A Regret Theory Approach to Assessing Customer Satisfaction When Alternatives Are Considered

ABSTRACT - Contrary to the findings of behavioral science and decision-making theory, many market researchers assume that the customer satisfaction rating depends solely on the quality of the product that is actually consumed. The consumer¦s conception of the fitness of alternative products is mostly neglected. The Aregret@ theory delivers an explanation for the phenomenon that alternative choices are re-evaluated during the post purchase phase and thus influence the satisfaction rating. Hence the confirmation/disconfirmation model of customer satisfaction has to be enlarged by the expectations about the alternative choices. This causal model of satisfaction is verified by means of an empirical study in the travel industry.

Citation:

Andreas Herrmann, Frank Huber, and Christine Braunstein (1999) ,"A Regret Theory Approach to Assessing Customer Satisfaction When Alternatives Are Considered", in E - European Advances in Consumer Research Volume 4, eds. Bernard Dubois, Tina M. Lowrey, and L. J. Shrum, Marc Vanhuele, Provo, UT : Association for Consumer Research, Pages: 82-88.

European Advances in Consumer Research Volume 4, 1999      Pages 82-88

A REGRET THEORY APPROACH TO ASSESSING CUSTOMER SATISFACTION WHEN ALTERNATIVES ARE CONSIDERED

Andreas Herrmann, University of Mainz, Germany

Frank Huber, University of Mainz, Germany

Christine Braunstein, University of Mainz, Germany

ABSTRACT -

Contrary to the findings of behavioral science and decision-making theory, many market researchers assume that the customer satisfaction rating depends solely on the quality of the product that is actually consumed. The consumer¦s conception of the fitness of alternative products is mostly neglected. The "regret" theory delivers an explanation for the phenomenon that alternative choices are re-evaluated during the post purchase phase and thus influence the satisfaction rating. Hence the confirmation/disconfirmation model of customer satisfaction has to be enlarged by the expectations about the alternative choices. This causal model of satisfaction is verified by means of an empirical study in the travel industry.

ASSESSING CUSTOMER SATISFACTION WHEN ALTERNATIVES ARE CONSIDERED

MarketingBdefined as a market-oriented management conceptBplaces the problems and wishes of actual and potential customers at the cente of all considerations. According to this principle, the satisfaction of consumers’ wants is what lays the foundation for market success. The requirements of the buyers must form the basis for a supplier’s marketing activities, since it is the response of the market that in the final analysis determines a firm’s success or failure (Urban, Hauser 1993).

Following this line of argument, the challenge facing any company is thus to achieve the highest possible level of customer satisfaction. The significance of the satisfaction rating is that it can serve as an indicator of future sales. Empirical studies have proved that customers who are satisfied with a purchased product will buy the same product again, and often also practise word-of-mouth advertisingBa form considered to be particularly credible. The costs of doing business with a buyer with whom a relationship has already been built up are moreover significantly lower than those entailed by canvassing a new customer (Fornell 1992, Anderson, Fornell 1994). Furthermore, a study by Anderson (1996) reveals that sensitivity towards price rises declines the more satisfied the customer is (Krishnamurthi, Raj 1991).

In view of the considerable relevance of customer satisfaction for the success of a company, it comes as no surprise to find that a large number of marketing studies are devoted to measuring customers’ ratings of the fitness (quality) of corporate performance. The scope of these analyses ranges from simple mean-value calculations and variance analyses to complex causal models based on the notion that customer satisfaction must always be interpreted as a multidimensional phenomenon, linked upstream and downstream to hypothetical constructs, such as the perceived quality of a performance and customer retention (Anderson, Fornell, Lehmann 1994, Johnson 1998).

The following problem is encountered by practically all studies, regardless of the approach adopted for measuring customer satisfaction ratings: contrary to the findings of both behavioural-science and decision-making theory, many market researchers assume that the customer satisfaction rating depends solely on the quality of the commodity or service that is actually consumed (Fornell, Johnson, Anderson, Cha, Everitt Bryant 1996). The individual’s conception of the fitness of alternative products for their intended purpose merely plays a subordinate role in analyses of the satisfaction ratings accorded to products that have actually been purchased and consumed. On the other hand, we know from empirical studies that expectations regarding the fitness of alternatives are most definitely capable of influencing consumers’ judgements of purchased products. Taylor (1997, p. 230) has a very succinct description for this state of affairs: "... specifically we posit that the expectations about the unchosen alternatives affect satisfaction with one’s choice ...". Two examples serve to illustrate this idea (Gilovich, Medvec 1995):

