Marketing's Potential Contribution to Consumer Behavior Reserach: the Case of Diffusion Theory

Thomas S. Robertson, The Wharton School, University of Pennsylvania
ABSTRACT - The purpose of this paper is to show the value of marketing concepts and perspectives in consumer behavior research. Diffusion theory is used as a case example of how the integration of marketing concepts can enrich the consumer behavior research base and lead to more actionable findings.
[ to cite ]:
Thomas S. Robertson (1984) ,"Marketing's Potential Contribution to Consumer Behavior Reserach: the Case of Diffusion Theory", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 482-489.

Advances in Consumer Research Volume 11, 1984      Pages 482-489


Thomas S. Robertson, The Wharton School, University of Pennsylvania


The purpose of this paper is to show the value of marketing concepts and perspectives in consumer behavior research. Diffusion theory is used as a case example of how the integration of marketing concepts can enrich the consumer behavior research base and lead to more actionable findings.


The field of marketing generally has been viewed as a client or user of consumer behavior theory and research. There has been a lack of theoretical and methodological reciprocity and few expectations on the part of consumer behavior scholars that the marketing discipline could contribute in any substantive way to the conduct of consumer behavior research.

The typical work flow in the consumer behavior field is to borrow inputs from the behavioral sciences, to conduct research in a consumer environment, and then to suggest outputs of value to the practice of marketing or, sometimes, to the formulation of social policy. There are a number of problems with this sequence, some of which have been documented recently. For example, the theoretical inputs borrowed may not be appropriate to consumer analysis, since they may have been developed for other forms of behavior with distinguishable characteristics from consumer behavior. Kassarjian (1971) and Wells (1974) have made this particular point with regard to the use of personality theory in consumer research. Similarly, the borrowing of inputs places consumer behavior forever in a stepchild position and ignores the need to develop a comprehensive and systematic account of consumer actions in the marketplace (Jacoby 1976).

The perspective in this paper is to probe the interaction between consumer behavior research and marketing actions. The thesis is that much of consumer behavior research, when driven by borrowed behavioral theory, does not provide results of sufficient value to marketing practice. Such research simply does not take marketing's capabilities and limitations into account. Indeed, the present flow of the consumer behavior research effort ignores the potential contributions which the marketing discipline might offer for the conduct of more original consumer behavior research of greater value to the practice of marketing.

The purpose of this paper is to argue for more integration of marketing concepts in the conduct of consumer behavior research. The intent is to demonstrate how marketing inputs can enrich consumer behavior research and how the integration of marketing and consumer behavior at an earlier point in the research process can lead to more actionable findings. For purposes of illustration the diffusion paradigm will be used as the theoretical domain.


The diffusion literature encompasses a set of research traditions, although it is primarily identified with rural sociology (Rogers 1983). Such concepts as the S-shaped diffusion pattern, the categorization of adopter categories, the innovator/opinion leader distinction, and the specification of the adoption process have been widely accepted and adhered to in consumer research and are familiar topics in consumer behavior texts. These concepts also form the backbone of diffusion and adoption models (Bass 1969; Mahajan and Muller 1979; Pringle, Wilson and Brody 1983).

Consumer behavior scholars have contributed to the cumulative development of diffusion theory but the conceptual impact has been limited. For the most part the consumer diffusion literature reflects the same concepts as the general diffusion literature and researchers for the most part have been content to perform a translation and demonstration function of how diffusion theory can be used in consumer research. Indeed, this is a valuable first step but is insufficient.

The current malaise of consumer diffusion research reflects the lack of new insights and methods by consumer behavior scholars. Indeed, the very directions for future research on consumer diffusion have been identified by Rogers, rather than consumer behavior scholars. In a JCR review paper, Rogers posited three "biases" in diffusion research: (l) a lack of process orientation such that research has not tracked the individual's decision process over time; (2) a pro-innovation bias which assumes that all innovation is desirable; and (3) a lack of sociometric analysis (Rogers 1976). These observations have validity but do not derive from a framework relevant to consumer research.


Diffusion theory in consumer research has two central limitations. First, the general paradigm used is based on a set of assumptions that do not necessarily hold in consumer behavior. Secondly, the paradigm ignores the role of marketing actions and the corresponding impact on the diffusion process. A review of extant theory and assumptions together with needed conceptual developments is summarized in Table 1.

Restrictive Assumptions

The development of diffusion theory in rural sociology and in medical sociology is based on a set of assumptions which do not generally hold in consumer research. These assumptions are not necessarily explicit in the diffusion literature but underlie most of the research in the field.

