Advances in Consumer Research Volume 17, 1990 Pages 494-500
THE FRAMING OF SALES PROMOTIONS: AN APPROACH TO CLASSIFICATION
William D. Diamond, University of Massachusetts
Robert R. Johnson, College of William and Mary
Research on sales promotions has been hindered by the absence of a theoretical approach to categorizing promotions. In this paper, the case is made for categorizing promotions by whether they are framed as gains or as reduced losses. A "cost/benefit" approach suggests that non-monetary promotions are most likely to be framed as gains whereas discounts and rebates will be framed as reduced losses. A perceptual map of promotions, derived from the similarity judgments of two groups of consumers, is consistent with this categorization. Attribute judgments show some heuristics subjects use to evaluate Promotions.
Promotions take many forms including coupons, rebates, reduced shelf price, and premiums. Behavioral theories tend to differentiate promotions on the basis of their magnitudes alone. A few of these theories have implications which could distinguish forms of promotion, but these implications have not been explored.
One reason for this is the absence of constructs providing rationales for differentiating promotions. As a result, behavioral researchers have tended to either confine empirical work in this area to one type of promotion at a time or select promotions atheoretically. The tendency toward very narrow categorization of promotions has been reflected in recent articles which deal with only a single type of promotion, such as couponing (e.g. Henderson 1985; Irons, Little, and Klein 1983; Narasimhan 1984; Neslin and Shoemaker 1983; Schindler and Rothaus 1985).
We argue here that a taxonomy based upon theory and consumer perceptual data will be useful in the application and extension of the behavioral theories of sales promotions. There is a close link between theory and practicality in this area.
The first section of this paper outlines a theoretical approach to the categorization of sales promotions. This approach is derived from prospect theory (Kahneman and Tversky 1979, 1984). Specifically, sales promotions can be categorized by whether they will be framed as gains or as reduced losses. This section concludes by discussing how behavioral theories apply differently to promotions framed in different ways.
The second part of the paper sets forth a perceptual map of promotions based upon the judgments of two consumer groups. This provides a taxonomy which is very similar to the theoretical categorization. By labelling the stimulus space with attribute judgments, we explore the heuristics that people use in evaluating types of Promotions.
A THEORETICAL CATEGORIZATION OF SALES PROMOTIONS
Prospect Theory and Framing
Recent research (Kahneman and Tversky 1984; Puto 1987; Thaler 1985) shows that when alternatives are presented as gains, subjects choose and judge them differently than when equivalent alternatives are presented as reduced losses. Thaler (1985) argued powerfully that promotions may be framed as gains or losses. The promotion framed as a gain will have benefits which are segregated from the original purchase price-whereas the promotion framed as a loss will be seen as merely reducing the initial purchase price. Thaler presented automobile rebates as a case where a promotion could be perceived as a separate gain rather than a mere reduction of the purchase price. In Thaler's language, a separate "mental account" may be established for the rebate. According to Thaler, an equivalent saving resulting from a sale offer should be seen as a reduced loss rather than a gain. Thaler used the shape of the prospect theory value function to deduce that a promotion framed as a segregated gain should be perceived as more valuable than one framed as a reduced loss. Although several other researchers (Klein and Oglethorpe 1987; Monroe and Chapman 1987; Puto 1987) have studied the effects of framing on consumer information processing, these researchers have not considered the implications of differences in the framing of sales promotions.
Predicting the Framing of Sales Promotions
There are several ways of hypothesizing whether a particular promotion will be framed as a segregated gain or as a reduced loss. One of Thaler's assumptions was that the physical or temporal separation of a rebate check from the price quotation leads to the framing of the rebate as a gain. He did not test this hypothesis.
Thaler and Johnson (1986) hypothesized that situations would be framed in accordance with a hedonic mechanism. Gains would be either segregated from or integrated with losses depending on which form would produce the most happiness for the subject. The general principle of hedonic framing led Thaler and Johnson to postulate several rules of framing. Because sales promotions are small gains presented in the context of a larger loss (the purchase price), the same rule should apply to all of them. This rule, known as the "silver lining effect" is to segregate small gains from larger losses. Accordingly, most promotions should be seen as gains rather than as reduced losses. This principle does not distinguish among types of sales promotions.