* A car owner is requested to rate his satisfaction with regard to the quality of a recently purchased vehicle (e.g. Mercedes C Class). Although the automobile in question is unrestrictedly operational, the buyer considers the unpurchased alternative (e.g. BMW 3 Series) to have significant advantages. When formulating his satisfaction rating, he will therefore also take account of this rival car which he did not purchase, but which he nevertheless finds attractive. It thus appears an almost impossible task for the car manufacturer concerned to satisfy the customer completely, despite offering an impressive range of products and services.

* An individual decides to spend his annual holiday at a particular location (e.g. Majorca). Before reaching his final decision, he also considered another destination (e.g. Crete) as an equivalent alternative in every respect. When he actually arrives in Majoca, the quality of the accommodation leaves something to be desired, though all the other services come up to his expectations without reservation. After his return, the holiday-maker sends an extremely forthright letter of complaint to the travel operator and is annoyed with himself for not having chosen Crete as his holiday destination instead.

In the first example, the expectations concerning the unpurchased car exert a considerable influence on the satisfaction rating given to the purchased car. The buyer’sBpossibly highly exaggeratedBconceptions with regard to the fitness of the other option for its intended purpose evidently play a central role. There are consequently grounds for supposing that satisfaction with the consumed product declines the higher the expectations placed in the alternative. The second example describes a defective service, which moves the holiday-maker to issue a protest and to enter into a dispute on the basis of what he subjectively feels to be a wrong decision. In the eyes of this consumer the perceived shortcoming in the performance appears to enhance the presumed fitness of the unchosen option. We can thus reasonably assume that the expectations regarding the alternative exert an especially great influence on the satisfaction rating accorded to the purchased product in the event that this product exhibits a performance deficit (Zeithaml, Parasuraman, Berry 1990, Herrmann, Huber, Gustafsson 1997).

In the following we shall attempt to discuss the implications for marketing policy that result from this empirical phenomenon. In particular, we shall focus on the question of the extent to which expectations regarding the alternative influence satisfaction with the consumed commodity. The following approach is adopted: first of all, the fundamental principles of the "regret" theoryBthe basis for all further observationsBare outlined. Next, hypotheses are derived for describing the formation of a customer satisfaction rating when alternatives are considered. The formulated hypotheses are then verified by means of an empirical study, whereby our attention is concentrated above all on the relevance of alternative products to the fitness rating given to the consumed product. Finally, the implications of the study findings for marketing policy and for research into purchasing patterns are discussed.

AN APPROACH FOR ASSESSING CUSTOMER SATISFACTION WHEN ALTERNATIVES ARE CONSIDERED

Socio-psychological theories provide an explanation, among other things, for the formation of judgements about experiences with consumed products on the basis of various cognitive processes. Interest is focused on what a person actually experiences when purchasing and consuming a product. The cognitive dissonance theory, which is concerned with evaluating product consumption events, provides particularly helpful clues. In his cognitive dissonance theory, Festinger (1957) postulates the existence of cognitive tensions that develop as a result of striving to achieve harmony and consistency. A system founded on individual psychological forfeits its stability because doubt is cast on the relationships that exist between certain system elements as a result of contradictory information or unfulfilled expectations. Immediately following the purchase or consumption act, the alternative choice that was originally considered is re-evaluated to the detriment of the rejected products. If these rejected products also exhibit important merits, cognitive tensions will be produced in the consumer. They are expressed in the individual’s misgiving that it might have been better to plump for one of the alternatives instead. This is especially true if the decision entailed a substantial amount of effort, in other words if it had a certain uniqueness about it. In addition, cognitive tensions are particularly common following the purchase of a very expensive or difficult-to-obtain poduct. Empirical studies have moreover shown that the likelihood of cognitive dissonances increases the greater the number of alternatives, the greater the attractiveness of the rejected options and the greater the similarity between the alternative commodities (Peter, Olson 1990). Since tensions of this kind occur at various stages in the purchase and consumption process, the approach developed by Festinger represents a universal theory for describing, explaining and predicting cognitive dissonances.

