Progress and Limitations of Social Marketing: a Review of Empirical Literature on the Consumption of Social Ideas.

Katryna Malafarina, University of Minnesota
Barbara Loken, University of Minnesota
[ to cite ]:
Katryna Malafarina and Barbara Loken (1993) ,"Progress and Limitations of Social Marketing: a Review of Empirical Literature on the Consumption of Social Ideas.", in NA - Advances in Consumer Research Volume 20, eds. Leigh McAlister and Michael L. Rothschild, Provo, UT : Association for Consumer Research, Pages: 397-404.

Advances in Consumer Research Volume 20, 1993      Pages 397-404


Katryna Malafarina, University of Minnesota

Barbara Loken, University of Minnesota

Social marketers have suggested limitations on the application of the marketing concept to the consumption of social issues. Several of these concerns were explored through a content analysis of empirical literature over the past decade. In general, findings suggest that although some criticisms were substantiated in the literature, others were not supported. Constraints that did occur in social marketing included negatively predisposed target segments and complex cost considerations. Concerns over lack of good secondary data, and unreliable or invalid measures; difficulties in identifying consumer behavior determinants, in identifying target segments, in defining effectiveness of measures and in evaluating programs; and the need to use personal, in-depth forms of communication were not substantiated. Study implications and limitations are discussed.


Social marketing, defined by Kotler and Zaltman (1971) as "the design, implementation, and control of programs calculated to influence the acceptability of social ideas and involving considerations of product planning, pricing, communications and marketing research" (p. 5), encompasses a broad range of consumer behavior topics from smoking cessation to AIDS prevention to environmental awareness. Because of the social nature of these topics, consumer and marketing researchers have raised important questions about whether the marketing of social ideas is compatible with the general marketing concept. In contrast to earlier conceptual work in social marketing, the present paper attempts to analyze the extent to which empirical articles on the consumption of social issues support prior claims made about the limitations of applying the marketing concept to social issues. It is our contention, which the data appear to support, that some of the assumed limitations of conducting social marketing programs and research occur for some, but certainly not most, social marketing programs and research.

The particular issues that we address in this paper, which are by no means comprehensive, can be divided into two general types of problems faced by social marketers that previous authors have discussed: (a) issues relating to the manner in which social marketing research and evaluation are conducted and interpreted, and (b) issues relating to the use of social marketing program elements such as pricing and communications.

Research and Evaluation Issues

With regard to the research and evaluation problems, several authors have argued that obtaining high quality data is more difficult to achieve in social marketing than in product marketing. Limitations on data quality result primarily from fewer financial resources available to social marketers but also can occur due to less continuity between social marketing research studies on a particular topic. Among the specific data quality concerns are (a) an absence of good secondary data available to social marketers, (b) lower reliability and validity of measures, (c) difficulty in identifying the determinants of the social consumer's behavior, and (d) difficulty in identifying through marketing research particular segments to target. Each of these concerns is elucidated below. See also Table 1.

First, good secondary data are often unavailable, since social marketers have limited syndicated services or panels that provide data on social marketing issues. Academic journals may be consulted, but these sources often may contain narrowly focussed research which is harder to generalize to another realm of social marketing (Bloom and Novelli 1981, Fox and Kotler 1980).

Second, problems with reliability and validity of primary data collected is assumed to frequently occur for social marketing studies, since people give more socially desired responses to sensitive issues such as smoking, charitable donations or sex (Bloom and Novelli 1981, Fox and Kotler 1980).

Third, researchers have argued that social marketing issues are more complex than non-social marketing issues, making it difficult for social marketers to isolate factors that affect an individual's behavior (Barach 1984, Bloom and Novelli 1981, Fox and Kotler 1980). Therefore, the determinants of the social consumer's behavior are more difficult to identify.