A third approach to predicting framing utilizes a "cost/benefit principle" (Beach and Mitchell 1978; Payne 1982). Beach and Mitchell posit that decision makers are motivated to choose the strategy which requires the least investment to achieve a satisfactory solution. From this perspective, the relatively unimportant problem of whether to take advantage of a sales promotion may not be allocated very much attention. The "cost/benefit" approach to predicting framing sterns from the amount of effort necessary to integrate gains with losses in different situations.
Tasks are more complex when different pieces of information about an alternative are in different units which may not be commensurable (Abelson and Levi 1985). Klein and Oglethorpe (1987) suggest that it should be more difficult to integrate multiple attributes of purchases if these attributes are in different metrics.
Adapting this to the domain of sales promotions, one might hypothesize that when promotions are in the same units as pricing information, they will be more easily integrated with the price. The promotion will then be framed as a reduced loss. Conversely, when promotions are in units other than money, they will be more difficult to integrate with the price. The consumer may not expend the effort required to integrate these noncommensurable promotions into the price paid, and such promotions should be more often considered as separate gains.
This reasoning presents theoretical grounds for the dichotomy between reduced-cost and value-added promotions recently discussed by Sawyer and Dickson (1984). Price-off promotions, including rebates, are expressed in the same units as price and are most likely to be framed as reduced losses. Value-added promotions, including either bonus amounts of the same product or other products as premiums, are in units other than price. These promotions are most likely to be framed as gains. Thus, the two classes of promotions should have fundamentally different properties if the consumer does not have the time, inclination, or ability to "recode" the nonmonetary benefits of a premium or purchase as a saving on the original product.
It is likely that other factors will influence the framing of sales promotions. For instance, continuity programs like frequent nier programs may be framed as gains because the promotion sets up very obviously separate "mental accounts" for the purchase price and for the promotional benefits. Similarly, promotions where money is contributed to a charity or worthy organization may be seen as gains because of the large segregation of the promotional benefit from the purchaser and the purchase price. The context of the promotion-whether other brands are being promoted--may well affect whether a particular offer is seen as a gain or a reduced loss. Finally, the various aspects of the presentation of a promotion may affect whether it is perceived as a gain or a reduced loss.
Implications of Framing for Behavioral Theories of Promotion
Once sales promotions are categorized in terms of how they are framed, it is possible to extend previous theories of sales promotions. Wt will examine the implications of this classification for two behavioral approaches to promotion: prize perception theories and noncognitive theories of induced behavior.
Price perception theories include theories c reference price and theories of the acceptability of promotions of different sizes. Reference prices, generally, are the amounts consumers expect to p; or will pay for a product or brand. The foundation for most theories of reference price is adaptation-level theory (Helson 1964; Klein and Oglethorpe 1987). The theory posits that consumers integrate all the pricing information they have seen to form adaptation-levels or reference prices. Consumers evaluate a specific price by comparing it with a reference price.
Monetary promotions, which are in the same units as price, are most likely to be framed as reduced losses. These promotions are likely to be integrated with reference price, leading to a downward modification of reference price. Inferences of reduced quality, which may result from a lower reference price (Monroe and Chapman 1987), should more likely result from promotions in monetary units than from promotions in nonmonetary unit!
Diamond and Campbell (1988) tested the effect of different promotional forms on reference price in a laboratory simulation. Subjects were exposed to 20 weeks of the pricing and promotion information on a fictitious brand. Discounts, but not nonmonetary promotions, lowered the reference price of the product.
Campbell and Diamond (1989) hypothesized that because nonmonetary promotions are less likely to be compared with the original price of the product, large nonmonetary promotions should be more likely to be within the consumer's "latitude acceptance" than large monetary promotions. Da from two experiments confirmed this hypothesis.
Sawyer and Dickson (1984) call a second s of behavioral theories of promotion "noncognitive theories of induced behavior." The best known o these is behavioral shaping or operant condition Rothschild and Gaidis (1981) and Peter and Nord (1982) conceived of promotions as rewards or reinforcements, but did not differentiate among promotions. Promotions perceived as separate g (rather than reduced losses) may be most readily perceived as rewards for a purchase. If this is so, nonmonetary promotions would be the most effective reinforcement, and would be the most effective at increasing purchase rates through operant conditioning.
As the preceding discussion shows, a classification of sales promotions based on framing can have significant implications for hypotheses regarding the effects of promotions on consumer! We now consider whether such a classification is consistent with the manner in which consumers actually differentiate promotions.