This work concentrates, however, on one specific anomaly in the formation of customer satisfaction ratings when alternatives are considered during the post-purchase phase. The "regret" theory, an approach derived from general utility expectation theory, is a suitable instrument for examining this phenomenon (Bell 1982, Loomes, Sugden 1982). Compared with the cognitive dissonance theory, the regret approach enables the phenomenon in question to be comprehended more thoroughly and simplifies the process of deriving recommendations for action (Loomes/Sugden 1987a).

The basic idea of the regret approach, which was developed by Bell and by Loomes/Sugden independently of one another, is described by Loomes (1988, p. 464) as follows: "The psychological intuition behind regret theory is that if an individual chooses A (and therefore rejects B) and state j occurs, the overall level of satisfaction he experiences will depend not simply upon x_Aj, but also upon how x_Aj compares with x_Bj. If what he gets is worse than what he might have had, it is suggested that the satisfaction associated with x_Aj will be reduced by a decrement of utility due to regret ...". By this principle, an individual compares the consequences associated with the various options before choosing a particular alternative (Loomes, Sugden 1987b, Harless 1992). According to the postulate of the utility expectation theory, he chooses the product he anticipates will bring him the greatest utility. After he has experienced the product, this person then evaluates the consequences that have resulted from its purchase and consumption. In order to arrive at this judgement, he undertakes a comparison of his actual experience and the consequences that, in his view, would have arisen from purchasing and consuming the alternative instead. If his experience with the actually purchased commodity surpasses the presumed consequences of consuming the alternative, the individual will be satisfied with his decision. If, on the other hand, he believes that the consequences he has actually experienced are inferior to those that would have resulted from purchasing and consuming the rejected options, he will regret his decision (Bell 1983). In this case, the feeling will set in that to have chosen one of the other alternatives would probably have been a better option.

This approach is taken as a basis in the following for explaining the formation of a satisfaction rating when alternatives are considered. Before we can attempt to analyse the formation of the satisfaction rating in detail, however, with particular emphasis on the consideration of alternatives, we must expound the concept of satisfaction a little further: according to Anderson (1994, p. 20), a satisfaction rating can be described as a decision regarding the fitness of a chosen product for its intended purpose. He postulates that "... consumer satisfaction is generally construed to be a postconsumption evaluation dependent on perceived quality or value, expectations, and confirmation/disconfirmationBthe degree (if any) of discrepancy between actual and expected quality ...". It thus follows that (dis)satisfaction is the result of a complex information-processing process, which essentially consists of a desired/actual comparison of a consumer’s perception of a product or service (actual) and his expectations with regard to its fitness for its intended purpose (desired). The congruence or divergence yielded by this comparison between the perceived product quality and the anticipated quality is expressed as (dis)confirmation (Anderson, Sullivan 1993). Since this construct is directly linked to the target variable, confirmation of these xpectations results in satisfaction, while their disconfirmation leads to dissatisfaction (Oliver, DeSarbo 1988).

Whether or not an individual considers his expectations to have been confirmed by a purchaseBand thus whether or not he is satisfied with the performance of the supplierBis dependent above all on perceived quality (Johnson, Gustafsson 1997, Zeithaml 1988). Quality perception is directly linked to the purchase and consumption experience, and can be characterised as a consumer’s global judgement regarding the fitness of a product for its proper purpose. The person concerned evaluates each of the attributes of the purchased product he considers to be relevant in relation to their intended application, and then combines the partial ratings in accordance with a defined rule to form an overall quality judgement. The buyer’s expectations represent a specific quality standard he hopes to find in a particular commodity. He uses them as a yardstick for assessing the fitness (quality) of the perceived performance. The level of expectation is determined firstly by previous consumption experiences, in other words by past encounters with the product in question. SecondlyBand this is especially true if the product is being purchased and consumed for the first timeBthe consumer derives his conception of the product’s quality from the price of the available alternatives, as well as from various other kinds of advance information.

If we accept this view, satisfaction can be interpreted as the result of a comparative process, in which the "desired" component serves as a measure for evaluating perceptions of a given situation. In addition, there are grounds for presuming the existence of an asymmetrical relationship between (dis)confirmation and (dis)satisfaction with a particular performance (Simonson 1992). This means that, in the event of his expectations being disconfirmed, a customer’s dissatisfaction is considerably greater than his satisfaction should the anticipated product quality be confirmed to exactly the hoped-for degree.

