Special Session Summary Tracing the Daisy Chain: Post-Purchase Affect and Its Influences on Subsequent Consumption


Anirban Mukhopadhyay (2005) ,"Special Session Summary Tracing the Daisy Chain: Post-Purchase Affect and Its Influences on Subsequent Consumption", in NA - Advances in Consumer Research Volume 32, eds. Geeta Menon and Akshay R. Rao, Duluth, MN : Association for Consumer Research, Pages: 495-498.

Advances in Consumer Research Volume 32, 2005     Pages 495-498



Anirban Mukhopadhyay, Hong Kong University of Science and Technology


Session Objective

Purchase decisions made at one point in time can often have an impact on subsequent behavior. The purpose of this special session was to present ongoing research that examines one aspect of the dynamics of such intertemporal behavior patterns, and to provide an integrative platform for a discussion of fruitful research directions.


Although much of consumer research addresses single purchase decisions, from need recognition to post-purchase evaluation (e.g. Howard and Sheth 1969), there are a few streams of research that study behavior across purchase occasions. One area, researched in depth, has looked at repeated behavior in the context of brand loyalty and variety seeking (e.g. Tucker 1964). A second well-researched perspective addresses a priori preferences on configurations of purchase decisions (e.g. Loewenstein and Prelec 1993). There is, however, a third stream that concerns itself with how consumers feel contingent on having made a purchase, and how this purchase decision affects behavior t a subsequent decision point. Research on the processes underlying such interdependence of purchase decisions across periods is surprisingly scant.

Recent work in marketing (Dhar, Huber, and Khan 2003) and psychology (Muraven and Baumeister 2000) has shown that behavior at one point in time can have varied and interesting effects on subsequent behavior. The three papers in this session add to this stream by looking at affective consequences of purchase decisions, and how they influence responses to subsequent consumption opportunities. As discussed below, the affect considered here is distinct from decision satisfaction and cognitive dissonance reduction, and the subsequent behavior is more generalized than repeat purchases of the same brand.


The three papers in this set demonstrated that purchases cause affect, which then influence responses to subsequent consumption opportunities. In the first paper, Patrick and Park looked at "consumption chains" that arise from "trigger" purchases of products that don’t fit with one’s existing possessions. Such purchases may cause unhappiness with one’s existing possessions, and hence set off a series of additional compensatory purchases. The circumstances under which the trigger purchase was acquired, for instance a self-bought item versus a gift, influence the level of dissatisfaction and consequently the buying behavior. In the second paper, Cheema, Chakravarti, and Sinha studied differences in the regret that auction participants feel, contingent on the result of the auction. They demonstrated cognitive and motivational antecedents of regret, which when combined with the auction’s outcome influence behavior in a subsequent auction. Here the purchase decision has been made, but the outcome, and hence the outcome-dependent affect, may not be in ones control. Nevertheless, the outcome and the affect influence subsequent behavior. Finally, Mukhopadhyay and Johar examined responses to unintended purchase opportunities, such as a sale encountered unexpectedly. They found that decisions to either buy or not buy in such situations lead to complex affective responses, with buying causing happiness, guilt, and remorse more than not buying, which causes pride. These mixed emotions then color responses to different types of emotional appeals in advertising for other unrelated offers. The underlying theme is that across all three papers, purchase decisions of different sorts lead to affect, which then influences subsequent consumption directed behavior.


The topic of the interdependence of purchase decisions is at the cutting edge of consumer research. Each of the three papers in this session presented novel and interesting results. Taken together, they constituted a significant step in the study of intertemporal behavior, across domains such as shopping, auctions, and ad response. At the conclusion, the audience and authors engaged in a discussion of the ideas presented as well as future research directions in the session’s context.




Vanessa M. Patrick, University of Georgia

C. Whan Park, University of Southern California

This research is concerned with the interdependence of consumption experiences. Most traditional models of decision-making in consumer behavior begin with consumer problem/need recognition and end with post-consumption evaluation and behavior (Engel et al. 1968; Howard and Sheth 1969; Nicosia 1966). In these models, and in the majority of consumer research, a purchase is assumed to be a single and isolated event. In this research we reexamine this assumption by investigating the interdependence of consumption experiences. The basic premise of this research is that consumption rarely occurs in isolation and the purchase of one item often leads to a series of successive purchases, forming what we refer to as a consumption chain.

