Consumer Price and Promotion Expectations

7/27/2019 Consumer Price and Promotion Expectations

1/12

/AANOHAR U. KALWANI and CHI KIN YIM*

The authors report results from a controlled experiment designed to investigatethe impact of a brand's price promotian frequency ond the depth of promotionalprice discounts on the price consumers expect to poy for that brand. A key featureof the work is that expected prices elicited directly from respondents in the exper-iment are used in the analysis, os opposed to the latent or surrogote measures ofexpected prices used in previous studies. As hypothesized, both the promotion fre-quency and the depth of price discounts are found to have a significant impact onprice expectations. Evidence also supports a region of relotive price insensitivityaround the expected price, such that oniy price changes outside that region havea significant impact on consumer brand choice. Further, the outhors find that con-sumer expectations of both price and promotional activities should be consideredin explaining consumer brand choice behavior. Specifically, the presence of a pro-motional deol when one is not expected or the obsence of a promotional deal whenone is expected may hove a significont impact on consumer brond choice. Finally,as in the case of price expectations, consumer response to promotion expectations

is found to be asymmetric in that losses loom larger than gains.

Consumer Price and Promotion Expectations:An Experimental Study

The role of expectations in predicting the behavior ofeeonomic agents has long been widely accepted in theeconom ics literature (Muth 1961; Nelson 1977). In mar-keting, the study of the impact of price expectations onconsumer choice behavior has begun to receive increas-ing attention In recent years. The intuitively appealingproposition that consumers form price expectations anduse them in evaluating price infonnation when makinga purchase has both theoretical and empirical support{Gurumurthy and Little 1989; Kaiwani et al. i990; Lat-tin and Bucklin 1989; Puller 1990; Raman and Bass 1988;Rinne 1981; Thaler 1985; Winer 1986). Introducing aproduct at a lower than regular price and then raising theprice afterward to its regular level has been shovi'n to

Manohar U. Kalwani is Professor ot" Management . Krannen Grad-uate School of Management, Purdue University. Chi Kin Yim is As-sistant Professor of Administrative Science. Jesse H. Jones GraduateSchool ot" Adm inistration. R ice U niversity.The authors thank Randy Bucklin, Jim Lattin, Dan Putler. the ed-i tor , and three anonymous JM R reviewers for their helpful commentsand suggest ions.

have an adverse effect on subsequent sales. The reasis that consumers come to adopt the low introductoprice as a reference and consider the regular price to unacceptably greater than the price they expect to p(Doob et al. 1969). From a managerial viewpoint, uderstanding how consumers form and use price expetations in making purchase decisions is important bcause the failure to incorporate price expectations hbeen shown to result in a misestimation of price elastity, which can lead to nonoptimal pricing decisions (Doyand Saunders 1985).Recently, the price expectations hypothesis has beused to provide an altemative explanation for the oserved adverse long-term effect of price promotions brand choice (Kaiwani et al. 1990). Previous researhas shown that repeat purchase probabilities of a braafter a promotional purchase are lower than the corrsponding values after a nonpromotional purchase (Doson, Tybou t, and Sternthai 1978; Guada gni and L it1983; Shoemaker and Shoaf 1977). Dodson, Tybout, aSternthai evoke self-perception theory to predict that a purchase is induced by an external cause (such asprice promotion) as opposed to an internal cause (e.90

Journal of Marketing ResearchVol. XXK (Febniary 1992), 90-1


2/12


3/12

JOURNAL OF MARKETING RESEARCH, FEBRUARY 1" of price promotions of the brand. Further, the rela-tionship between the expected price and the priceprotnotion frequency can be approximated by a sig-moid function.

Tversky and Kahneman (1974) have shown that peo-ple rely on a limited number of heuristic principles thatreduce complex tasks of assessing probabilities and pre-dicting values to simpler judgmental operations. In somecases, people may anchor and adjust their forecasts bystarting with a preconceived point and weigh that pointheavily in arriving at a jud gm ent. When the frequencyof past price promotions is "very low," consumers iden-tify a price promotion offer as an exceptional event andmay not modify the brand's expected price.' The brand'sexpected price then will be anchored around the regularprice because of insufficient adjustment. In other cases,people m ay arrive at a judgm ent on the basis of howsimilar or representative the event is to a class of events.Therefore, when a brand is price promoted "too often."consumers come to expect a deal with each purchase andhence expect to pay only the discounted price on the ba-sis of its representativeness. Clearly, given a certain levelof price discount, the brand's expected price will bebounded by the regular price and the implied sale price.That tine of reasoning suggests that the relationship be-tween the price promotion frequency and the expectedprice can be approximated by a sigmold function.

Impact of Depth of Price Discounts on Expected PriceOur second hypothesis pertains to the impact of thedepth of price discounts offered during a brand's pricepromotions on its expected price.

H^: The price con.sumers expect to pay for a brand de -crease.s with an increase in the observed depth ofprice discounts of the brand. Further, the relationshipbetween the expected price and the depth of pricediscounts can be approximated by u concave func-tion.