This generally accepted customer satisfaction concept can be extended with the aid of the regret theory. The central idea of this modified approach is that satisfaction with a product or service is dependent not only on its fitness for its intended purpose, but also on the anticipated quality of the rejected alternative. Consequently, an individual will be satisfied with a consumed product, providing its perceived quality corresponds to the expected quality and his experience with it surpasses the consequences he presumes would have arisen had he purchased and consumed the rejected option (Ritov, Baron 1995). If, on the other hand, the person concerned considers his experience with the consumed commodity to be inferior to the consequences he associates with purchasing and consuming the alternative, he will end up dissatisfied regardless of his judgement about the fitness of the purchased and consumed product (Imman, Dayer, Jia 1997). This applies equally if the product proves to be fully functional and coincides with his preconceptions. The first hypothesis can thus be formulated as follows:

* The higher the expectations placed in the alternative, the lower the level of satisfaction with the consumed product, assuming the product quality to be identical.

A glance at the example described earlier is sufficient to illustrate this hypothesis. To recapitulate, the buyer of the Mercedes C Class believes that the unpurchased alternative (the BMW 3 Series) has significant advantages. When formulating his satisfaction rating, he therefore also takes account of the BMW car he did not purchase, but which in his estimation is extremely attractive. If this alternative did not exist, the customer would base his rating solely on the fitness of the Mercedes for its intended purpose and proclaim his satisfaction with this vehicle.

In addition, Ritov and Baron (1995) postulate the existence of an asymmetry in the formation of customer satisfaction ratngs when alternatives are considered. It is their view that presumed superior or inferior consequences of rejected options do not influence satisfaction to the same extent as actual experience with the consumed product. If a person’s experience with a consumed product matches his expectations of it, the consequences that are presumed to arise from purchasing and consuming the alternative are unlikely to play a central role in the formation of his satisfaction rating. On the other hand, the expectations regarding the fitness of the rejected option for its intended purpose will be crucial to the formation of a judgement if the buyer’s conceptions of the actually purchased and consumed product are not fulfilled. This brings us to the second hypothesis:

* If the expectations placed in the consumed product are (not) confirmed, the expectations with regard to the alternatives will have a moderately strong (very strong) influence on satisfaction.

The second of our two examples illustrates this hypothesis effectively. It centres around an individual who chooses Majorca as his holiday destination and rejects the other option (Crete). Since it emerges during the course of his holiday that the quality of the accommodation does not come up to his expectations, this customer regrets his wrong decision and proclaims the excellence of Crete as a travel destination. If he had not been so disappointed with the quality of the accommodation in Majorca, the holiday-maker would probably not have been annoyed about his choice of destination, nor would he have stressed the benefits of Crete in this way.

EMPIRICAL STUDY

The hypotheses deduced above on the basis of the regret theory concerning the formation of a satisfaction rating when alternatives are considered must now be verified by means of an empirical analysis. The central aim of the study conducted for this purpose in the travel industry was to determine the effect of the presumed attractiveness of a rejected alternative on satisfaction with the actual holiday destination. According to the findings of behavioural science, whereby the correlation between the two alternative choices can only be determined by studying them in relation to other relevant variables, a role is also played by expectations regarding the chosen travel destination, previous experience with this resort and the confirmation/disconfirmation of these expectations.

The work carried out by Anderson (1994) gives us grounds for supposing that the expectations placed in a travel destination have a negative influence on their confirmation. The higher a consumer’s expectations about the individual dimensions of a product or service, the more difficult it is for the supplier to meet them. Previous experience with the chosen holiday resort, on the other hand, has a positive effect on the confirmation of expectations. The fitter the individual dimensions of the consumed product or service, the more likely the customer is to consider his expectations fulfilled. According to the confirmation/disconfirmation paradigm discussed above, the confirmation of expectations about the chosen travel destination has a positive effect on satisfaction with this resort. The more closely the offered product or service corresponds to the holiday-maker’s expectations, the better will be his satisfaction rating. Finally, the regret theory asserts the existence of a negative correlation between expectations about the alternative travel destination and satisfaction with the resort that is actually chosen (Larrick, Boles 1995). The higher the expectations placed in the individual dimensions of the unchosen offer, the lower will be the level of satisfaction with the product or service that is actually consumed.