As regards the concept of the innovation, for example, Katz (1961) compares the classic hybrid corn study of Ryan and Gross (1943) and the medical diffusion study of Coleman, Katz and Menzel (1966) and notes that both innovations:" .. came highly recommended by competent scientific authority ... (were) of central importance to the groups for whom they were intended ... (and) can be measured with a rational yardstick which enables users to see for themselves, more or less, whether the innovation serves better than its predecessor" (p. 71). These conditions would be the exceptions in research on the diffusion Of a consumer innovation.

The diffusion process itself is almost invariably conceptualized as an S-shaped (logistic) pattern. Although this pattern is reasonably prevalent in consumer research and in diffusion models at the product category level, a number of other patterns also occur, particularly exponential patterns (Cox 1967, Robertson 1971). It is essential to specify the conditions under which one or another pattern will be expected (Mahajan and Muller 1979). It will also be necessary to take into account the heterogeneity of the population such that different diffusion patterns may occur by market segment (Wind, Robertson and Fraser 1982; Eliashberg, Tapiero and Wind 1983).



The adoption process in the diffusion literature is viewed as a hierarchical sequence from awareness to knowledge, evaluation, trial and adoption. Although this may be a useful schematic for high involvement products, this "rational' or "learning" bias may be inappropriate for low involvement products where awareness may be sufficient to encourage trial. Other restrictive assumptions include the focus on individual level adoption decisions, despite the group nature of many consumption decisions; the focus on initial adoption only, ignoring issues of the width and depth of adoption; and the focus on a single innovation while ignoring its relationship to the consumption system or inventory.

Communication in the diffusion literature tends to de-emphasize advertising and personal selling and to emphasize "scientific" sources and change agents. The general assumption is that information is available and is unbiased. By contrast, most consumer innovations rely on mass media advertising, personal selling, and sales promotions for their dissemination. Marketing information is inherently biased and objective information may be unavailable.

Opinion Leadership and personal influence are accorded major attention in the traditional diffusion literature and a two-step flow of influence on a horizontal basis among members of a given social system is assumed. However, personal influence may not be important for all consumer innovations (particularly low involvement products). Additionally, influence flow may occur in multiple processes (not just a two-step) and individuals may belong to multiple social systems which are relevant to consumption behavior.

Finally, the appealing notions of innovators and other adopter categories are central to diffusion theory. The characteristics generally attributed to innovators (Rogers 1983), however, such as venturesomeness and greater social participation may not be particularly salient in consumer research, especially for low involvement products. The identification Or innovators in consumer research is, in fact, problematic and may not be worth the cost, except for major industrial innovations.

Marketing Impact

Diffusion theory assumes a certain passivity. A logistic diffusion pattern, for example, is to be "expected." Yet, marketing actions by their very nature are designed to change the diffusion process to the firm's advantage. Recent refinements of diffusion models recognize this objective (Horsky and Simon 1983; Simon and Sebastian 1982). Diffusion theory is quite incomplete unless it recognizes the proactive nature of marketing and competitive actions as suggested in Table 2.



The role of marketing actions begins with the very concept of the innovation. Marketing, in conjunction with Research & Development, designs and positions the innovation in order to achieve a certain perception. Many new products are designed in order to achieve a particular set of benefits based on consumer research input. It is not sufficient for consumer behavior researchers to study consumer reactions after product introduction and to ignore the before introduction research which created the product.

Marketing actions are,also important in influencing the pattern and speed of diffusion, as well as the diffusion process by market segment. Indeed, in most cases marketing actions are designed to achieve faster penetration in order to achieve a quicker return on investment, to block competition, and to establish a market franchise. Commitment of sizeable marketing expenditures in advertising, sampling programs, and distribution programs is likely to result in a diffusion function more similar to an exponential curve than a logistic curve if consumer learning needs are low and the product is not high on social visibility. Here we see the interaction of marketing and consumer behavior variables. Marketing actions also have a major bearing on diffusion by segment given allocation of resource decisions by segment. For many new appliances with "upscale" early diffusion, this segmentation is a result of both potential proclivity to innovation by this segment and marketing emphasis directed to this segment.