Fourth, attitudinal and behavioral data used to identify target segments are assumed to be less accurate when the issue pertains to social marketing. For example, it may be difficult to identify "users" and "nonusers", and differentiate groups from one another. Self-report measures may be misleading when measuring attitudes and behaviors pertaining to breast self-examination or contraceptive usage, and it may be impossible to use other behavioral measures such as observation in these circumstances (Bloom and Novelli 1981, Fox and Kotler 1980, Kotler and Zaltman 1971, Sheth and Frazier 1982).

Finally, it has been suggested that social marketers have difficulty evaluating their programs effectively. Determining the nature of the effectiveness variables may be difficult. Is the social marketing program designed to create awareness of an issue, alter consumers' knowledge or attitudes toward a social issue, change people's behavior, or save lives (e.g. Barach 1984, Bloom and Novelli 1981, Bloom and Ford 1979, Fox and Kotler 1980)? Further, the contribution of the marketing program in accomplishing certain objectives may be difficult to estimate (Barach 1984, Fox and Kotler 1980, Bloom and Novelli 1981, Bloom and Ford 1979). Part of the reasoning behind this latter train of thought is the assumption that social marketers have a limited ability to implement randomized experiments or quasi experiments. Often they are forced to use after only or before and after with no control group studies (Bloom and Novelli 1981, Fox and Kotler 1980).

Program Implementation Issues

In addition to the research and evaluation issues noted above, questions concerning implementation of program components have been raised (see also Table 1). For example, researchers have argued that key target segments in a social marketing program may be negatively predisposed to the offering, which, of course, results in the added difficulty of reaching these target segments (Barach 1984, Bloom and Novelli 1981). For example, a program that tries to promote safe sex through use of condoms will often have trouble reaching an audience of teenagers who may be embarrassed or unwilling to address the issue. Bloom and Novelli (1981) have suggested that this difficulty results from target markets who are frequently highly involved with their negative feelings, making them much more resistant to changing their views.

A second program implementation concern is that pricing strategies in social marketing often involve more than monetary costs; psychic, energy, and time costs are also involved (Andreasen 1984, Bloom and Novelli 1981). Because of this, social marketers may have less control over consumer costs (Barach 1984, Bloom and Novelli 1981, Fox and Kotler 1980, Kotler and Zaltman 1971, Rothschild 1979).



Finally, communications strategies may cause problems for social marketers. For example, the information to be communicated may be complex. Social marketers may need to educate consumers about certain types of behaviors and may also need to describe benefits of the behavior and recommendations for change (Bloom and Novelli 1981, Fox and Kotler 1980, Rothschild 1979). Due to the complexity of the social marketing issue, more personal forms of communication such as using health care professionals or other intermediaries may be necessary. Several articles in the literature have proposed suggestions on successfully implementing communications strategies (Kotler and Zaltman 1971, Fox and Kotler 1980, Sheth and Frazier 1982, Barach 1984, Andreasen 1984, Rothschild 1979).

Objectives of the Study

As indicated by the above discussion, concerns have been raised about the implementation, research, and evaluation of social marketing programs. The objective of the present research was to determine if the relevance of these issues is borne out in recent literature. Much anecdotal evidence exists on these issues, but no one has performed a content analysis to determine the extent to which these problems exist. In this paper, we attempt to remedy this omission and present the results of a content analysis of social marketing studies that have appeared in major marketing journals in recent years.