Empirical studies have shown that the multitude of variants of the attribute leels render it extremely difficult to operationalise and measure these constructs. As far as the object of our investigation is concerned, the aim was to determine which feature dimensions provide the basis for the formation of the satisfaction rating. The preliminary study which was conducted for this purpose revealed that the expectations placed both in the chosen holiday destination and in the alternative refer above all to the accommodation, the catering and the surrounding programme of activities. When modelling the hypothetical 'experience’ construct, it seemed reasonable to use the same indicators (i.e. accommodation, catering and programme of activities). The 'satisfaction’ construct was operationalised by means of two indicatorsBglobal satisfaction and plans to visit the same resort again. The latent 'confirmation of expectations’ variable, on the other hand, was defined identically to its indicator.

The data was collected by a German travel operator in collaboration with a firm of marketing consultants. A total of 408 customers in the Munich area were interviewed between July and September 1996, both before and after their holiday. All the respondents had booked a package trip to a tourist island in southern Europe. Directly prior to their departure, the interviewees were asked to formulate their expectations concerning the chosen travel destination and the alternative resort. 'Accommodation’, 'catering’ and 'programme of activities’ were used as the indicators, each on a scale from 1 to 8. Directly following their return from holiday, the respondents then had to provide information about their experience with the chosen travel destination, the extent to which their expectations had been confirmed and their satisfaction. The same indicators were used again for this part of the survey, once more on a scale from 1 to 8.

The causal model of satisfaction with the chosen travel destination takes account of the explanatory variables 'expectations about the chosen travel destination’, 'expectations about the alternative travel destination’, 'experience with the chosen travel destination’ and 'confirmation of expectations about the chosen travel destination’. The relationships that exist between these dimensions can only be analysed using a method which allows account to be taken of such hypothetical constructs in the form of latent variables (Bagozzi 1995, Bagozzi, Baumgartner 1995). This requirement is met by linear structural equation models such as LISREL and EQS (Bollen 1989, Bentler 1993), which in addition allow us to verify an entire system of hypotheses based on theoretical and objective/logical observations. Such methods also have the advantage that they enable measurement errors to be modelled explicitly and do not presuppose any mutual independence of the explanatory variables (Hair, Anderson, Tatham, Black 1995).

Version 8 of the LISREL software enables the parameters to be estimated with the aid of the Maximum Likelihood method. This method is held to be particularly effective whenever (as in our case) only a small sample is available and there is no indication of any values with a normal frequency distribution (Fornell, Tellis, Zinkhan 1982, Bagozzi, Phillips 1991). Beside the Goodness of Fit-Index and the Adjusted Goodness of Fit-Index (Kaufman, Dant 1992), with a value of 0.96 and 0.92 respectively, the following fit measures are considered: The Root-Mean-Square Error of Approximation (RMSEA) (Steiger 1990) of 0.067 was determined for our study, what Browne and Cudeck (1993) and J÷reskog and S÷rbom (1993) consider as a fairly good model. Some reference works also recommend the Normed Fit Index (NFI) (Bentler, Bonett 1980). Since our study yielded a value of 0.95, we can claim a satisfactory approximation (Bollen 1989). In addition to the overall model, the approach was also verified with regard to the reliability of its indicators and the validity of its convergence. In the case of the first of these detailed criteria, all the values exceeded the level of 0.4 suggested by market researchers. The calculation of the convergence validity yielded results higher than the value of 0.5 consideredto be adequate for each individual construct.

Figure 1 shows the results of the parameter estimate. The greatest influence on satisfaction with the chosen travel destination is exerted by the confirmation of expectations about this resort (0.48), whereby experience with the actual destination in turn has the greatest effect on the confirmation of expectations (0.47). In the opposite direction, however, the expectations placed in the chosen holiday resort also influence the consumption experience (-0.20), a correlation which has already been backed up by numerous empirical studies. Furthermore, the expectations regarding the chosen resort have a negative effect on their confirmation (-0.35). It is moreover striking to note that the expectations about the alternative travel destination exert both a direct and an indirect influence on satisfaction with the chosen resort. 'Direct’ means that an increase of one unit in the expectations placed in the alternative leads to a decrease of 0.14 units in satisfaction, assuming that all other factors remain unchanged. The fact that the expectations regarding the unchosen option also have an effect on the expectations about the alternative destination, and consequently an indirect influence on satisfaction, has however proved to be significant (0.01; -2.39). An increase of one unit in the expectations placed in the alternative travel destination leads to a rise of 0.16 units in the expectations regarding the chosen resort, assuming that all other factors remain unchanged. This positive correlation expresses a facet of individual behaviour that has also been revealed by other empirical studies: after deciding in favour of a particular commodity, people adhere to their conviction of having chosen the best of all possible options, at least until they come to use or consume it. If expectations about the rejected alternative are increased as a result of new information, the expectations regarding the chosen commodity will also rise. It is not until the commodity has been used or consumed that the individual begins to believe he would have been better off choosing the other product instead. As one can see, there is a relationship between these two constructs, nevertheless based on computations of the discriminant validity the constructs discriminate very well. So, as one could think, the total effect we measured is not an artifact.