Whether a price discount will affect the brand's ex-pected price depends on how consumers perceive thediscount. Uhl and Brown (1971) postulate that the per-ception of a retail price change depends on the magni-tude of the price change. They report results from anexperiment indicating that 5% deviations were identifiedcorrectly 64% of the time wherea s 15% deviations w ereidentified correctly 84% of the time. Delia Bitta andMonroe (1980) find that consumers" perceptions of sav-ings from a promotional offer do not differ significantlybetween 30% , 40 % , and 50% discount levels. H owever,they find significant differen ces betwee n the 10% and

'Terms such as "low" and "of ten," which are used to descr ibe Ihcfrequency of price promotions, are all relative. Please refer lo thesection discussing the ncatment levels to put such qualifiers into properperspect ive.

30 to 50% levels . They also discuss some m anagersliefs that at least a 15 % discou nt is needed to attract sumers to a sale. Apparently, small price changes not be noticed and even a large price reduction (sayor 70%) may not be assimilated to affect the braexpected price if it is considered exceptional. Henceimpact of the depth of price discounts on loweringbrand's expected price is likely to occur when the pdiscount offered by the brand is relatively large butso large that it is seen as an exceptional event.

In our research, we chose to examine price discoranging from !0 to 40%', a range commo nly used inresearch on price discounts in the consumer packgoods categories (Berkowitz and Walton 1980; Cuand Kopp 1986). That was also the range of price counts available in stores in the local area where theperiment was conducted. Within that range, the findof Uhl and Brown (1971) and Delia Bitta and Mo(1980) suggest that it is reasonable to expect the tionship between the brand's expected price and the dof price discounts to be concave.Consumer Price Expectations and Brand Choice

Previous research provides plenty of evidence to port the hypothesis of a standard price serving as a erence level for price judgments (Helson 1964; Kaman and Tversky 1979; Monroe 1973). When modeconsumers' evaluation of a purchase, in addition toelement of consu mer surplus or acquisition utility as gested by standard economic theory. Thaler (1985) ttjlates a transaction utility that pertains solely to the mof the "deal" available or the perceived savings ofoffer. Spec ifically, a negativ e transac tion utility (wthe retail price exceeds the expected price) will redthe brand purchase probability, whereas a positive traction utility (when the retail price is less than thepected price) will enhance the brand purchase probaity.As a further refinement of how consumers responprice changes, Raman and Bass (1988) apply assimtion-contrast theory (Sherif 1963) to postulate a regof price insensitivity around a brand's expected price that changes in price within that region produce no machange in perceptions. Hence, only price changes are outside that region are assumed to have a signifiimpact on consumer brand choice. We identify threshold values representing the boundaries of thegion of indifference empirically on the basis of the qity of the model fit.As mentioned, there is a growing amount of empievidence on the role of reference price in determiconsumer response to retail prices. Two recent stuby Winer (1985. 1986), based on linear probability moof consumer response to price information for the chase of seven durable goods and three brands of coreveal that a retail price higher than ihe reference p("sticker shock") has a statistically significant nega


4/12

CONSUMER PRICE AND PROMOTION EXPECTATIONS 93

Gurumurthy and Little (1989), Kalwani et al. (1990),

n exponentially smoothed measure of a reference priceo model its effect on brand choice, but they find thathe reference effect of price is not significant. Never-theless, all of the authors cited, who rely on consumerpurchase data to calibrate their models, only use a sur-rogute measure of the expected price. Our approaeh dif-fers from previous research in that we investigate theimpact of price expectations on brand choice by using airect measure of the expected price from a controlledxperiment.pecification of the Brand Choice Model

We propose a binary logit model to represent the im-pact of a brand's expected price, in relation to its retailprice, on a consumer's probability of purchasing thatbrand. Following McFadden (1973). we assume the con-sumer chooses the brand that offers more utility and therandom errors on utility are independent and identicallyGumbel distributed; the probability of choosing brand iis given by(I ) Pik -

where V,* is the deterministic component of utility thatconsumer k derives trom the purchase of brand /.We model the deterministic compt^nent of the utilitythat consumer k derives from the purchase of brand ( asa linear function of the consumer's preference of brand( (directly m easured in the expe rim ent ), the retail priceof brand /, a O-I dummy variable indicating whether ornot brand / is available on a promotional deal, and theperceived price gain and loss variables. In operational-izing the price gain and loss variables, we take into ac-count two factors, (I) a postulated region of price in-sensitivity around the expected price and (2) theproptisition that the utility consumers derive from pur-chasing a brand may not vary linearly with the magni-tude of the difference between the brand's expected andretail prices. In particular, we posit that the price gainvariable takes a value of I if a brand's expected priceexceeds its retail price by a certain threshold level, 6^.,and takes the value 0 otherwise. Analogously, the priceloss variable is assumed to take the value ! if a brand'sretail priee exceeds its expected priee by a certain thresh-old level, 6,, and is assumed to take the value 0 other-

w is e. ' The th reshold lev els, 6 . and 0,, are mea sured inpercentage terms as a fraction of a brand's average non-promotional price. We subsequently estimate the thresh-old parameters, 6^ and 6,. empirically through a grid searchon the basis of the fit of our brand choice model. Math-ematically, we represent the deterministic component ofutility, V,t, as(2)

where:

= a,,, +

i= 1 ,2 ,

brand-specific constant for brand / (onlyone brand-specific constant is specified toavoid singularity in the maximum likeli-hood estimation),consumer k's preference for brand /,the retail price of brand / as observed byconsumer A,a dummy variable that captures the effectof a promotion of brand / observed by con-sumer k.