Empirical articles that have appeared in marketing journals since 1980 were included in the analysis. In particular, articles that appeared in Journal of Consumer Research (1980-1991), Journal of Marketing Research (1980-1991), Journal of Marketing (1980-1991), Advances in Consumer Research (1981-1991), and Journal of Public Policy and Marketing (1982-1991), were evaluated and selected on the basis of the article's title and abstract. Conceptual articles and empirical articles designed to test only theoretical issues (e.g. studying categorization theory in the context of a health care issue) were excluded from our sample. Further, Fox and Kotler's (1980) interpretation of the scope of social marketing was used to determine the nature of the issues to be included: "Social marketing should be distinguished from 'societal marketing' on the one hand and 'nonprofit organization marketing' on the other" (p. 25). Societal marketing refers to marketing's social responsibility and its impact on society. Nonprofit organization marketing includes the marketing of issues that are not relevant to social causes (e.g. marketing of political candidates or urban police departments). We used this definition in order to eliminate some of the multitude of articles that we encountered. Thus, articles pertaining to cause-related marketing and many nonprofit issues (e.g. political marketing) were excluded from our sample. Finally, since the number of articles pertaining to energy conservation and environmental issues was proportionally so much greater than other social issues, only 51.1% of the former articles were sampled for inclusion in analyses. [The analyses were conducted with and without the total number of energy conservation/environmental articles as well as the 51.1% reported. Interestingly, results were not changed substantively by omitting 48.9% of the articles.] The final sample included 76 articles, shown in Table 2.

Two judges evaluated 35% of the articles, based on content relating to the nine research, evaluation, and program implementation issues discussed previously and the scoring procedure shown in Table 1. Each article was scored as either (a) supporting the proposition, (b) providing negative support for the proposition, or (c) not addressing the issue. Analysis of these scores resulted in an inter-rater reliability of .844. While this reliability index is somewhat lower than standard levels, it is consistent with other reliability indices found for content of research articles in social marketing (cf. Ford et al 1990). The remaining articles were each evaluated by only one of the two judges.


Results, shown in Table 3, are reported by percentage of articles supporting each issue. In addition, results are presented as a function of specific content areas, including energy conservation and other environment issues, nutrition, smoking, alcohol and drug use, and, finally, other miscellaneous issues (including organ and the charitable donations, elderly social issues, shoplifting, gambling, homeless, patriotism and the performing arts).

Research and Evaluation Issues

The first issue addressed was whether empirical studies on social marketing had secondary data sources available that were relevant to hypothesis generation and/or interpretation of results (see Table 1). The vast majority of articles (89%) were able to cite sources of data (generally, previously published results of studies) that were relevant to the study at hand. Availability of secondary data was particularly evident for studies pertaining to smoking cessation, alcohol and drug use, nutrition, and energy and the environment, where, especially recently, an abundance of research findings has been accumulating. Not surprisingly, secondary sources of data were less prevalent for "miscellaneous" issues for which few studies have been published (e.g. compulsive behaviors like gambling, shoplifting, and studies of the homeless).

The second issue addressed whether, although people may be willing to be interviewed, they may give biased or socially desirable responses, resulting in lower reliability and validity of measures. While this issue was difficult to evaluate based solely on stated reports of issue sensitivity, many authors nevertheless provided indices of reliability and/or validity of measures or discussions of measure appropriateness. Results indicated that, in most cases (61% of the articles), authors reported measures of acceptable reliability and/or validity of their measures, or specifically mentioned the use of questions or measures that did not rely on socially desirable responses. For example, in some cases, self-report measures (e.g. of energy conservation) were verified through other data sources. Analyses of specific content domains suggest that, perhaps not surprisingly, articles pertaining to the sensitive issues of smoking, alcohol and drug use (41%), and other miscellaneous issues (38%, but especially AIDS prevention and use of contraceptive devices) yielded the lowest rates of reported reliability of measures.

The third issue tested was whether empirical studies supported the proposition that social marketing studies have difficulty in ascertaining the determinants of behavior due to the greater complexity of the issues. In examining whether each research article was able to identify specific determinants of the cognitive and behavioral criteria examined, we found overwhelming evidence against this proposition. In fact, all of the articles that investigated variables relevant to this issue were able to identify significant determinants of social attitudes and behavior. For example, age and cognitive style were found to be determinants of the processing of nutrition information (Cole and Gaeth 1990). It might be argued, of course, that had significant determinants of attitudes and behavior not been found, these studies may have gone unpublished. This question will be addressed later. Similarly, in examining the question of whether it was possible to differentiate user or preference groups by means of self-reports, observational data, or by whatever other method was used in the article, again, all of the studies that reported subgroup analyses reported significant differences between target groups. Again, these results suggest that the problem of obtaining useful data for identifying target segments may be overrated, or, alternatively, has been remedied in recent years.