The direct and indirect effects together make up the total influencing effect exerted by expectations about the alternative on satisfaction with the chosen travel destination. In our case the direct effect amounts to -0.14, while the indirect effect is -0.03 (=0.16 $ (-0.35) $ 0.48 + 0.16 $ (-0.20) $ 0.47 $ 0.48). The total effect is thus -0.17 (=-0.14 + (-0.03)). Hypothesis 1, which postulates a negative correlation between the expectations placed in the alternative and satisfaction with the actual travel destination, is evidently confirmed.

The most obvious method of analysing hypothesis 2, which asserts that the intensity with which expectations about the alternative influence satisfaction with the chosen travel destination is dependent on the extent to which the performance is considered to have been accomplished, is to make use of the data collected for the survey. The respondents were subdivided into two groups based on the degree to which they claimed their expectations about the chosen travel destination had been fulfilled. The first group comprised all those who stated a level between 1 and 4 for this indicator. This group of 134 persons is characterised by the fact that their expectations regarding the holiday resort were on the whole not satisfied. The second group is made up of all the respondents who specified a level between 5 and 8 for the same indicator. The expectations of these 274 interviewees concerning their travel destination were basically fulfilled.

The extent to which the asymmetry that is presumed in the formation of satisfaction ratings when alternatives are considered actually exists can be established by means of a group-specific analysis. Once again, the LISRELapproach was used for both studies to estimate the relationships between the relevant variables. Although the number of cases per study was considerably lower (134 respondents in group 1 and 274 in group 2), the values obtained for the various fitness indices are satisfactory. The stipulated threshold values were exceeded in both cases for all the examined quality yardsticks. We can therefore assume that the two models fit the real situation accurately.

The following results were obtained: in the case of group 1, in other words those holiday-makers whose expectations were not confirmed, the total effect of expectations about the alternative travel destination on satisfaction with the chosen resort was -0.26. There is a conspicuously sharp increase in the direct effect from -0.14 (in the overall model) to -0.21. The indirect effect rose only slightly from -0.03 (in the overall model) to -0.05. The total effect for group 2, comprising those respondents whose expectations were confirmed, was a mere -0.10. The significant drop in the direct effect from -0.14 (in the overall model) to -0.08 is noticeable here. The indirect effect was only slightly lower, namely -0.02 instead of -0.03 (in the overall model). Hypothesis 2, which claims that if expectations about the consumed product are (not) confirmed, the expectations regarding the alternative will exert a moderately strong (very strong) influence on satisfaction, is thus correct. The asymmetry that is presumed in the formation of customer satisfaction ratings when alternatives are considered is confirmed.

FIGURE 1

RESULTS OF THE PARAMETERS ESTIMATE

IMPLICATIONS

The central aim of this work was to derive an approach for analysing the formation of satisfaction ratings when alternatives are considered. An empirical study conducted in the hotel business served to verify the hypotheses formulated for this phenomenon on the basis of theoretical observations. The outcome was as follows: the higher the expectations about the unpurchased alternative, the lower the level of satisfaction with the product that is actually consumed. If, moreover, the expectations placed in the chosen product are (not) confirmed, expectations about the alternative will exert a moderately strong (very strong) influence on satisfaction.

A more detailed examination of the formation of satisfaction ratings would be of little significance for marketing managers were it not for the fact that the majority of them view the information gathered from their customers as an important starting point for specifying marketing activities. Account is taken of satisfaction ratings, for instance, when agreements on targets are fixed within the framework of employment contracts, when the variable components of payment systems are defined and when whole divisions of a company are controlled according to market-oriented principles. A major role is also played by customer satisfaction ratings in the concrete design of marketing mixes (Hauser, Simester, Wernerfeldt 1994).