= 1 if

Loss,,= > 6,. 0 otherwise.

0 otherwise.EPit - the expected price of brand / directly elic-ited from consumer k, andRP n = the average nonpromotional price of brand( as observed by consumer k.

Because the coefficient ai captures the effect of brandpreference on brand choice, we expect it to have a pos-itive sign. The coefficient a^ represents the direct effectof price and so is expected to be negative. The promo-tion variable can be interpreted as representing the short-term effect of promotions and we expect a, > 0. Finally,we expect a perceived price gain (perceived price loss)to have a positive (negative) effect on brand choice and,hence, we expect 04 > 0 and a^ < 0.DESIGN OF THE EXPERIMENT

OverviewThe experiment consisted of two parts. The first part

was designed to study the impact of different price pro-motion schedules on brands' expected prices. The sec-ond part was aimed at investigating the role of con-sumers' price expectations in a brand choice situation.

"We are indebted 10 one of the anonym ous JM R reviewers for theconstruction of a region ot" price insensitivily around the bran d's e x-pected price. Raman and Bass (1988) and Gurumurthy and Little (1989)also report empirical evidence in support of that latitude of acceptableprice difterences.


5/12

94 JOURN AL OF MAR KETIN G RESEARCH. FEBRUARY 1Briefly, a group of respondents were exposed to priceand promotion infonnation on two competing l iquidlaundry detergent brands and then were asked to putthemselves in a hypothetical buying scenario in whichthey needed to purchase laundry detergent from a localsupermarket. The laundry detergent product category wasselected on the basis of the results of a pilot study anda pretest of the actual experiment. The pilot study wasused to identify key product attributes for characterizingthe brands in the experiment, the level of regular brandprices, and the hypothetical names of brands.

T w o hypothetical brands of liquid laundry detergentwere used to minimize the potential respondents' ten-dency to invoke an image or previous experience, whichcan cause variability (Monroe 1977). Descriptions andperformance ratings of the two brands were distinct. Forinstance, for one brand we emphasized versatility or highperfonnance in all temperatures and for the other brandwe emphasized whitening power. The assignment of a t-tributes to the hypothetical brands was based on two ma-jor design considerations: (1) each brand should have somedistinct attributes (attributes not shared by the other brand)to allow a variation in the brand preference and (2) toallow the possibility of switching between the two brands,neither of the brands should be dominated by the otherbrand. The regular prices of the two brands were set atlevels consistent with actual retail store data and re-spondents' perceptions. The higher priced brand was usedas the target brand and the lower priced hrand was usedas the control brand in terms of the manipulation of theprice promotion frequency treatment variable (describedin detai l subsequen tly). A randomly determined 1 %variation was added to the regular price to represent mi-nor changes in store prices and to simulate the effect ofl imited memory of consumers.Research Design

The research design for testing the impact of pricepromotions on brands ' expected prices was essentia l ly a4 (price promotion frequencies) by 4 (depth of price dis-coun ts) bet ween -subject factorial de sign . The four levelsof price promotion frequencies were 1, 3, 5, and 7 pricepromotions over 10 weeks. The four levels of price dis-counts were 10, 20. 30, and 40% off the average non-promotional price. The pretest revealed that applying thesame manipulations to both the target brand and the con-trol brand appeared to result in respondent informationoverload. Respondents were not able to separate theidentity and the past price and promotion information(especially the frequency of price promotions) of the twobrands. Brand choices seemed to be made randomly.Therefore , in the experiment, the i6 treatments were ap-plied only to the target brand and the treatments appliedto the control brand were restricted to oniy one promo-tion over 10 weeks and the four levels of price discounts.Recall that the target brand was also the higher pricedof the two brands and, hence, our experimental setup canbe viewed as a market comprising a national brand that

is price promoted frequently and a private label bthat is seldom price promoted.Sample Design and E.xperimenlal Procedures

Two hundred undergraduate students enrolled introductory marketing management classes at a large western university served as respondents for the eximent. From the 200 respondents, we collected 188 of completed responses. Responses were collectedteractively on a computer. The computer was usedstimuli presentation and response recording. The usan interactive computer experiment was also helpfuthe implementation of individualized randomization enhancing respondent involvement in the experim(Aronson, Brewer, and Carlsmith 1985). Purchase text, retail outlet, and package size were specifiedcontroi for variability of price perception due to tcontextual variables.