The speed of diffusion, of course, is related to the speed with which individual consumers proceed through the adopt- ion process. Marketing actions directed to the consumer have a considerable bearing on the speed of adoption (for example, advertising expenditures) and on the form of the adoption process. For example, marketing actions in sampling may short circuit the adoption process such that awareness and trial occur simultaneously followed by a repeat purchase or rejection decision. Marketing actions in personal selling are equally important in influencing the speed and form of organizational adoption.

The communication process for a consumer innovation may largely be marketer-controlled. There may be no readily available scientific authority or impartial source as in farming and medicine. Even if available, such impartial sources as Consumer RePorts do not reach the mass market. Diffusion for consumer products, therefore, depends critically on marketer-controlled information sources.

The operation of opinion leadership and personal influence, although very real, is influenced (sometimes heavily) by marketing efforts. Additionally, personal influence may not be relevant for many consumer products of a low involvement nature, Movies are a classic case of how marketing actions affect personal influence. Movies of dubious quality open to maximum initial distribution in an attempt to outdistance unfavorable personal influence. Alternatively, sneak previews and limited distribution seek to build personal influence for movies considered to be good.

Finally, marketing efforts affect who the innovators will be as a result of market segmentation decisions. However, there is also a question of how meaningful the innovator concept is for many consumer goods' categories where there is not much discrimination in the profile of the user versus the non-user, let alone the early user of a particular brand.


Diffusion theory has the potential to undertake major new research initiatives. The future directions should be based on explicitly taking the marketing perspective into account and recognizing the proactive role which marketing plays in new product development and diffusion. What follows are some ideas to guide consumer behavior scholarly research.

The Innovation

The existing concept of an innovation in the diffusion literature, as discussed, is limited. An initial need is simply for a broader framework for classifying innovations from continuous innovations (such as line extensions) to discontinuous innovations (new technologies). The familiar set of variables for assessing innovations (relative advantage. compatibility, complexity, trialability and observability) simply may not apply uniformly for the range of consumer product innovations. One contribution would be to specify an elaborated set of assessor variables and the types of innovations for which they are appropriate. Ostland (1974) took a step in this direction. It may also be that new psychometric methods offer unexplored ways for assessing new products.

In particular need of development are new concepts and methods for dealing with trivial continuous innovations. These represent a major component of new packaged goods innovations. A low involvement diffusion theory is needed to account for adoption and diffusion of continuous innovations.

In parallel, an enriched diffusion theory is needed in the high technology area -- especially when innovations are not an obvious extension of existing products. For a number of new products recently, including personal computers and videotext, the consumer lacks experience with the concept (Wilton and Pessemier 1981). Our research methods are embryonic in conducting consumer needs assessments when the consumer has to be educated about the concept, and perhaps even taught how to use it, before being able to venture an opinion as to its desirability. Methods of new product assessment, such as conjoint analysis and multidimensional scaling, are very much dependent on existing consumer product familiarity and experience and are most appropriate when studying reasonably minor product modifications.

The Diffusion Process

Consumer behavior and marketing researchers are well aware that the diffusion process does not follow a single form. Yet the overriding assumption is that a logistic pattern should be expected, and, in a certain sense, reacted to. Although there has been endless discussion in the marketing literature about the product life cycle, and although other functions, such as exponential and Gompertz curves have been found, future research could benefit from three considerations: (l) specification of expected diffusion patterns depending on a set of variables, (2) recognition of varying diffusion patterns by segment, and (3) acknowledgement of how marketing actions shape the diffusion process.

The diffusion pattern to be expected would seem to vary as a function of five overall variables. Their particular effects and interactions are in need of research. What follows are the categories of variables and their possible impact.

Innovation characteristics -- A major class of variables affecting takeoff point and speed of diffusion is the characteristics of the innovation. Rogers'(1983) set of five dimensions requires further testing in consumer research, including operationalization and factor analysis. For many consumer innovations the diffusion pattern seems to depend on learning requirements and social visibility.

Marketing variables -- As a general expectation we should expect more rapid diffusion given high marketing expenditures. However, the impact of short-run marketing expenditures seems to be much more pronounced for innovations with low learning requirements and low social visibility.

Competitive variables -- Diffusion for the product category will be more rapid as the level of competition increases. Competitors help undertake the consumer education function and make the product more visible and available.

Social contagion effort -- If the innovation is high on social visibility, a more pronounced social contagion effect leading to an S-shaped diffusion pattern should be expected, as in the Bass (l¦69) mode:

Environmental factors -- Government regulation may slow diffusion, as in expensive medical technology, whereas government funding may speed diffusion, as in the agricultural extension services.