It is on the basis of such data, for example, that companies decide whether or not to extend their set of services, develop new products or reposition existing products. The far-reaching implications of customer feedback of the kind described here explain why a careful analysis of the processes involved in the formation of satisfaction ratings is indispensable. The task of market researchers is to identify the influence of respondents’ expectations about an alternative on their satisfaction with the actually purchased commodity. This is essential if recommendations for action that are meaningful for marketing practice are to be derived from the information gathered about satisfaction. The following measures are also worthy of consideration.

To begin with, a supplier can lessen the regret effect by mentioning the shortcomings and problems associated with alternative commodities during the sales talk. In addition, after-sales activitiesBsuch as mailing further information and congratulating the buyer on his choice of bradBalso help to prevent feelings of regret from developing in the first place. Money-back guarantees and goodwill payments in the event of complaints are other possible ways of alleviating regrets that another alternative was not chosen instead.

As far as research strategy is concerned, future attention should focus more strongly on unpurchased products. Above all, it is important to determine how many different alternatives a consumer normally considers and the means by which he arrives at his preferred option. A deeper understanding of the process of determination and the make-up of the evoked set could be of assistance for carrying out a comprehensive analysis of the formation of satisfaction ratings when alternatives are considered. It must also be remembered that as the consumption or use phase progresses, the evaluation spectrum is enriched with experience components. As a result, the purchased alternative tends to be cognitively revalued, so that the regret effect is lessened. Moreover, consumers do not constitute a homogeneous unit, but rather exhibit different behavioural patterns, for instance, when it comes to forming judgements. The set of customers is thus a complex structure made up of different individual types that behave in different ways when evolving satisfaction ratings. If it is possible to identify the characteristic attributes of these types, a segment-specific explanation of their evaluation behaviour can be attempted.

LITERATURE

Anderson, E. W. (1994): Cross-Category Variation in Customer Satisfaction and Retention, Marketing Letters, 5, 19-30.

Anderson, E. W. (1996): Customer Satisfaction and Price Tolerance, Marketing Letters, 7, 19-30.

Anderson, E. W., Sullivan, M. W. (1993): The Antecedents and Consequences of Customer Satisfaction for Firms, Marketing Science, 12, 125-143.

Anderson, E. W., Fornell, C. (1994): A Customer Satisfaction Research Prospectus, in: Rust, R. T., Oliver, R. L. (eds.), Service Quality, Thousand Oaks, 241-268.

Anderson, E., Fornell, C., Lehmann, D. (1994): Customer Satisfaction, Market Share, and Profitability: Findings from Sweden, Journal of Marketing, 58, 53-66.

Bagozzi, R. P. (1995): Structural Equation Models in Marketing Research, in: Bagozzi, R. (ed.), Advanced Methods of Marketing Research, Cambridge, 317-385.

Bagozzi, R. P., Baumgartner, H. (1995): The Evaluation of Structural Equation Models and Hypothesis Testing, in: Bagozzi, R., (ed.), Advanced Methods of Marketing Research, Cambridge, p. 386-422.

Bagozzi, R., Yi, Y., Phillips, A. (1991): Assessing Construct Validity in Organizational Research, Administrative Science Quarterly, 36, 421-458.

Bell, D. E. (1982): Regret in Decision Making under Uncertainty, Operations Research, 30, 961-981.

Bell, D. E. (1983): Risk Premiums for Decision Regret, Management Science, 29, 1156-1166.

Bentler, P. M. (1993): EQS, Los Angeles.

Bentler, P. M., Bonett, D. G., Significance Tests and Goodness of Fit in the Analysis of Covariance Structures, Psychological Bulletin, 88, 588-606.

Bollen, K. (1989): Structural Equations with Latent Variables, New York.

Browne, M. W., Cudeck, R. (1993): Alternatives Ways of Assessing Model Fit, in: Bollen, K., Long, J. S. (eds.), Testing Structural Equation Models, Newbury Park, 120-136.

Festinger, L. (1957): A Theory of Cognitive Dissonance, Stanford.

Fornell, C. (1992): A National Customer Satisfaction Barometer: The Swedish Experience, Journal of Marketing, 56, 6-21.

Fornell, C, Johnson, M. D., Anderson, E. W., Cha, J., Everitt Bryant, B. (1996): The American Customer Satisfaction Index: Nature, Purpose, and Findings, Journal of Marketing, 60, 7-18.