PERCEPTUAL MAP OF SALES PROMOTIONS
The plan of the study was to develop stimulus and subject weight spaces using a weighted multidimensional scaling of similarity judgments of sales promotions. The universe of sales promotions was developed by adding categories found in area supermarkets to those found in previous taxonomies. No references to specific products or prices were made in the sentence-long descriptions of the promotions. This approach has advantages and disadvantages. Specifics are very important determinants of whether a particular promotion is an effective incentive. However, our purpose was to explore the heuristics which people use to classify and evaluate general types of promotions. If the perceptual map were developed from a set of specific promotions, the dimensions of the promotions might be heavily dependent on the specifics of the stimulus set. The effects of specific promotional and contextual attributes on the evaluation of promotions must be examined in subsequent research.
To reduce the number of judgments subjects had to make, and to keep the stimulus space simple, we eliminated sweepstakes and contests from the preliminary list of stimuli. Sweepstakes and contests are complex phenomena which can not be described with a simple stimulus. The final list consisted of 16 "generic" sales promotions. Abbreviated descriptions of these promotions (which were not presented to subjects) are listed at the bottom of Figure 1.
Twenty students and twenty non-student women judged the similarities of all pairs of the promotions. Such a small sample size has precedents in the literature. Green and Rao (1971) used 30 matrices to evaluate scaling programs. Jones and Young (1972) used the similarity judgments of 31 subjects to describe the longitudinal development of a social environment. Green and Carmone (1969) used the similarity and preference judgments of four different groups, each ranging from 15 to 20 in size, to scale business school images. These methodological studies generated interesting stimulus configurations with small sample sizes.
The students were juniors and seniors at a northeastern state university. The sample of nonstudent women was a diverse convenience sample. They ranged in age from 21 to 65 with a median age group of 35-39 years. The subjects were paid $8.
A personal computer presented all pairs of stimuli in random order to each subject and recorded similarity judgments. Subjects rated the paired promotions from 1 (very similar) to 7 (very different). The similarities data were analyzed using the ALSCAL procedure in SPSSX. A weighted multidimensional scaling (INDSCAL) model provided stimulus and subject weight spaces.
CRITERIA FOR THE RATING AND CLASSIFICATION OF THE SALES PROMOTIONS
A second group of twenty undergraduates rated the promotions on 12 attributes. The 16 promotions were presented in random order on a personal computer. The experimenters also classified the promotions according to two criteria suggested by the theory developed earlier in this paper. Promotions were distinguished by whether the benefits were monetary or nonmonetary and by whether they used another product as a premium. The criteria for classification are presented in Table 1.
At 2, 3, and 4 dimensions, the S-Stress describing the fit of the solution was .351, .282, and .244, respectively. The fit improved far more from 2 to 3 dimensions than from 3 to 4 dimensions. The three dimensional solution is the most interpretable of these solutions, and is very consistent with the classification based on framing. Therefore, the three dimensional solution is presented in Figure 1.
To label the stimulus configuration, we regressed the mean attribute ratings upon the dimensions of the stimulus space (Kruskal and Wish 1978). These regression results are presented in Table 1. Two types of information must be used to interpret this table. First, the multiple correlation between an attribute ratings of the various promotions and the coordinates of the stimulus space (perceptual map) must be high. For instance, the multiple correlation between the attribute "other product as premium" and the coordinates of the stimulus space is .99 . This means that the position on the perceptual map virtually perfectly reflects whether a promotion does or does not use another product as a premium. The second important piece of information is the regression weight of an attribute expressed as a direction cosine. The closer this number is to 1.0 or -1.0, the closer one of the three dimensions of the stimulus space reflects the ratings of the promotions on a particular attribute. The direction cosine of "Other Product as Premium" is nearly 1.0 for the first dimension. Therefore the first dimension almost perfectly reflects whether the promotion uses another product as a premium or not.
The first dimension of the stimulus space correlates most highly with the objective rating of whether or not the promotion utilizes another product (a premium) as an incentive. The second dimension is most correlated with the attribute distinguishing monetary promotions (including discounts and rebates) from all other offers (including extra amounts of product and other product as premiums). The multiple correlations between each of the objective attribute ratings and the positions of the promotions in the stimulus space were greater than .95 .
The attributes leading to the easiest interpretation of the third dimension were the ratings of the time and effort required to utilize the promotional offer. This dimension distinguishes promotional incentives which are immediate and low-effort from delayed and high-effort incentives such as rebates and mail-in offers.