The experiment was conducted in a computer laratory. Upon arrival , the respondents were assigned dom ly to one of the 16 treatment co nditio ns. They wgiven a handout describing the objectives of the eximent, the phenomena represented in the experiment, how to use the computer. A practice session at the ouof the experiment familiarized respondents with the chase si tuat ion and the use of the computer. Two shpoo brands were used in the practice session.Respondents were debriefed after the experiment. debriefmg was used to explore with each respondentimpact of the experimental events and to seek answto such questions as whether the instructions were clwhether or not the respondent was suspicious, and w heor not particular suspicions would invalidate the resuThe true objective of the experiment then was reveato the respondent .Measures

Three of the variables directly measured in the expiment were (1) brand preferences, (2) expected brprices, and (3) brand choice. Consumers' preferences the brands were measured hy a constant-sum preferequestion in which respondents were asked to allocate ipoints between the two brands of laundry detergent areading a description of the benefits and features (c luding price) and performance rat ings of each braThis measure was obtained before the respondents wexposed to any price and promotion information on two brands.The brands' expected prices were measured by askthe respondents to answer the fol lowing open-enquestion: "Based on the prices of Brand X over the p10 weeks, how much do you expect the price of BrX to be this week (week ID ? $ " Puto (198Rowe and Puto (1987), and Zeithaml and Graham (19have used similar questions to obtain measures of pected prices. Prior to being asked to give their perct ions of brands" expected prices, the respondents wexposed to 10 wee ks (week 1 throug h week 10) of pr


6/12

A N D PROMOTION EXPECTATIONS 95

$ Off" (flashing on the com pute r screen inFinally, the two brands' prices and promotion infor-

Brand choice was obtained by requesting the re-

Selected other questions were asked during and after

DISCUSSION OF FINDINGSManipulation checks reveal that the manipulations of

Price Promotions on Consum ers' PriceWe tested our two price expectations hypotheses in

-{I) the frequency

The results of the ANOVA are summarized in Table

Table 1ANOVA RESULTS WITH BRAND'S EXPECTED PRICE AS

THE DEPENDENT VARIABLE

'Tw o 4 x 4 factorial analyses of variance were used in the manip-

a discounted price and the average amount of price discount

SourceFrequency of pricepromotionsDepih of price discountsInteraction of promotionfrequency and depth ofprice discountsErrorTotal

d.f.33

9172187

SSH.067.592.3840.0361.06

MS3.692.53

.2 6.2 3

F15.8410.87

M 4

p-vatue(two'taitedi.0001.0001.3382

Next, we tested whether the relationships between thebrands' expected prices and each of the two independentvariables were nonlinear as hypothesized. Results of theF-tests for linearity show that we cannot reject the hy-potheses of linear relationships between the brand's ex-pected price and both the frequency of price promotionsand the depth of price discounts at the a = .10 level.However, the results of a multiple comparison by theTukey test indicate that, among all the contrasts of thebrand's expected prices at different price promotion fre-quencies, only the contrast between the second level (i.e.,3 promotions over 10 weeks) and the third level (i.e., 5promotions over 10 weeks) of the price promotion fre-quency is significant at a = .05. A plot of the averageexpected prices for the different levels of price promo-tion frequency also provides some directional support fora sigmoid relationship between the expected price andthe price promotion frequency (see Figure 1).

F i g u r e 1EXPECTED PRICE AS A F U N C T I O N OF THE FREQUENCY

OF PRICE PROMOTIONKxpectedPrice4.34.24,14,03.93.83.7

3.5Frequency of Price Prumotion


7/12

96 J O U R N A L O F AAARKETING RESEARCH, FEBRUARY 1Similarly, the testing of contrasts amon g different lev-els of price discounts reveals that only the difference inthe brand's expected price between the 30% and 40%discount levels is significant at a ^ .05 . Fu nhe r. thedifference of the differences in the brand's expected pricebetween the 30% and 40% price discount treatments andbetween the 10% and 30% price discount treatments issignificant at a = . 10. Again, a plot of the average ex-pected prices for the different levels of price discountsprovides some directional support for a concave rela-tionship between the expected price and the depth of pricediscounts (see Figure 2).

Price Expectations Mode l of Brand ChoiceFindings from the calibration of the price expectationsmodel representing the impact of price expectations onbrand choice are reported in Table 2. As shown in thefirst column, the brand preference variable is the moststatistically significant variable in the determination ofbrand choice, which is consistent with the fmdings of

previous research on modeling brand choice (e.g.,Guadagni and Little 1983). The retail price variable issignificant at a = .05 but the short-term promotion vari-able is not significant at conventional significance lev-els. For the impact of price expectations on brand choice,we find that both the price gain and the price loss vari-ables have the hypothesized signs, but only the effect ofprice loss is significant at the a = .01 level. Further, weobserve that the magnitude of the coefficient of the priceloss variable exceeds the corresponding value for the pricegain variable, a finding consistent with the loss aversionphenomenon reported in the research cited previously.For the thre shold par am eter s, 6^ and 6/, deline ating theboundaries of the price insensitivity region, we find that

F i g u r e 2EXPECTED PRICE A S A FUNCTION OF THE DEPTH O F

PRICE DISCOUNTSE'^xptP r4.34.24.14.03.93.83.73.63.5

ectedic e

-

-----

io *

i, , ^T - ,

' 1 ' .