Fornell, C., Tellis, F. G., Zinkhan, B. J. (1982): Validity Assessment: A Structural Equations Approach using Partial Least Squares, AMA Chicago, 405-409.

Gilovich, T., Medvec, V. H. (1995): The Experience of Regret: What, When, and Why, Psychological Review, 102, 379-395.

Hair, J. F., Anderson, R. E., Tatham, R. L., Black, W. C. (1995): Multivariate Data Analysis with Readings, 4. edition, Englewood Cliffs.

Harless, D. W. (1992): Actions versus Prospects: The Effect of Problem Representation on Regret, American Economic Review, 82, 634-649.

Hauser, J. R., Simester, D. I., Wernerfeldt, B. (1994): Customer Satisfaction Incentives, Marketing Science, 13, 327-350.

Herrmann, A., Huber, F., Gustafsson, A. (1997), From Value-oriented Quality Improvement to Customer SatisfactionBA Case Study for Passenger Cars, in: Johnson, M.D., Herrmann, A., Huber, F., Gustafsson, A. (eds.), Customer Retention in the Automotive Industry: Quality, Satisfaction and Loyalty, Wiesbaden, Germany, 93-115.

Imman, J. J., Dayer, J. S., Jia, J. (1997): A Generalized Utility Model of Disappointment and Regret Effects on Post-Choice Valuation, Marketing Science, 16, 97-111.

J÷reskog, K. G., S÷rbom, D. (1993): LISREL 8: Structural Equation Modeling with the SIMPLIS Command Language, Chicago.

Johnson, M. D. (1998): Customer Orientation and Market Action, Upper Saddle River.

Johnson, M. D., Gustafsson, A. (1997): Bridging the Quality-Satisfaction Gap II: Measuring and Prioritizing Customer Needs, in: Gustafsson, A., Bergman, B., Ekdahl, F. (eds.), Proceedings of the Third Annual International QFD Symposium, Volume 2, Link÷ping, Sweden, 21-34.

Kaufman, P., Dant, R. (1992): The Dimensions of Commercial Exchange, Marketing Letters, 3, 171-185.

Krishnamurthi, L., Raj, S. (1991): An Empirical Analysis of the Relationship between Brand Loyalty and Consumer Price Elasticity, Marketing Science, 10, 172-183.

Larrick, R. L., Boles, T. L. (1995): Avoiding Regret in Decisions with Feedback: A Negotiation Example, Organizational Behavior and Human Decision Processes, 63, 87-97.

Loomes, G. (1988): When Actions Speek louder than Prospects, American Economic Review, 78, 463-470.

Loomes, G., Sugden, R. (1982): Regret Theory: An Alternative Theory of Rational Choice under Uncertainty, The Economic Journal, 92, 805-824.

Loomes, G., Sugden, R. (1987a): Some Implications of a more General Form of Regret-Theory, Journal of Economic Theory, 41, 270-287.

Loomes, G., Sugden, R. (1987b): Testing for Regret and Disappointment in Choice under Uncertainty, The Economic Journal, 97, 118-129.

Oliver, R. L., DeSarbo, W. (1988): Response Determinants in Satisfaction Judgement, Journal of Marketing, 14, 393-402.

Peter, J. P., Olson, J. C. (1990): Consumer Behavior, 2. edition, Homewood.

Ritov, I., Baron, J. (1995): Satisfaction and Retention, Organisational Behavior and Human Decision Processes, 64, 119-127.

Simonson, I. (1992): The Influence of Anticipating Regret and Responsibility on Purchase Decisions, Journal of Consumer Research, 19, 105-118.

Steiger, J. H. (1990): Structural Model Evaluation and Modification: An Interval Estimation Approach, Multivariate Behavioral Research, 25, 173-180.

Taylor, K. (1997): A Regret Theory Approach to Assessing Consumer Satisfaction, Marketing Letters, 7, 229-238.

Urban, G. L., Hauser, J. R. (1993): Design and Marketing of New Products, 2. edition, Englewood Cliffs.

Zeithaml, V. A. (1988): Consumer Perceptions of Price, Quality, and Value: A Means-End Model an Synthesis of Evidence, Journal of Marketing, 52, 2-22.

Zeithaml, V. A., Parasuraman, A., Berry, L. L. (1990): Delivering Service Quality, New York.

----------------------------------------