investigation.

The Role of Purchasing Criteria

Earlier we argued that purchasing criteria (price, brand, package design, quality marks, instructional books, etc.) are not product characteristics since they are not intrinsic to the product. However, the relevance of these criteria to consumer choice cannot be dismissed.

Purchasing criteria might be conceived as "super-A" characteristics. It is well know that some consumers use price or brand name as surrogate indictors of durability, economy, convenience, fashionability and other "A" level characteristics. This suggests that if purchasing criteria are investigated in conjunction with characteristics, the relationship between purchasing criteria and characteristics should be examined, e.g. the relationship between price and durability.

Characteristics and "Efficiency of Choice"

The relationship between a consumer's "efficiency of choice" and specific characteristics used by consumers is, perhaps, the most interesting aspect of the hierarchy. For example, if "consumer I" uses knowledge of durability and "consumer II" uses knowledge of weave and fiber content when purchasing blankets, will this lead to a situation in which one consumer consistently selects a blanket of higher quality (as measured via product tests) than the other? If it does, the implications for consumer education are significant. We are currently exploring this issue.

Theoretical Development

AS we indicated at the beginning of this paper, our conception of the hierarchical concept is in the preliminary stages of development. At present, the primary contribution of the concept is in synthesizing into a single framework the levels, types and dimensionality of product characteristics. Using such a framework, the problem of characteristic specification can become more relevant to research on consumer choice, both at theoretical and applied levels.

Further theoretical refinements to the concept of the hierarchy should now be accomplished in the following areas:

1. Identification and definition of levels in the hierarchy, across categories of consumer products.

2. Specification of interrelationships between levels of the hierarchy.

3. Proposal of a universal set of "A" level characteristics such as durability, convenience, economy, fashionability, etc. which could form the basis for a theory of product characteristics at a general level of abstraction.

4. Further exploration of theoretical and practical relations between objectively measured characteristics, the consumer's factually-based knowledge of the characteristics, and the consumer's subjective judgment (attitudes, opinions, beliefs) which are formed for the characteristics. Ultimately, these dimensions must also be tied, by a theory of consumer decision-making, to the actual acts of consumer choice.


Keith Cowling and John Cubbin, "Price, Quality and Advertising Competition: An Econometric Investigation of the United Kingdom Car Market," Economica, 38 (November 1971) 378-394.

Loren V. Geistfeld, "A Technical Efficiency Approach to Consumer Decision Making," unpublished Ph.D. dissertation, University of Minnesota, 1974.

Zvi Griliches, Ed., Price Indexes and Quality Change, (Cambridge, Mass.: Harvard University Press, 1971).

Kelvin Lancaster, Consumer Demand, A New Approach, (New York: Columbia University Press, 1971).

Kelvin Lancaster, "Hierarchies in Goods-Characteristics Analysis," Advances in Consumer Research; Volume III, Proceedings of Association for Consumer Research Sixth Annual Conference, 1976, pp. 348-352.

E. Scott Haynes, Decision-Making for Consumers, An Introduction to Consumer Economics, (New York: Macmillan Publishing Company, 1976).

E. Scott Maynes, "The Concept and Measurement of Product Quality," in Nester E. Terteckyj, Ed., Household Production and Consumption, Studies in Income and Wealth, Volume 40, (New York: National Bureau of Economic Research, 1976).

G. Miller, "The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information," Psychological Review, 63 (1956), 81-97.

Brian T. Ratchford, "The New Economic Theory of Consumer Behavior: An Interpretive Essay," Journal of Consumer Research, 2 (September 1975) 65-77.

Hans B. Thorelli, Helmut Becket and Jack Engledow, The Information Seekers, (Cambridge, Mass.: Ballinger Publishing Company, 1975).

William L. Wilkie, How Consumers Use Product Information, An Assessment of Research In Relation to Public Policy Needs, Report prepared for the National Science Foundation, Research Application Directorate, Government Printing Office, Washington, D.C., no date.

William L. Wilkie and Edgar A. Pessemier, "Issues in Marketing's Use of Multi-Attribute Attitude Models," Journal of Marketing Research, 10 (November 1973) 428-441.