While previous research has examined how the purchase of a product can lead on to a stream of repeat purchases of the same product (McConnell 1968; Tucker 1964) and even how the "momentum" of shopping can lead to making other unrelated purchases (Dhar et al. 2003), this research examines the stream of related purchases generated from a single "trigger" purchase. Moreover, this research posits that consumption chains are activated in order to enhance the overall satisfaction derived from the "trigger" purchase. Specifically, we propose a broadened conceptualization of satisfaction to suggest that satisfaction with a product is dependent not only on product performance but also on the "fit" of the product with one’s existing possessions (for a similar notion of satisfaction as active, dynamic and context-dependent, see Fournier and Mick, 1999).

Three studies have been conducted to examine the notion of consumption chains and to discern the potential causes and consequences of the interdependence of consumption experiences. The first qualitative study using real-world examples of consumption chains (N=314) reveals the universality of this phenomenon and provides real-world instances of consumption chains. In this study, participants were asked to illustrate an example of "one purchase leading to another." The objective of this study was to understand whether and how consumers relate to the notion of consumption chains, and to determine the characteristics of "trigger" purchases that initiate consumption chains. Moreover, this study provides evidence of the complexity of consumption chains by examining the number of "chains" of consumption that arose from the "trigger" purchase, the number of purchases generated by a single trigger purchase and the diversity of the consumption chain.

A second study (N=171) examines the conditions under which consumption chains are likely to be generated. In this study, the nature of the consumption (painting (hedonic) vs. toaster (functional)), the degree of controllability (self-bought vs. gifted) and the "fit" of the product (high vs. low) with existing possessions (in this case the decor of one’s home) are manipulated. We find that satisfaction with products that "fit" is higher than satisfaction with products that do not fit (main effect of fit). Moreover, for items that have low "fit", satisfaction with hedonic purchases is higher than satisfaction with utilitarian purchases and satisfaction with purchases that do not fit is lower than satisfaction with gifts that do not fit (main effect of controllability) and obviously satisfaction with products that fit better is higher than satisfaction with those that do not (main effect of "fit"). Examining the effect on the likelihood of purchasing additional items, a three item "trigger likelihood" index was created. Results reveal that hedonic purchases are more likely to trigger consumption chains than are functional purchases (main effect of nature of purchase), purchases are more likely to trigger consumption chains than are gifts (main effect of controllability) and a significant nature of product X controllability interaction in which hedonic purchases are most likely while utilitarian purchases are least likely to trigger consumption chains.

In the third study (N=83), we look at hedonic purchases only and examine one potential explanation as to why one purchase often leads to other related purchases. Our primary proposition for this study is that hedonic purchases that have low "fit", instead of increasing overall satisfaction, in fact decrease overall satisfaction with one’s possessions by making the deficiencies or "what one does not have" salient. These "trigger purchases" initiate a string of successive purchases that aim to enhance the overall satisfaction that may be derived from the trigger purchase itself. Thus, in this study we look at a hedonic trigger purchase (a shirt) and manipulate the degree of controllability (self-bought versus gifted) and the "fit" of the product (high versus low) with existing possessions (in this case one’s wardrobe). This study reveals that whether purchased or gifted, satisfaction with the product is higher when the fit with existing possessions is good versus bad (main effect of "fit"). In addition, purchases are more likely than gifts to trigger consumption chains, and, when a purchase is hedonic, the satisfaction with the purchase itself is high but the satisfaction with other possessions decreases thereby triggering a chain of consumption to enhance overall satisfaction. In sum, a new hedonic purchase that has a low fit, decreases the satisfaction with one’s existing possessions, and simultaneously increases the likelihood of buying additional items to complement the new purchase.

In conclusion, this research investigates the "consumption chains" phenomenon, explores the characteristics of the "trigger" purchase that initiates the consumption chain, examines potential explanations underlying this phenomenon and discusses the implications for marketing and consumer behavior.



Amar Cheema, Washington University in St. Louis

Dipankar Chakravarti, University of Colorado B Boulder

Atanu Sinha, University of Colorado B Boulder

This paper examines how motivational (i.e., a winning-oriented or a prudent, value-oriented focus) and cognitive antecedents (precision of value information and its salience during the auction) influence first consumers’ experienced regret contingent on auction outcome (winning versus losing), and then behavior in a subsequent auction. We compare the pattern of direct and interactive effects in two experiments (involving a descending and an ascending auction respectively) that embed manipulations of the available deliberation time (short/long) at each auction step. The contrasting patterns of effects are interpretable in terms of how mechanistic differences in the auction procedures influence how the manipulated variables drive the evolution of bidding strategies and value.