^ ^ '

20% Mm ^ 40%De pth o f Pr ice D i iscounts

the values providing the highest adjusted U^ for the bchoice model are each equal to 6% of the brand's erage nonpromotional price. This Onding implies that pchanges of 5% or less of the brand's average nonpmotional price did not produce a significant changeconsumers' price perceptions. The finding of a latitof acceptable price differences around the brand's pected price is also consistent with the empirical fmdreported by Raman and Bass (1988) and Gurumurthy Little (1989).

A comparison of the goodness of fit of our pricepectations model with that of a traditional brand chomodel (see column 3 in Tabie 2) by the likelihood rtest yields a x ' value of 7.596 w ith 2 d.f., which is nificant at a = .025. We also estimated a price exptations model without threshold effects in which pgains and losses were measured as differences betwa brand's expected and retail prices. A test of non-neshypotheses (Ben-Akiva and Lerman 1985, p. 171-1between the price expectations model with threshold fects (shown in the first column of Table 2) and the pexpectations model withottt threshold effects shows the former provides a significantly better fit to the dat the a = .025 level. These fmdings suggest that incorporation of price expectations, with threshold fects in consumers" perceptions of price gains and loscan contribute significantly to the explanation of csumer brand choice behavior.Comb ined Price and Promotion Expectations Model Brand Choice

Lattin and Bucklin (1989. p. 229-300) report fmdifrom analyses of the IRI academic coffee dataset, sgesting that "consumer expectations about future pmotional activity are just as important to understandconsumer choice behavior as consumer expectationsprice." Though we did not set out to elicit expectatifor promotional activity as we did for price, our datapermit inference of consumers' promotion expectation the basis of their past exposures to promotional tivities. The question arises: How specifically do exptations of promotional activity affect consumer brchoice? We suggest that consumers who are exposedprice promotions on a brand to a degree beyond a pscribed threshold come to expect a discount every tithey buy that brand. When a brand does not offer a pmotional deal that a consumer has come to expect, refer to the choic e occasion as an "unfulfilled prom otexpectation event." Clearly, unfulfilled promotion pectation events will have an adverse effect on consumbrand purchase probability. Analogously, we define "unexpected promotion event" as the choice occaswhen a consumer encounters a price promotion on a brathat he or she did not expect because the brand has gaged in minimal promotional activity. Obviously, expected promotion events will affect brand purchprobability positively.

To model the impact of promotion expectations


8/12

PRICE AN D PR OMO TION EXPECTATIONS 97Table 2

CALIBRATION OF THREE ALTERNATE CONSUMER BRAND CHOICE MODELS

VariableCoefficient estimates'

Price expectationsmodel- .100 ( - .284).066 (6.049)'-1.268 (-1.758)^.685 (.8(M).571 (1.077)-1 .510 ( -2 .547) '

.5152.4692-63.169

-130.310

Price and promotionexpectations model.252 (.616).068 (5.950)'-1.223 (-1.633)".237 (.250).491 (.798)-.946 (-1.440)"

-1.589 (-2.168)'^.721 (1.128).5379.4765

-60.215-130.310

Traditionalmodel- .373 ( -1 .146).062 (6.096)''-1 .315 ( -1 .898) '1.460 (1.950)'

.4861.4554-66.967130.310

rand-specific consiamrand preference {PREF)etail price (P)romotion {PROM)Loss (9, = .06)Unfulfilled promotionexpectations eventUnexpected promotioneventAdjusted U^L^ig likelihood(at convergence)Log likelihood(at zero)

'Entries in parentheses are ihe /-statistics for the estimated eoefftcienis.''Significam at a = ,01 in the one-tailed asymplotic /-test.'Significant at a = .05 in the one-tailed asymptotic f-test."significant al a = .10 in the one-tailed asymptotic /-test.

)if brand i promotes 50% or moreof the time as observed by consumer it,otherwise.

ote that the promotion frequency threshold of 50% ormore of the time is adopted in the operationalization be-ause it provides the highest adjusted U' for the brandhoice model in a grid search. A similar result is re-ported by Lattin and Bucklin (1989)they fmd that whenonsumers observe a brand to be on price promotion morethan half of the time, they come to expect a price dis-ount on it and arc reluctant to buy it when it is not pricepromoted. In this connection, to represent the effects ofpromotion expectations on brand choice in our price ex-pectations model (equation 2). we operationalize (I) anunfulfilled promotion expectation event variable as ETP^k= I and PROM,I, ='- 0 and (2) an unexpected promotionevent variable as ETP,^ - 0 and PROM^^ 1. Thus wean capture the potential asymmetric effects on brandhoice of not getting an expected promotion and gettinga surprise or unexpected promotion.The findings from the calibration of the combined priceand promotion expectations model are reported in thesecond column of Table 2. A likelihood ratio test of thecombined price and promotion expectations model againstthe price expectations model yields a x' value of 5.908with 2 d.f. of freedom and is significant at the a = .10level. Thus, the incorporation of the promotion expec-tation effects is found to contribute significantly to the