The attribute ratings cast light on the subjects' stereotypes toward promotions. Most prominently, the subjects were not fond of rebates or other mail-in offers. Dimension 3 was negatively correlated with liking of promotion, feelings of savings and value, feeling that the promotion is a reward, and ratings that the promotion is for young people .
The attribute ratings also provide evidence about how promotions are framed. Subjects rated promotions on a 7-point scale from "makes me feel that I am losing less than usual" to "makes me feel that I am gaining something extra." Promotions framed as gains should be rated on the high end of the scale. The regression coefficient relating dimension 3 (high time and effort) and framing showed indicated that promotions perceived as requiring much time and effort were rated as reduced losses (t(12)=3.6, p<.01). The regression coefficient relating dimension 2 (monetary/ nonmonetary) and the perception as a gain was also significant (t( 12)=2.7, pc.OS). Nonmonetary promotions (including extra amounts of products and premiums) tended to be perceived more as gains than monetary promotions.
To aid in interpretation, the points in the stimulus space were clustered using the centroid method. The dimensions were weighted by the average squared subject weights. The coefficient representing the distance within clusters has the values .52, .23, .16 and .13 for one, two, three and four clusters. The three-cluster solution appears to be helpful in interpreting the stimulus space.
Cluster A (triangles in Figure 1) comprises the promotions which use another product as a premium. Cluster B (squares in Figure 1) contains promotions which pride an extra quantity of the promoted product. Cluster C (circles in Figure 1) includes the price-off or cash rebate offers. Thus, the three-cluster solution distinguishes promotions on the basis of the units of benefit: whether these units are monetary, extra amounts of the promoted product, or units of another product as a premium. (If the two-cluster solution is adopted, the two nonmonetary clusters collapse into one cluster. The two remaining clusters, monetary units vs. nonmonetary promotions, is equivalent to the framing distinction presented earlier.)
In this paper, we have developed a theoretical classification of sales promotions. This classification is not fundamentally different than some classifications which have been used by managers . However, the theoretical foundation of the present distinction may lead to elaboration and extension of behavioral theories of sales promotions.
The classification is based on the hypothesis that since nonmonetary promotions are more difficult to integrate with the cost of the product, they are more likely to be framed as separate gains. Nonmonetary promotions include extra amounts of the promoted product and premiums using another product. The logic of this hypothesis derives from Thaler's (1985) work on mental accounting.
Consumer similarity judgments produced a perceptual map which strongly resembles the theoretical classification. Subjects did discriminate between monetary and nonmonetary promotions. Unexpectedly, promotions which offered extra amounts of the promoted product were seen as different than promotions offering premiums. Moreover, subjects discriminated and evaluated promotions on the basis of whether they required much effort and time.
Ratings of the promotions support the basis for the classification based on framing. Subjects tended to rate nonmonetary promotions as gains and monetary promotions as reduced losses. Our research thus supports Thaler's approach. However, it differs in the classification of rebates. Thaler hypothesized that rebates would be framed as segregated gains if they were physically or temporally segregated from the purchase price quotation. Our research found that subjects considered monetary rebates to be reduced losses. This may be due to consumers' increasing familiarity with such offers or the increasing practice of expressing prices as "after rebate." The framing of rebates as gains might re-emerge in future research, when the specifics and context of the promotions are described in more detail.
Other attribute ratings suggest some of the subjects' stereotypes toward promotions. For instance, nonmonetary promotions and immediate incentives were rated as more appropriate for young people. These were also rated as feeling the most like a reward, and may be the most appropriate for behavioral shaping (Rothschild and Gaidis 1981). It may be rewarding to investigate these and other connotations of sales promotions.
Future research in this area should address two limitations of the present work. First, these promotions were "generic." Consumer reactions to promotions may depend on the specifics of the products being promoted, and the magnitude and type of the incentives. Secondly, the samples used here were not large, nor were they necessarily representative of specific populations. The generalizations from these groups must be considered exploratory.
Despite these limitations, we believe that the correspondence between a classification of promotions based on framing and the empirical results reported here is sufficient to encourage the study of the different behavioral properties of promotions the consumer frames as gains and those the consumer frames as decreased costs.
Abelson, Robert P. and Ariel Levi (1985), "Decision Making and Decision Theory," in Handbook of Social Psychology, Gardner Lindzey and Eliot Aronson, eds. New York: Random House, 231-310.