Finally, with only two exceptions, results indicated most studies (97%) did not have the expected difficulty in defining effectiveness measures. Such measures ranged from memory for information on nutrition labels, to increased knowledge of AIDS-prevention measures, to decreased sales of alcoholic beverages over time. With regard to the contribution of findings to a social marketing program, most articles (68%) did not discuss results in the context of a particular social marketing program. For example, articles on nutrition and smoking often measured the effectiveness of warning labels, and included experimental designs that examined factors that were hypothesized to enhance warning effectiveness. However, of the articles that did report results of a social marketing program, most (92%) were able to determine whether the program element contributed significantly (whether successfully or unsuccessfully) to the program goals.

Program Implementation Issues

With regard to program implementation, we first examined, where relevant, whether any of the segmented targets was negatively predisposed to the program offerings. This proposition was also supported if the article specifically mentioned difficulty in reaching a particular segment that was "high risk" for the social issue in question. Of the articles for which this issue was relevant, the majority (64%) did report or imply that potential problems exist in reaching at least one high-risk segment that is negatively predisposed to the social marketing issue. The content areas in which these concerns were most dramatic pertained to smoking, alcohol and drug use. However, all content areas appeared to be affected by this problem to some extent.

The second program implementation issue examined was whether costs of changing the social consumer's behavior involved more than monetary costs. In over 90% of the articles, and across all content areas, non-monetary constraints on behavior (including psychic, energy, and/or time costs) were considered in the studies, thus supporting this proposition.

Finally, we examined whether, for studies that included a marketing communications component, more personal forms of communication (e.g. home visits, physician appointments) were used (and/or required) rather than mass media or direct mail, to convey information. We found that almost all (96%) of the articles that included a marketing communication component reported the use of mass media or direct mail, rather than a more personal form of communication to convey information. Furthermore, the information conveyed (whether about the amount of sugar content in food or a plea to reduce one's thermostat) was often in the form of a simplified, rather than a complex, message.




Research and Evaluation Issues

The results of our study suggest that although some assumed limitations do occur in social marketing programs, many of these limitations occur for relatively few of the social marketing issues addressed in empirical research. Most of the research and evaluation issues examined in the present analysis were not problematic for most social marketing research studies. While studies of newly researched social marketing topics (e.g. AIDS, the homeless) yielded a paucity of secondary data sources, studies of most topics (smoking cessation, alcohol use, nutrition, energy conservation) did not indicate these concerns. Furthermore, concerns about identifying determinants of attitude and behavior, or identifying differences between preference or user groups, were not supported by the literature. Similarly, authors of research articles were able to successfully define their effectiveness measures (whether awareness, attitudes, behavior, etc.) and, where relevant, determine whether research results could be interpreted as successfully or unsuccessfully contributing to the social marketing program.

Even with regard to the reliability and validity of measures, problems were not encountered as frequently as expected. In most cases, researchers sought creative solutions to problems of reliability or self-report measures that were anticipated for the issues examined. For example, measuring a respondent's self-reported views on energy conservation could be verified through electricity costs (Fritzsche 1981).

Given that the research and evaluation problems anticipated in the social marketing literature were not supported by the present literature, the question remains as to how the discrepancies between the conceptual and empirical literatures can be resolved. One possibility is that, while research problems exist for social marketers, researchers have striven to overcome these obstacles in their research efforts. A contributing factor for these discrepancies may be that most criticisms levied on social marketing were made in the early 1980's. Since then, and included in the present review, the lion's share of the research on social marketing has been reported in the literature. Thus, the earlier problems anticipated for social marketers were not borne out to the degree anticipated. On the other hand, a possible limitation to the present results is that only published empirical studies were examined or that information relevant to our analysis was not included in the article. One might argue that numerous unpublished findings would reveal a different set of conclusions. Published data may show greater reliability of measures, greater ability to identify target segments, and greater ability to evaluate program effectiveness. Furthermore, the criteria for determining support or nonsupport for a proposition tested in the present research may have been unduly restrictive, although, as noted earlier, the present analysis was not meant to be exhaustive in addressing all potential problems faced by social marketers.