ability to explain brand choice above and beyond themodeling of just the price expectation effects. All vari-ables in the combined price and promotion expectationsmodel have the hypothesized signs. However, of the twopromotion expectation effects arising from the unful-filled promotion expectation and unexpected promotionevents, only the effect of the unfulfilled promotion ex-pectation event is significant at a = .05. Because thecoefficient estimate for the unfulfilled promotion expec-tation event variable bas a larger absolute magnitude anda larger lvalue than the ctwfficient for the unexpectedpromotion event variable, we conclude that, as in thecase of price expectations, for promotion expectationslosses loom larger than gains. This fmding is consistentwith prospect theory (Kahneman and Tversky 1979, p.279) , which states that the "value function for losses issteeper than the value function for gains." It suggeststhat though offering frequent price promotions will in-crease the chance of consumers observing unexpectedpromotions, which then may produce short-tenn salesgains, those sales increases may be offset by the saleslosses resulting from consumers not getting promotionsthey have come to expect.

Our individual-level data enable us to track purchasebehavior, in the presence and absence of a price pro-motion, separately for respondents whom we posit havecome to expect a promotion versus those whom we posithave not. Recall that, on the basis of a grid search, thepromotion frequency threshold at which the respondentsare posited to form promotion expectations is 50% ormore of the time. We now compare the impact of thepresence of a price promotion on the respondents who


9/12

98 JOURNAL OF MARKETING RESEARCH, FEBRUARY 1have come to expect a promotion on the brand with theimpact on those who have not. Similarly, we also com-pare the impact of the absence of a price promotion onthe respondents who have come to expect a promotionwith the impact on those who have not. Table 3 showsaverage purchase probabilities with and without pricepromotions when a promotion on the target brand is ex-pected and when it is not expected. We see that 59% ofthe 51 respondents who were offered a price promotionon the target brand and expected a promotion on it pur-chased the brand in contrast to 79% of the 42 respon-dents who were offered a price promotion on the targetbrand but did not expect it. The difference in the pur-chase probabilities is significant at a = .05. Correspond-ingly, the absence of a price promotion on the targetbrand is found to cause a larger reduction of brand choiceprobability among respondents who expected a promo-tion on the target brand than among those who did notexpect a promotion. Specifically, only 14% of the re-spondents who expected a promotion on the target brandpurchased it without a price promotion in contrast to 40%of the respondents who did not expect a price promotionon the target brand. The difference in the purchase prob-abilities is significant at a = .0 1 . In sum , these findingsreinforce our inference from the calibration of the com-bined price and promotion expectations brand choicemodel (see column 2 of Table 2) of a loss aversion phe-nomenon in the case of promotion expectations. There-fore, we conclude that consumers' response to the lossof an expected promotion is stronger than their responseto a surprise or unexpected promotion.

CONCLUSIONSW e set out to inves tigate the im pact of different pricepromotion schedules on brands' expected prices. Ourfindings from the tests of the two hypotheses pertainingto the impact of exposure to price promotions on con-sumers' price expectations reveal that the impact is sig-nificant. However, our experimental findings do not

Table 3CLASSIFICATION OF BRAND PURCHASE PROBABILITIES

OF THE TARGET BRAND WITH AND WITHOUT PRICEPROM OTIONS BY PROM OTION EXPECTATIONSPromotionexpectations'

Expecta promotionDo not expecta promotion

'In this study, rcsptindents are assumed to expecl a brand to beavailable an price promotion if they have seen il offered on pricepromoUon 50% or more of the time.""A total of 30 of the 51 resp ondents wh o were offered a price pro -motion that they expected purchased the brand.

Promotionavailable5 9 % '(=30 of 5l>79 %(=33 of 42)

Promotionnot available14%(=6 of 42)4 0%( = 21 of 53)

suggest rejection of the hypotheses that the brand's pected price is a linear function of the price promofrequency and the depth of price discounts at convtional significance levels. Nevertheless, the results pvide .some directional support for nonlinear relationshbetween the expected price and the two elements oprice promotion schedule. Given the important imcations of such potential nonlinear effects of price pmotions on brands' expected prices, further reseatesting those nonlinear effects of price promotions shoprove fruitful for the design of optima l price prom otpolicies.

We find evidence in support of a region of price sensitivity around a brand's expected price within whprice changes do not produce a significant changeconsumers' price perceptions. Price differences outsthat region, in contrast, are found to have a significimpact on consumer brand purchase probability. Fther, we observe that the effects of price gains mayoffset by the effects of price losses when the price pmotion is retracted because losses loom larger than gaThe former finding on the region of price insensitivaround a brand's expected price implies, as Gurumurand Little (1989, p. 21) note, "marketers wishing to crease prices should nibble, not bite. Small price creases are less hazardous if they stay within the latitof acceptance." The latter finding on the loss aversphenomenon suggests that marketers of products suchcoffee and sugar, who experience wide fiuctuationsraw commodity costs, should attempt to .smooth out pfiuctuations instead of passing along entire increases decreases In commodity costs to consumers. Our fiings suggest that the sales gains realized from a prdecrease that is outside the region of price insensitivmay nol compare favorably with the sales losses incuras a result of an equal price increase.