In contrast with prior work in economic auction theory (Klemperer 2000) the conceptual premise of this work is that consumers do not enter auction situations with known and immutable values (Bazerman 2001; Chakravarti et al. 2002). Rather, they sometimes have no prior value for the object, or know it only imprecisely. Moreover, what is known about value may be more or less salient during the auction, depending on the cognitions activated by prior circumstances. Furthermore, consumers may approach the auction in different mindsets related to a chronic or induced regulatory focus (Higgins 1997) on winning the auction versus acquiring the object at a prudent, value-focused price.

These motivational and cognitive variables, combined with the amount of deliberation opportunity available at each auction step should influence how the objects’ value evolves for the consumer. For example, the variables may influence consumers’ attention to during-auction events (e.g. competitive bidding and entry/exit behavior), what they infer about their own and competitors’ values, how they bid, and their probabilities of winning (losing) the auction. Consequently, they may have direct and interactive effects on the regret that consumers experience as a function of winning or losing.

Regret, a sense of disappointment or distress experienced as a function of a goal-incongruent outcome, generates a sense of dissonance that needs mitigation (e.g., Looms and Sugden 1987; Cooper and Fazio 1984). Therefore, contingent on the auction outcome, consumers may attempt to mentally manipulate their own values, as well as their perceptions of others’ values relative to their own (or their opponent’) winning bid. Such labile values (Fischoff 2000) provide a means to managing regret, whether it is a winner’s fear of having paid too much or a loser’s sense of having compromised their ownership option. Therefore, the experienced outcome-contingent regret may mediate the post-auction values that consumers report and also influence their behavior in a second, subsequent auction.

We offer formal propositions regarding how such processes may operate, and then examine these propositions in two laboratory auction experiments. The auction environment is computer-simulated and subjects bid for a set of objects (calibrated for their aesthetic and functional properties) against a set of pre-programmed robot bidders. The subjects’ goals are manipulated using conceptual primes that elicit primarily a winning versus a prudent, value-oriented focus during the auction. The precision with which the object’s price is known is manipulated by providing an estimated price range that is either wide or narrow (centered on the same mean). Value salience is manipulated by the presence or absence of a pre-measurement procedure in which subjects are asked to estimate their own and others’ value for the product.

Experiment 1 examines a descending (Dutch) auction environment in which prices drop from a high value in fixed 50-cent steps with deliberation time at each step manipulated to be either short (2 seconds) or long (5 seconds). Subjects participate in two consecutive auctions. Following the first auction, they provide measures of experienced regret, followed by measures of their post-auction value for the auction object, and task reactions. We examine the direct and interactive effects of the manipulated variables on experienced regret and the extent to which these patterns are contingent on the auction outcome (win / loss). Experiment 2 repeats the study in an ascending auction (Japanese variant of an English auction, i.e., with open bids and irrevocable exit). The focal variables are manipulated identically, with deliberation time pre-calibrated to be either short (5 seconds) or long (10 seconds) to allow for procedural differences.

The studies reveal starkly different patterns of direct and interactive effects of the manipulated variables on regret. On average, regret levels are higher in the descending (versus the ascending) auction. The manipulated variables show few main effects in either auction, but influence regret interactively. The interaction patterns are qualitatively different. For example, in the ascending auction, the goal manipulation has few direct or interactive effects with other manipulated variables. However, in the descending auction, it shows complex higher order interactions, suggesting that its impact is significantly moderated by variables such as the deliberation time, value salience and value precision.

Moreover, while the win/loss outcome affects regret directly in the ascending auction, its impact is moderated by a selected set of the manipulated variables (e.g., bidder goals and deliberation time) in the descending auction. Thus, the regret experienced in the ascending auction (which reveals more information and allows subjects more control over the outcomes) involves a simple outcome contingency (winning/losing). However, the sudden-death and low information characteristic of the descending auction creates more complex contingencies for the regret experience, involving not just the win/loss outcome, but also selected other manipulated variables. Among subjects for whom value was pre-measured, the regret experience was more predictable for losing bidders than winning bidders as a function of the gap between the pre-measured values (self and others) and the winning bid.