Finally, a possible explanation for the discrepancy between earlier conceptual articles and the empirical data reported here is that problems for social marketers exist for only a subset of social marketing issues. In particular, novelty of the issue, the size of the program, or the nature of the resources (financial, expertise, etc.) available to the program, may be moderating variables that affect whether research problems exist. Clearly, novelty of the issue (e.g. AIDS, the homeless) may affect the extent to which secondary data are available to the social marketer. Size of the program, and financial and expert resources available, may affect whether the appropriate amount and level of research data are collected, and the resulting reliability and validity of data findings. Although these moderating variables may create problems for social marketers, such problems are not specific to the marketing of social ideas. Future research might address the role of such moderating variables in creating limitations for social marketing research and program evaluation.

Program Implementation Issues

In contrast to the research and evaluation issues, concerns examined about program implementation were often supported by the empirical literature. For most social marketing topics, empirical articles suggested that a key market segment exists that is negatively predisposed to messages directed toward them or toward changing their behavior. These concerns were particularly important for smoking, alcohol and drug use studies, as well as miscellaneous other health issues. For example, alcohol abuse programs target chronic alcoholics who are so highly involved with drinking that changing their behavior would require a drastic change in their lifestyles. Unlike marketing of products, where marketers target those consumers most interested in the product, social marketing often requires targeting messages to segments that are initially unfavorable toward the message.

Both monetary and nonmonetary costs were found to be examined in social marketing research, consistent with prior conceptual arguments about the complexity of social marketing issues. Thus, for example, in addition to considering the price of cigarettes, the consumer is also likely to consider the social, health, and other psychological costs in his or her decision to quit smoking (or to continue smoking). However, it is important to note that non-monetary costs are also important for virtually all marketing decisions, and are not unique to social marketing. For example, purchasing an automobile involves consideration of a complex array of factors, and, in fact, may be a more complex decision than deciding whether to turn down the thermostat in one's home an additional two degrees during the winter months. Moderating factors such as novelty of the decision or social marketing issue, the consumer's felt involvement in the decision, or, more generally, the extensiveness of information processing required for the decision, may be greater determinants of the complexity of the decision than whether the marketing decision is social or non-social (e.g. relating to a product). Future research might address the effects of such moderating influences.

The need for personal, educational, or in-depth, forms of communication in social marketing was not supported by the empirical literature. Generally, messages about social marketing issues (nutrition, smoking, energy conservation) were designed to be simple and easily conveyed to a wide audience and through mass media or direct mail. For example, hazards associated with smoking are presented on package labels (Bhalla and Lastovicka 1984). Hearts printed next to heart healthy items on restaurant menus, or colored tags next to items in grocery stores, have been used to convey to the consumer that these items are low-fat selections. Such simple messages were often quite effective. In fact, the extent to which a complex, "educational", message is required may be determined by the same moderating factors as noted previously. That is, the greater the novelty or innovativeness of the social marketing topic, or the greater the consumer involvement in the topic (cf. Rothschild 1979), the greater the need for high informational messages. In this sense, social marketing may yield more similarities than differences with non-social (e.g. product) marketing, where these moderating factors are likewise influential.