Diffusion will also vary by market segment. This can be seen in contrasting patient vs. profession-oriented doctors in the classic Coleman, Katz and Menzel ( 1966) study More recently, it has been shown that different categories of hospital ( such as profit vs. nonprofit) have different diffusion rates for new technologies (Wind, Robertson and Fraser 1982).

It is necessary in conducting research to recognize that marketing actions are instrumental in affecting the shape and speed of the diffusion process. The segmentation decision for initial emphasis is critical to early diffusion. The commitment level of marketing resources behind the innovation affects rate of diffusion. And, marketing efforts may be important in guiding the diffusion process to avoid saturation and decline via product enhancement and product migration strategies.

The Adoption Process

Much as for the diffusion process, consumer behavior researchers recognize that multiple forms of the adoption process are possible (Ray 1973). However, research is needed in order to specify the expected pattern of adoption as a function of a particular set of variables. When can we expect the traditional hierarchy of effects to occur? When is the low involvement or dissonance/attribution model more appropriate?

A further need is for research which recognizes that most adoption involves multiple participants in the decision making process. Again, this is recognized but for the most part ignored in the research literature. There is only limited research on organizational adoption (Zaltman, Duncan and Holbek 1973) and even less which takes into account the family context of household adoption.

The concept of adoption itself has been used in a rather limited way: to refer to a single decision. Yet, for consumer product diffusion, adoption should be conceptual,zed more robustly. Obviously, for many consumer products, repeat purchase is the key to adoption. For other consumer products, however, it may be important to assess adoption as to both width and depth. By width is meant the number of people within the adoption unit who use the product or the number of different uses for the product. By depth is meant the amount of usage or the purchase of accompanying software ( as in personal computers). Further, diffusion may well be a function of both width and depth in adopting new units ( perhaps next generation) or additional units.

Finally, future research could contribute by focusing on how an innovation fits into existing consumption system and inventory patterns. A major barrier to diffusion occurs if consumers already have a similar product in inventory, such as a camera. Should the innovation then be positioned as more extreme in order to maximize the disparity with existing inventory? A further barrier to diffusion is if the innovation does not fit within the existing consumption system-- for example, if a new product requires the consumer to change how the entire consumption event is conducted

The Communication Process

The traditional diffusion literature has not accorded marketing-controlled communication sources much attention. Indeed, these sources may be less important for farm goods and drugs. However, for most consumer goods, advertising, personal selling and sales promotion will be the dominant means of consumer influence. Unlike farming and medicine where objective, scientific information may be available, most consumer goods may not have such information available. The communication environment, therefore, may be quite biased and information may be incomplete.

Research is needed which better integrates the advertising and personal selling literatures with diffusion theory. How does the consumer behave when information sources lack objectivity? How does the consumer behave in an environment of incomplete information or where information is difficult to access? What is the role of such traditional communication variables as source credibility, one versus two-sided arguments, conclusion drawing, etc. in the diffusion of new consumer products? What roles do these variables play in the diffusion of varying types of innovations?

Opinion Leadership

For many consumer products of a low involvement nature, opinion leadership and personal influence may simply not be relevant. The occurrence of personal influence would seem to depend on social visibility and whether objective product evaluation standards are available. An interesting research need is simply to document when personal influence occurs and as a function of what variable.

When personal influence is relevant, there remains the need to better document the interaction between marketer controlled communication sources and personal influence. Mass media advertising for many consumer products is too pervasive to allow much faith in a two-step model. What roles do these sources play, for what products, and what are the interactions of marketer versus personal sources?

Furthermore, we need an enriched model of personal influence. The two-step flow is still cited, despite its limitations (Weimann 1982). A more valid model of personal influence would have to recognize that influence may be verbal or visual, source or recipient-initiated, and that one or both parties may be influenced. Influence may also occur for different reasons, such as the normative versus informational distinction, and may go beyond the boundaries of the social system.

Finally, marketing actions may affect the process of personal influence--perhaps reducing the need for it or influencing what information is communicated in personal channels. An interesting research program might study the impact of marketing actions (as to level and type) on the personal influence process.

Adopter Categories

A normal distribution of adopter categories may be a useful heuristic. However, this again assumes an S-shaped diffusion ra.tern. Since the usual marketing objective is for more rapid diffusion, the S-shaped pattern may not be achieved. Thus, what categorization of adopter categories is most useful? How should the categorization vary for continuous versus discontinuous innovations? Is the innovator concept useful? Is the innovator for one consumer product likely to innovate for similar consumer products? Can marketing actions result in a desired innovator profile? In particular, it would be interesting to document how "innovator" characteristics can change as a function of market segmentation and product positioning decisions.