Our findings on promotion expectations suggest tunfulfilled promotion expectation event.s among csumers who have come to expect promotions on a brabecause of frequent exposure to them will have an verse impact on the brand. Analogously, unexpecpromotion events wili enhance the probability of pchasing a brand among consumers who have not bexposed to many price promotions and therefore do as a rule expect the brand to be available on a promtional deal. We suggest that those results are consistwith the rational expectations view that "any policy rthat is systematically related to economic conditions, example, one observed with stabilization in mind, wbe perfectly anticipated, and therefore have no effect output or employment" (Maddock and Carter 1982,43) . Policy actions that come as a surprise to people,contrast, will generally have some real effect. Cleathe design of optimal price promotion schedules requiconsideration of the fact that an increase in the useprice promotions could erode long-term consumer mand by lowering the prices that consumers anticippaying for the brand. Price promotional deals may co


10/12

AND PROMOTION EXPECTATIONS 99

to be "perfectly anticipated" and have much less impacton consumer response than they do when they come asa surprise to consumers.

Evaluation of the tradeoff between the short-term salesgain from a price promotion and the adverse effeel onfuture sales because of consumers forming price and pro-motion expectations requires knowledge of how pricepromotions affect the formation of consumers" expec-tations under different market conditions. We cannot, fromthe findings of our study, make more definitive recom-mendations about the optimal price promotion schedule.Such recommendations must come from analysis that ex-plicitly considers cost data and information about com-petitive promotional activities.

Our study and the others cited previously have begunto atte.st to the important role of price and promotionexpectations in determining consumer purchasing behav-ior. Therefore it is critical for us to study how price andpromotion expectations are formed. Experimental stud-ies such as ours, we believe, afford an opportunity toconduct systematic investigations of how such expecta-tions are formed. In our study, the direct measurementof brands' expected prices enabled us to examine the im-pact of different price promotion schedules on price ex-pectations. However, we did not measure consumers'promotion expectations and had to rely on a grid searchprocedure to infer the promotion frequency threshold atwhich consumers come to form promotion expectations.In future research, a probabilistic measure of consumers'promotion expectations should be employed to obtainfurther insights into the formation of promotion expec-tations and their impact on brand choice behavior.

REFERENCESAronson. Elliot. Marilynn Brewer, and J. Merrill Carlsmith(1985), "Hxperinienlation in Social Psychology," TheHandhooti of Sociat Psychology. Gardner Lindzey and ElliotAronson, eds. New York: Alfred A. Knopf, Inc. Chapter8.Ben-Akiva. Moshe and Steven R. Lcmian (1985) . DiscreteChoice Analysis: Theory and Application toTravel Demand.Cambridge. MA: MIT Press.Berkowitz, Eric N, and John R. Walton (1980). "ContextualInfluences on Consumer Price Responses: An ExperimentalAnalysis," Journal of Marketing Research. 17 (August), 349-58.Curhan. Ronald C. and Robert J, Kopp (1986), "Factors In-

fluencing Grocery Retailers ' Support of Trade Promotions."Report No. 86-104. Boston: Marketing Science Institute.Delia Bina. Albert J. and Kent B. Monroe (1980), "A Mul-tivariate Analysis of the Perception of Value Erom RetailPrice Advertisements," m Advances in Consumer Research,Vol. 8, Kent B. Monroe, ed, Ann Arbor. Ml: Associationfor Consumer Research, 161-5.Dodson, Jtie A., Alice M, Tybout, and Brian Sternthai (1 978),"Impact of Deals and Deal Retraction on Brand Switching,"Journat of Marketing Research. 14 (February), 7 2 - 8 1 ,D(K)b, Anthony N,, J, Merrill Carlsmith. Jonathan L. Freed-man. Thomas K. Landauer . and Saleng Tom. Jr. ( 196 9) ,