We model experienced regret as a function of the discrepancy between the winning bid (their own or an opponent’s) and prior value (their own or perceived others’). Patterns of results demonstrate that the importance of own and others’ perceived values differ as a function of the auction outcome. Moreover, the outcome from the first auction (and the following regret) affects behavior in a second, subsequent auction. Interestingly, thi pattern holds for the descending auction but not for the ascending auction. These effects are interpreted in terms of differences in the information availability and inferential dynamics permitted by the two auction mechanisms.



Anirban Mukhopadhyay, Hong Kong University of Science and Technology

Gita V. Johar, Columbia University

Consumers spend substantial proportions of their expenditures on products they had not intended to buy. Correspondingly, marketers spend billions of dollars trying to influence purchase incidence. How do consumers feel after they give in to such temptations, or hold back from them? And how does this affect influence responses to subsequent consumption opportunities? This paper investigates consumers’ mixed emotions deriving from their decisions to either buy or not buy at an unintended purchase opportunity, such as a sale encountered by chance, and how these reactions can affect responses to subsequent tempting offers.

Stern (1962) defines "suggestion impulse buying" as a purchase prompted by a need activated on exposure. That is, seeing a product on offer may activate a certain goal, which may be achieved by acquiring and using the product. However, buying the product also involves the temptation to spend money where one had not intended to, and hence exposure to an attractive offer could also simultaneously cue a goal of not spending ones money unnecessarily (Fishbach, Friedman, and Kruglanski 2003). These two goals could be in conflict, as either decisionCto buy or not buyCwould result in protecting or furthering one goal at the cost of the other. Buying furthers the goal served by acquisition while violating the goal of not spending, and not buying protects the goal of not spending while sacrificing the goal served by acquisition. Now, progress towards a goal can generate positive affect, while lack of progress towards or giving up on a goal can generate negative affect (Carver and Scheier 1990). Hence both buying and not buying should generate positive affect from one source (acquisition or not spending), but also negative affect from the other source. Further, specific to unintended purchase, hypothesized emotions would be: pride at protecting the goal of not spending or guilt for having spent money unplanned on the one hand, and happiness at having acquired something inherently pleasurable or regret at passing up a tempting object (Carmon and Ariely 2000) on the other (Rook and Gardner 1988).

The first two studies in this sequence tested the above predictions. These studies supported the hypotheses that consumers’ responses to unintended purchase opportunities can have complex affective consequences, combining feelings of pride, happiness, guilt, and remorse. Specifically, faced with an unintended purchase scenario (a sale on software products encountered unexpectedly), participants who chose to buy the products on offer reported significantly stronger feelings of happiness, marginally more guilt and remorse, and significantly less pride, than those who decided not to buy. This finding was robust across the studies, and did not depend on the nature of the software on sale, or whether attention was drawn to the product involved (i.e. the goal served by acquisition) or the act of spending.

The aim of the next study was to investigate how these mixed emotions play out across purchase opportunities. In Experiment 3, participants were shown an advertisement for an attractive sale (for software products) being held by their university bookstore. The ad appeal was manipulated to be a pride appeal, a happiness appeal, or a mixed pride-and-happiness appeal. The timing of the ad exposure was also manipulated, such that participants either saw the ad after a no-buy decision at a scenario involving a sale on books (thereby presumably generating mixed emotions of pride and unhappiness), or were shown the ad without any shopping history, when the would have been at a baseline affective state. Results showed that the nature of the advertising appeal had sharply differing effects depending on when the ad was viewed. At a baseline affective state, with no recent shopping history, appeals that emphasized only pride or only happiness were seen to be more effective than mixed appeals. However, mixed appeals were more effective if viewed after a no buy decision, which presumably had caused both pride and unhappiness. This is in contrast to past research as well as traditional advertising wisdom, that mixed emotional appeals are not favored by Western cultures (Williams and Aaker 2002; Ries and Trout 1981). Indeed, when viewed in a state of mixed emotions generated by a recent decision to not buy, pure happiness appeals tended to perform worse than the pride or mixed appeals. This pattern of results held down the hierarchy of effects from ad attractiveness through attitude to the ad, attitude to the product, and purchase intention.