In this paper it has been argued that marketing may contribute to consumer behavior research leading to an enriched set of research directions of greater value to marketing practice. Until now the flow of research ideas has been rather one-sided from consumer behavior to marketing. For purposes of illustration diffusion theory has been used to demonstrate the potential value of a marketing perspective and marketing concepts.

Extant diffusion theory, as related to consumer products, has a built-in set of limitations, since it has been borrowed from dissimilar environments in farming and medicine. Consumer diffusion scholars to this point have not developed new concepts and methods of their own. With a strong infusion of a marketing perspective, a set of new research directions has been suggested. These directions have the potential to develop theory unique to the diffusion of consumer products and to provide implications of more value to the practice of marketing.


Bass, Frank M. (1969), "A New Product Growth Model For Consumer Durables," Management Science, 15, 215-227.

Coleman, James S., Katz, Elihu and Menzel, Herbert. (1966), Medical Innovation: A Diffusion Study, Indianapolis : The Bobbs-Merrill Company, Inc.

Cox, William E. Jr. (1967), "Product Life Cycles as Marketing Models," Journal of Business, 40, 375-384.

Eliashberg, Jehoshua, Tapiero, Charles S. and Wind, Yoram (1983), "Diffusion of New Products in Heterogeneous Populations: Incorporating Stochastic Coefficients," A working paper, The Wharton School, #83-018.

Horsky, Dan and Simon, Leonard S. (1983), "Advertising and the Diffusion of New Products," Marketing Science, 2 1-17

Jacoby, Jacob (1976), "ACR Presidential Address -- Consumer Behavior: Telling It Like It Is," in Advances in Consumer Research, Vol. 3., ed. Beverlee B. Anderson, 1-11.

Kassarjian, Harold H. (1971), "Personality and Consumer Behavior: A Review," Journal of Marketing Research, 8, 409-418.

Katz, Elihu (1961), "The Social Itinerary of Technical Change: Two Studies on the Diffusion of Innovation," Human Organization, 20, 70-82,

Mahajan, Vijay and Muller, Eitan. (1979), "Innovation Diffusion and New Product Growth Models in Marketing," Journal of Marketing, 43, 55-68.

Ostlund, Lyman E. (1974), "Perceived Innovation Attributes as Predictors of Innovativeness," Journal of Consumer Research, 1, 23-29.

Pringle, Lewis G., Wilson, R. Dale and Brody, Edward I. (1983), "NEWS: A Decision-Oriented Model for New Product Analysis and Forecasting," Marketing Science, 1, 1-29.

Ray, Michael L. (1973), "Marketing Communication and the Hierarchy of Effects," in Peter Clarke, ed., New Models for Mass Communication Research, Vol. 2, Beverly Hills: Sage Publications, 147-176.

Robertson, Thomas S. (1971), Innovative Behavior and Communication. New York: Holt. Rinehart & Winston.

Rogers, Everett M. (1976), "New Product Adoption and Diffusion," Journal of Consumer Research, 2, 290-301.

Rogers, Everett M. (1983), Diffusion of Innovations, 3rd Edition, New York: The Free Press.

Ryan, Bryce and Gross, Neal (1943), "The Diffusion of Hybrid Seed Corn in Two Iowa Communities," Rural Sociology, 8, 15-24.

Simon, Hermann and Sebastian, Karl-Heinz (1982), "Diffusion and Advertising: The German Telephone Campaign," A working paper, The Marketing Science Group of Germany, WP 0.9.

Weimann, Gabriel (1982), "On The Importance of Marginality One More Step Into The Two-Step Flow of Communication. American Sociological Review, 47, 764-773.

Wells, William D.,ed. (1974), Life Style and Psychographics, Chicago: American Marketing Association.

Wilton, Peter C. and Pessemier, Edgar A. (1981), "Forecasting the Ultimate Acceptance of an Innovation: The Effects of Information," Journal of Consumer Research, 8, 162-171.

Wind, Yoram, Robertson, Thomas S. and Fraser, Cynthia (1982), "Industrial Product Diffusion By Market Segment," Industrial Marketing Management, 11, 1-8.

Zaltman, Gerald, Duncan, Robert and Holbek, Jonny (1973), Innovations and Organizations, New York: Wiley.