"Effect of Initial Selling Price on Subsequent Sales ," Jour-na t of Personatity and Social Psychology. 11 (4). 3 4 5 - 5 0 .Doyle, Peter and John Saunders (1985), "the Lead Effect ofMarketing Decis ions ," Joumal of Marketing Research. 22(February) 54-65.Guadagni, Peter M. and John D. C, Little (1983). "A LogitModel of Brand Choice Calibrated on Scanner Data," Mar-keting Science. 2 ( S u m m e r) . 2 0 3 - 3 8 .Gurumurthy , K. and John D. C. Little (1989), "A Price Re-sponse Model Developed From Perception Theories," work-ing paper, Sloan School of Management. MIT.Helson, H, (1964), Adaptation-Level Theory. New York: Har-pe r & Row Publishers. Inc.Johnston, J. (1984) , Econometric Methods. 3rd ed, McGraw-Hill Botjk Company.Kahneman, Daniel and Amos Tversky (1979), "Prospect The-ory: An Analysi.s of Decision Under Risk," Econometrica,47 (March), 2 6 3 - 9 1 ,Kalwani, Manohar U., Chi Kin Yim, Heikki J, Rinne, andYoshi Sugita (1990), "A Price Expectations Mtxlel of Cus-tomer Brand Choice." Journal of Marketing Research. 27(Au g u s t ) , 251-6 2 .Klein. Noreen M, and Janet E. Oglethorpe (1987). "CognitiveReference Points in Consumer Decis ion Making." in Ad-vances inConsumer Research, Vol. 14, Melanie Wallendorfand Paul Anderson, eds. Provo, UT: Association for Con-sumer Research . 183-7 .Lattin, James M. and Randolph E. Bucklin (1989). "Refer-ence Effects of Price and Promotion on Brand Choice Be-havior ," Journal of Marketing Research. 26 (August) . 299-310.Maddock, Rtxiney amd Michael Carter (1982), "A Child's Guideto Rational Expectations," Journat of Economic Literature,20 (March) , 3 9 - 5 1 .McEadden. Daniel (1973), "Conditional Logit Analysis ofQualitative Choiee Behavior." in Frontiers in Economet-rics, P. Zarembka, ed. New York: Academic Press, Inc.Monroe. Kent B. (1973). "Buyers' Subjective Perceptions ofPriee," Journat of Marketing Research. 10 (February) . 70-80 . (1977), "Objeetive and Subjective Contextual Influ-ences on Price Perception," Consumer and Industrial Buy-ing Behavior. Arch G. Woodside, Jagdish N. Sheth, andPeter D. Bennett, eds. NewYork: North-Holland PublishingC o m p a n y . 2 8 7 - 9 6 .Muth. John E. (1961), "Rational Expectations and the Theoryof Price Movement," Econometrica. 29 ( Ju ly ) , 3 15-3 5 .Nelson, Clarence W, (1977) . "Rational Expectat ionsFreshIdeas that Challenge Some Established Views of PolicyMaking," special report. Federal Reserve Bank of Minne-apolis,Neslin, Scott A. and Robert W. Shoemaker (1989) . "An Al-ternative Explanation for Lower Repeat Kates After Pro-motion Purchases," Journal of Marketing Re.search. 26(May) , 2 0 5 - 1 3 .Putler. Daniel S, (1990). "Incorporating Reference Price Ef-fects Into a Theory of Consumer Choice." working paper .Economic Research Service. USDA,Puto, Christopher P. (1987), "*The Framing of Buying Deci-sions," Journal of Consumer Research. 14 (December) , 30 1 -15 .Raman. Kalyan and Erank M, Bass (1988) , "A General Testof Reference Priee Theory in the Presence of Threshold Ef-


11/12

100 JOURNAL OF MARKETING RESEARCH, FEBRUARY 1fects," working paper. College of Business Administration,Universiiy of Florida.Rinne, Heikki J. (1981). "An Empirical Investigation of theEffects ol Reference Prices on Sale s." unpublished PhD dis-sertation, Purdue University.Rowe. Debra and Christopher P. Puto (1987), "Do Con-sumers' Reference Points Affect Their Buying Decisions?"Advances in Consumer R esearch, Vol. 14, Melanie Wal-lendorf and Paul Anderson, eds. Provo. UT; Association forConsumer Research , 188-92.

Sawyer. Alan G. and Peter H. Dickson (1984). "PsychologicalPerspectives on Consumer Response to Sales Promotion."Research on Sates Promotion: Collected Papers. Report No.84-104, Katherine E. Jocz, ed. Boston: Marketing ScienceInstitute. 1-21.Sherif, Carolyn W. (1963), "Social Categorization as a Func-tion of Latitude of Acceptance and Series Range." Journalof At^normal Psychology, 6 7 . 1 4 8 -5 6 .Shoemaker, Robert W. and F. Robert Shoaf (1977), "RepeatRates of Deal Purchases." Journat of Advertising Research,17 (April). 4 7 - 5 3 ,

Thaler, R. (1985), "Mental Accounting and Consumer ChoiMarketing Science. 4 (Summer) . 199-214.Tversky. Amos and Daniel Kahneman (1974), "Judgment der Uncertainty: Heuristics and Biases," Science, 185, 113 ! .Uhl. J. N. and Harold L. Brown (1971). "Consumer Perction of Experimental Retail Food Priee Changes," Jourof Consum er Affairs. 5 (Win te r ) . 1 7 4 -8 5 .Winer. Russell S. (1985). "A Price Vector Model of Demfor Consumer Durables: Preliminary Development," Mtaeling Science. 4 (Winter) , 74-90.(1986). "A Reference Price Model of Brand Chotor Frequently Purchased Products." Journat of ComuResearch. 13 (September) , 25 0- 6 .Zeithaml. Valarie A. and Karen L. Graham (1983). "The curacy of Reported Reference Prices for Professional Svices ," Advances in Con.sumer Research, Vol. 14. RichP. Bagozzi and Alice M. Tybout. eds. Ann Arbor, MI: sociation for Consumer Research, 6 0 7 - 1 1 .Rcprini No. JMIU9

You ccai h a v eit all.. .W hat you 're readin g now plusthe impo rtant articles in 800 other business a n dm anag em ent m agazines, in a matter ofminutes.

The ABI/INFORM^*" business database and yourcomputer give you acce ss to article sum ma riesfrom m aga zines worldwide.Ca ll the publishers of ABI/INFORM at 800/626-2823, today

Vr .


12/12

Consumer Price and Promotion Expectations

Documents