Beach, Lee Roy and Terence Mitchell (1978), "A Contingency Model for the Selection of Decision Strategies," Academy of Management Review, 3 (July), 439 449
Campbell, Leland and William D. Diamond (1989), "Effects of Framing on the Perception of Sales Promotions. Paper presented at the Annual Meeting of the American Academy of Advertising, San Diego.
Diamond, William D. and Leland Campbell (1988), "The Framing of Sales Promotions: Effects on Reference Price Change." Paper presented at the Annual Meeting of the Association for Consumer Research, Honolulu.
Green, Paul E. and Frank J. Carmone (1969), "Multidimensional Scaling: An Introduction and Comparison of Nonmetric Unfolding Techniques," Journal of Marketing Research, 6 (August), 330-341.
Green, Paul E. and Vithala R. Rao (1971), "Multidimensional Scaling and Individual Differences", Journal of Marketing Research, 8 (February), 71-77.
Helson, Harry (1964), Adaptation-Level Theory, New York: Harper Row.
Henderson, Caroline M. (1985), "Modeling the Coupon Redemption Decision", Advances in Consumer Research, Vol 12, Elizabeth C. Hirschman and Morris B. Holbrook, eds., Provo, UT: Association for Consumer Research, 138143.
Irons, Karl W., John D.-C. Little and Robert L. Klein (1983), "Determinants of Coupon Effectiveness," in Advances and Practices of Marketing Science 1983, Fred Zufryden, ed., Providence: ORSA/TIMS, 157-164.
Jones, Lyle E. and Forrest W. Young (1972), "Structure of a Social Environment: Longitudinal Individual Differences Scaling of an Intact Group," Journal of Personality and Social Psychology, 24 (Oct), 108-121.
Kahneman, Daniel and Amos Tversky (1979), "Prospect Theory: An Analysis of Decision Under Risk," Econometrica, 47 (March), 263-291.
Kahneman, Daniel and Amos Tversky (1984), "Choices, Values and Frames," American Psychologist, 39 (April), 341-350.
Klein, Noreen M. and Janet E. Oglethorpe (1987), "Cognitive Reference Points in Consumer Decision Making," Advances in Consumer Research, Vol 14, Mellanie Wallendorf and Paul Anderson, eds., Provo, UT: Association for Consumer Research, 183- 187.
Kruskal, Joseph B. and Myron Wish (1978), Multidimensional Scaling, Beverly Hills: Sage.
Monroe, Kenl B. and Joseph P. Chapman (1987), "Framing Effects on Buyers' Subjective Product Evaluations, Advances in Consumer Research, Vol 14, Mellanie Wallendorf and Paul Anderson, eds., Provo, UT: Association for Consumer Research, 193- 197.
Narasimhan, Chakravarthi (1984), "A Price Discrimination Theory of Coupons," Marketing Science, 3 (Spring), 128-147.
Neslin, Scott A. and Robert W. Shoemaker (1983), A Model for Evaluating the Profitability of Coupon Promotions," Marketing Science, 2 (Fall), 361-388.
Payne, John W. (1982), "Contingent Decision Behavior," Psychological Bulletin, 92, 382-402.
Peter, J. Paul and Walter R. Nord (1982), "A Clarification and Extension of Operant Conditioning Principles in Marketing," Journal of Marketing, 46 (Summer),,102-107.
Puto, Christopher P. (1987), "The Framing of Buying Decisions," Journal of Consumer Research, 14 (December), 301-315.
Rothschild, Michael L. and William C. Gaidis (1981), "Behavioral Learning Theory: Its Relevance to Marketing and Promotions," Journal of Marketing, 45 (Spring), 70-78.
Sawyer, Alan G. and Peter R. Dickson (1984), "Psychological Perspectives on Consumer Response to Sales Promotion," in Research on Sales Promotion: Collected Papers, Katherine E. Josz, ed. Cambridge MA: Marketing Science Institute, 47-62.
Schindler, Robert M. and Stacy E. Rothaus (1985), "An Experimental Technique for Exploring the Psychological Mechanisms of the Effects of Price Promotions," Advances in Consumer Research, Vol 12, Elizabeth C. Hirschman and Morris B. Holbrook, eds., Provo, UT: Association for Consumer Research, 133-137.
Thaler, Richard (1985), "Mental Accounting and Consumer Choice," Marketing Science, 4 (Summer), 199-214.
Thaler, Richard and Eric J. Johnson (1986), "Hedonic Framing and the Break-Even Effect," unpublished working paper.