Special Session Summary Extending Measures of Advertising Effectiveness: Ads’ Effects on Price Sensitivity

Ronald C. Goodstein, Indiana University
[ to cite ]:
Ronald C. Goodstein (1997) ,"Special Session Summary Extending Measures of Advertising Effectiveness: Ads’ Effects on Price Sensitivity", in NA - Advances in Consumer Research Volume 24, eds. Merrie Brucks and Deborah J. MacInnis, Provo, UT : Association for Consumer Research, Pages: 109-111.

Advances in Consumer Research Volume 24, 1997      Pages 109-111



Ronald C. Goodstein, Indiana University

Behavioral research on advertising effectiveness has typically been measured in terms of recall, attitudes, and purchase intentions. While these measures are vital determinants of an ad’s effects, they fail to incorporate a key economic measure, namely the influence of advertising on price sensitivity. Economists offer two different arguments on the relationship between advertising and price sensitivity. If ads simply associate brands, then they argue that advertising is simply a form of information about market alternatives (Nelson 1974). As such, it would serve to raise price sensitivity in the marketplace (Mitra and Lynch 1995). Much advertising, however, works to differentiate one product from others in the category. This form of advertising provides market power to the brand and decreases the price elasticity of demand (e.g., Comanor and Wilson 1979). Thus the brand builds equity as evidenced by the ability to demand a premium price in the marketplace (Aaker 1991).

Instead of trying to address whether advertising provides market power or information, the set of papers presented in this session assumes that advertising has the ability to do either depending upon the content of the ad. That is, there are certain types of advertisements that are intended to provide information and others that are meant to differentiate brands and build equity. However, the research presented goes beyond this traditional perspective to investigate the contingencies existing toward these economic predictions.

The paper by Mitra and Lynch, for instance, suggests that the effects of advertising on consumer price elasticity and prices paid depends on the degree to which the choice situation requires brands to be recalled in order to be considered. They find that when the choice context forces consumers to recall ads from memory, then the ads serve to increase price elasticity. However, when brand names are available in the choice context, advertising serves to lower elasticity, indicating that the ads differentiate the brands. Goodstein and Kalra illustrate the classic effects of advertising positioning on price sensitivity. Specifically, holding ad content constant, they find that relative to feature-based positioning (Aaker 1991), value positioning increases both brand and category price sensitivity, while unique attribute positioning decreases both. Minor brand versus premium brand positioning serves to decrease brand price sensitivity, but increase category-level sensitivity. Finally, premium versus premium brand positioning lowers both brand and category sensitivity.

Mela, Gupta, and Lehmann then examine the relative effectiveness of advertising versus promotions on price elasticity depending on the time frame used to investigate these effecs. Using data from a proprietary consumer panel incorporating eight years of weekly observations, they show that while promotions increase purchases over the short-run, they negatively affect elasticity over the long haul. This counters earlier predictions (e.g., Davis, Inman, and McAlister 1992) stating that promotions have no "bad" long-term effects. Their data concerning advertising are more encouraging. Advertising may have less of an effect on sales, relative to promotions, in the short-run. However, the long-term effects of advertising are to increase brand equity as evidence by a decrease in price elasticity over this time horizon.

Thus this session serves to take us "beyond the box" of what we currently know about the relationship between advertising and price sensitivity. The session uses both experimental and panel-based methodologies to explore these issues. Further, the session illuminates several important and rather common contingencies affecting this relationship. Our discussant, Steve Hoch, highlighted the important contributions of this research and suggested other contingencies that should be examined to test the boundaries of the economic models. Further, Steve worked with the authors and audience to develop an agenda for future research in this area, focusing on interesting ways to combine economic and behavioral methodologies to gain additional insights into this stream of research.



Anusree Mitra, American University

John G. Lynch, University of Florida

Two competing research streams in economics have postulated divergent theoretical accounts and have made conflicting predictions about the economic and welfare effects of advertising. The first model emphasizes the ability of advertising to "artificially" differentiate products, which lowers the price elasticity and makes consumers unwilling to switch from their favorite brands to cheaper alternatives. The second model stresses that advertising provides information about alternatives and actually increases price elasticity and lowers prices paid. This research contributes to the debate on the economic effects of advertising by identifying the two conflicting perspectives as parts of a generalized theoretical model of advertising effects on price elasticity at the level of the individual consumer. Results of an experiment conducted to test theoretical predictions based on our model showed that the effects of advertising on consumer price elasticity and prices paid depends on the degree to which the choice situation requires brands to be recalled in order to be considered. In accordance with our theoretical predictions, in this study we found support for both the schools of thought (Mitra and Lynch 1995).