Implications and Conclusions for Research on Social Issues

The findings reported here have implications for future social research programs. First, social marketing is a progressing field that has perhaps overcome many of its early research difficulties, through enhanced database sources and improvements in the quality of research. Second, the issues involved in social marketing may be susceptible to the same concerns as found in product marketing. Both social and product marketers need to research and evaluate more carefully complex offerings. Third, there is a tendency for important high-risk target segments to be negatively predisposed to social marketing offerings, and these segments cannot be ignored in many programs. Social marketers need to be aware of this problem, and tailor their programs to attempt to alleviate discomfort or difficulties that consumers face in relation to an offering. Fourth, social marketers should continue to emphasize the importance of research and evaluation procedures to program organizers. By making the effort to obtain high quality data at all stages of the program, social marketers can more effectively target communications and segments of the population in need of the offering.

Finally, social marketers need to continue to adopt techniques that facilitate behavior change. For example, energy conservation may be facilitated by thermostat monitors that automatically reduce temperatures during the night. Consumers concerned about AIDS may be more likely to seek information if an anonymous source, such as a counselor obtained through a toll-free number, is available for this sensitive issue. Examples in the literature abound (e.g. Hutton and Ahtola 1991, Levy et al 1985, Alexander and McCullough 1981) in which social marketing programs completed their goals by facilitating the social consumer's behavior change. As consumer researchers expand their work in these areas, we can probably expect continued success at impacting social change. Such findings should also continue to have significant political ramifications to enhance positive social change.


Alexander, Katherine and James McCullough (1981), "Cultural Differences in Preventative Health Care Choice: A Study of Participation in a Cervical Cancer Screening Program Among Mexican-Americans," Advances in Consumer Research, 8, 617-632.

Andreasen, Alan R. (1984), "A Power Potential Approach to Middlemen: Strategies in Social Marketing," European Journal of Marketing, 18 (4), 56-71.

Barach, Jeffrey A. (1984), "Applying Marketing Principles to Social Causes," Business Horizons, (July-August) 65-69.

Bhalla, Gaurav and John L. Lastovicka (1984), "The Impact of Changing Cigarette Warning Message Content and Format," Advances in Consumer Research, 11, 305-310.

Bloom, Paul N. and Gary T. Ford (1979), "Evaluation of Consumer Education Programs," Journal of Consumer Research, 6 (December), 270-279.

Bloom, Paul N. and William D. Novelli (1981), "Problems and Challenges in Social Marketing," Journal of Marketing, 45 (Spring), 79-88.

Cole, Catherine A. and Gary J. Gaeth (1990), "Cognitive and Age-related Differences in the Ability to Use Nutritional Information in a Complex Environment," Journal of Marketing Research, 17, 175-184.

Ford, Gary T., Debra J. Ringold, and Martha Rogers (1990), "Cigarettes in the Popular Press, 1930-1960: Preliminary Research," Advances in Consumer Research, 17, 467-473.

Fox, Karen F.A. and Philip Kotler (1980), "The Marketing of Social Causes: The First 10 Years," Journal of Marketing, 44 (Fall), 24-33.

Fritzsche, David J. (1981), "An Analysis of Energy Consumption Patterns by Stage of Family Life Cycle," Journal of Marketing Research, 18, 227-232.

Hutton, R. Bruce and Olli T. Ahtola (1991), "Consumer Response to a Five-Year Campaign to Combat Air Pollution," Journal of Public Policy and Marketing, 10 (1), 242-256.

Kotler, Philip and Gerald Zaltman (1971), "Social Marketing: An Approach to Planned Social Change," Journal of Marketing, 35 (July), 3-12.

Levy, Alan S., Odonna Mathews, Marily Stephenson, Janet E. Tenney, and Raymond E. Schucker (1985), "The Impact of a Nutrition Information Program on Food Purchases," Journal of Public Policy and Marketing, 4, 1-13.

Rothschild, Michael (1979), "Marketing Communications in Nonbusiness Situations or Why it's so Hard to Sell Brotherhood Like Soap," Journal of Marketing, 43 (Spring), 11-20.

Sheth, Jagdish N. and Gary L. Frazier (1982), "A Model of Strategy Mix Choice for Planned Social Change," Journal of Marketing, 46 (Winter), 15-26.