In a second study, we focused on the effects of advertising on consumer welfare. In Experiment 1 we had observed that when the effects of advertising on brand name recall for consideration set formation were superfluous, advertising decreased price elasticity and increased purchase prices. The market power view would argue that this occurred because advertising leads consumers to value and therefore pay a premium for differentiating attributes that would otherwise be seen as trivial; therefore, advertising has detrimental effects on consumer welfare. On the other hand, information economists contend that even if advertising and reduces price sensitivity and increases prices paid, consumers may nonetheless be better off. Rosen (1978) argues that advertising helps match brands and consumers in the presence of dispersion of the product attributes among brands and heterogeneity of preferences among consumers. Stated differently, attribute information in advertising allows consumers to concentrate search efforts on their most preferred brands and thereby enables consumers to choose brands that more closely reflect their personal tastes than would be true without advertising.

In Experiment 2, we examined the effect of advertising on consumers’ ability to choose in line with their personal tastes by closely replicating two conditions out of the six cells from Experiment 1. Subjects either saw no advertising or differentiating advertising, and then made choices. Our results suggest that advertising need not always have detrimental effects on consumer welfare as predicted by the market power view. In conditions nearly identical to those in Experiment 1 that showed differentiating ads decreased price sensitivity and increased prices paid, we showed that these ads provided useful information that allows consumers to make choices more in line with their personal tastes than would be possible without advertising (Mitra and Lynch 1996). Taken together, these experiments enabled us to develop new insights into the informative versus persuasive roles of advertising that have been central to the debate in the economics literature.



Ronald C. Goodstein, Indiana University

Ajay Kalra, Carnegie-Mellon University

Recent criticism argues that much of today’s advertising fails to differentiate the brands available to the consumer (e.g., Achenbaum 1992). In fact, many ads associate brands as being more alike than distinctive (Pechmann and Ratneshwar 1991). Other advertising, however, works by providing pricing information (e.g., Popkowski Leszczyc and Rao 1990) or by differentiating brands in the market (e.g., Comanor and Wilson 1979). We believe, like others, that these different advertising strategies affect the brand and market in very different ways, even though the ads may contain the same factual information (Kaul and Wittink 1995; Mitra and Lynch 1995). For instance, the predicted effects on price sensitivity are conceptualized to be that price-oriented and nondifferentiating advertising increases sensitivity, while differentiating ads reduce sensitivity (e.g., Bolton 1989; Popkowski Leszczyc and Rao 1990). To date, the limited empirical support for this conjecture has been criticized as being open to other interpretations (Mitra and Lynch 1995). While aggregate data have been used to study these effects, little experimental research has examined individual-level reactions to different types of advertising positioning on the price sensitivity of demand. We believe that this type of experimentation provides opportunities for examining micro-level ad differences that may be obscured by aggregate level data collection (i.e., scanners) and analyses.

This research is similar in spirit to others examining the impact of ad strategies on price sensitivity (e.g., Krishnamurthi and Raj 1985; Mitra and Lynch 1995) with several notable exceptions. First, we present a conceptual model that differentiates between information-based and power-based positioning strategies and test their effects on price sensitivity. Specifically, we test how different ad positioning tactics affect price sensitivity using a variety of measurements. The positioning tactics examined differ in terms of their highlighting price versus non-price attributes (though the ad information is constant across conditions), and in terms of how many brands in the competitive set contain the highlighted attribute (one in the case of noncomparative ads and two in the case of comparative ads). The positioning tactics examined include various comparative advertising appeals (e.g., Pechmann and Ratneshwar 1991) and unique attribute advertising (e.g., Nowlis and Simonson 1996). Second, other work in this area typically examines the price elasticity for an advertised brand within a product category. However, the competitive marketplace dictates that advertising affects both the sponsor’s brand as well as the competitive offerings in the category (e.g., Unnava and Sirdeshmukh 1994). Thus, our study examines price sensitivity at both the brand and category levels.

To investigate these issues, in Study 1 we test the effects of advertising content on willingness to pay for a brand and the relative importance of the prie attribute in category decisions. More details on both brand and category reactions are provided in Study 2 by adding choice level measures of price sensitivity to our dependent measures. Our measures extend those traditionally used to test advertising effectiveness by illustrating that advertising’s effects on price sensitivity may be a good indicator of effects when ad attitude measures fail to differentiate between ads.

Our two studies reveal that although many ads contain equivalent attribute information, the positioning adopted to present that information will differentially impact price sensitivity. Further, we show that the same ad positioning strategy can affect brand and category price sensitivity in opposite directions, much as end-of-aisle displays lower price sensitivity for the displayed brand and raise it for the product category (Bolton 1989). In Study 1, we found that positioning tactics associating a minor brand to a premium brand, as well as unique attribute positioning, raise consumers’ willingness to pay for a brand. Value-oriented positioning, however, lowers their willingness to pay for the advertised product. Additionally, we show that these same strategies have different effects on category sensitivity. While unique attribute advertising lowers the importance of price within the category, both value-oriented and associating positioning increase price sensitivity at the category level.

Study 2 replicated and extended our findings from the first experiment using additional measures, a different product, and different subjects. In particular, we again find that value-oriented ads increase price sensitivity at both the brand and category levels. We also replicate the finding that comparisons positioning the minor brand as more similar to the premium brand decrease price sensitivity at the brand level, while increasing sensitivity at the category level. When the positioning adopted compares two premium brands, however, we find decreased price sensitivity for the brand, but mixed evidence on category sensitivity.



Carl F. Mela, Notre Dame University

Sunil Gupta, Columbia University

Donald R. Lehmann, Columbia University

This paper examines the long term effects of promotion and advertising on consumers’ brand choice behavior. Specifically, we use eight years of panel data for a frequently purchased packaged good to address two questions: (1) Do consumers’ responses to marketing mix variables, such as price, change over a long period of time? (2) If yes, are these changes associated with changes in manufacturer’s advertising and retailer’s promotions? Based upon these results, we draw implications for manufacturers pricing, advertising, and promotion policies.

According to Progressive Grocer (1995), almost 70% of firms have increased their trade promotions between 1990 and 1995, mostly at the expense of advertising. Currently, total trade and consumer promotional expenditures represent $70 billion annually. The effect of this reallocation of marketing dollars is much debated in the industry. On the one hand, some believe the shift has made consumers more price sensitive (Brand Week 1993) while others believe the shift has helped brands (Wall Street Journal 1992). Implicit in all the speculation is that there exist short term effects (what is the effect of a price change on choice behavior this week?) and long term effects (how has years of price promotions or advertising affected the response to a price change?).

Academic researchers have achieved somewhat greater consensus regarding the defection of marketers from advertising. The long-term impact of advertising on choice behavior has been summarized by Kaul and Wittink (1995) and Mitra and Lynch (1995). In essence, advertising is hypothesized to reduce price sensitivity (and presumably price promotion sensitivity) to the extent advertising is i) brand rather than price oriented and ii) consideration sets are relatively stabe. Both conditions are likely true for mature product categories such as the one we analyze. We therefore hypothesize that advertising leads to a reduction of price and price promotion sensitivity in our data. Additional insights regarding advertising’s effects are obtained by considering possible heterogeneity in consumer response to advertising. As loyals are relatively price insensitive, we expect that any effects of advertising will likely be stronger for non-loyals than loyals. Finally, as mass media advertising has been shown to reduce consumer price sensitivity and build loyalty, we expect that advertising will increase the size of any loyal segment that exists.

Somewhat less consensus exists regarding the effect of price promotions on brand choice. Speculation that promotions hurt brands has been advanced by Dodson, Tybout, and Sternthal (1978) who use consumers are likely to attribute their behavior to the presence of promotions rather than brand preferences. Increased promotion in a category is likely to lead to a perception that the key differentiating feature of a brand is its price (Sawyer and Dickson 1983). However, many researchers (Davis, Inman and McAlister 1992, Ehrenberg, Hammond, and Goodhardt 1994, Hariharan 1992, Johnson 1984) have found little evidence of the negative long term effect of promotions. However, consistent with the consumer behavior theories and the converse effects of price promotions to those described in that advertising literature, it is our expectation that price promotions will have many of the opposite effects of advertising.

The effect of display advertising, on the other hand, may depend upon whether consumers perceive displays to function as advertising or as signals of price cuts. According to Inman, McAlister, and Hoyer (1990), promotional signals have a positive impact on the "low need for cognition" individuals. However, high need for cognition individuals react to a promotion signal only when it is accompanied by a substantive price cut. Loyals, as habitual buyers, are less motivated to actively process pricing information and may therefore focus less on price in response to display advertising. Non-loyals, as active processors, are likely to become think of displays as price promotions than advertising and therefore become more price sensitive in response to displays.

To capture these advertising and promotion effects, we use a two-stage modeling approach. In stage one, we estimate a segment level logit model for each quarter of the data to capture the effect of short term (weekly) variables on consumers’ brand choice. In stage two, we use the partial-adjustment model to see if consumers’ responses to short term marketing activity can be explained by medium term advertising and promotion effects (quarterly). We then calculate the long term advertising and promotion effects (e.g., over an infinite horizon).

Our results are consistent with the hypotheses that consumers became more price and promotion sensitive over time and that these changes correlated with reductions in advertising and increases in promotions over the eight year duration of the study. The effects were found to be significantly larger for non-loyal consumers then for loyal consumers. Decreases in advertising also correlated with a reduced number of loyals. In general, compared to the "good" effects of advertising, promotions were found to have significantly larger "bad" long run effects on consumer’s price and promotion sensitivities.