In Experiment 4, respondents were allowed to choose to buy or not at the first scenario, and then shown either the pride or the happiness appeal. The extent of discount at the first sale was also varied, such that it was either the same as in Experiment 3 ($55 marked down to $29.95), or heavier ($19.95). When the first sale was at the lower regular discount level, respondents who choose to not buy preferred pride and mixed appeals, replicating Experiment 3. Moreover, those who choose to buy preferred pure happiness appeals, thereby extending the above result. However this pattern did not hold for those who had seen the heavier discount. Participants in these conditions did not respond well to the pride appeals, and process measures indicated that this may be because these respondents report feeling angry, rather than proud, about their decision to not buy.

This research contributes in two ways to the literature on mixed emotions. First, it demonstrates how specific combinations of mixed emotions can result from situations that are not uncommon in the real world. And second, it demonstrates how these mixed emotions can themselves influence responses to different types of persuasion appeals. Mixed appeals are not always undesirable, and indeed may sometimes even be recommended. Essentially, this research shows that if as a marketer you know what your target has just done, you have a better idea as to how they’re feeling, and also guidelines on how to tailor your communication appropriately.


Bazerman, Max H. (2001), "Consumer Research for Consumers," Journal of Consumer Research, 27 (March), 499-504.

Chakravarti, Dipankar, Eric Greenleaf, Atanu Sinha, Amar Cheema, James C. Cox, Daniel Friedman, Teck H. Ho, R. Mark Isaac, Andrew A. Mitchell, Amnon Rapoport, Michael H. Rothkopf, Joydeep Srivastava and Rami Zwick (2002). "Auctions: Research Opportunities in Marketing," Marketing Letters, 13, 3 (August), 281-296.

Carmon, Ziv and Dan Ariely (2000), "Focusing on the Forgone: How Value can Appear so Different to Buyers and Sellers," Journal of Consumer Research, 27, 360-370.

Carver, Charles S. and Michael F. Scheier (1990), "Origins and Functions of Positive and Negative Affect: A Control-Process View," Psychological Review, 97, 19-35.

Cooper, Joel and Russell H. Fazio (1984), "A New Look at Dissonance Theory," in L. Berkowitz (Ed.) Advances in Experimental Social Psychology, 17, Orlando, FL: Academic Press. 229-266.

Dhar, Ravi, Joel Huber, and Uzma Khan (2003), "The Shopping Momentum Effect," Working paper, Yale University.

Engel, James F., David T. Kollatt, and Roger D. Blackwell (1968), Consumer Behavior, New York: Holt, Rinehart and Winston.

Fishbach, Ayelet, Ronald S. Friedman, and Arie W. Kruglanski (2003), "Leading Us Not Unto Temptation: Momentary Allurements Elicit Overriding Goal Activation," Journal of Personality and Social Psychology, 84 (February), 296-309.

Fischhoff, Baruch (2000), "Value Elicitation: Is there anything in there?" in T. Connolly, H.R. Arkes and K.R. Hammond (eds.), Judgment and decision making: An interdisciplinary reader, New York, NY: Cambridge University Press, pp. 517B543.

Fournier, Susan and David G. Mick (1999), "Rediscovering Satisfaction," Journal of Marketing, 63 (October), 5-23.

Higgins, E. Tory (1997), "Beyond Pleasure and Pain," American Psychologist, 52, 1280-1300.

Howard, John A. and Jagdish N. Sheth (1969), The Theory of Buyer Behavior, New York: Wiley.

Klemperer, Paul D. (2000), (ed.), The Economic Theory of Auctions, Cheltenham, UK: Edward Elgar.

Loewenstein, George and Drazen Prelec (1993), "Preferences for Sequences of Outcomes," Psychological Review, 100 (1), 91-108.

Loomes, Graham and Robert Sugden (1987), "Some Implications of a More General Form of Regret Theory," Journal of Economic Theory, 41, 270-287.

McConnell, J. Douglas (1968), "The Development of Brand Loyalty," Journal of Marketing Research, 5 (February), 13-19.

Muraven, Mark and Roy F. Baumeister (2000), "Self-Regulation of Limited Resources: Does Self-Control Resemble a Muscle?" Psychological Bulletin, 126 (March), 247-259.

Nicosia, Franco M. (1966), Consumer Decision Processes: Marketing and Advertising Implications, New Jersey: Prentice Hall.

Ries, Al and Jack Trout (1981), Positioning: The Battle For Your Mind, New York: Warner

Rook, Dennis W. and Meryl P. Gardner (1988), "Effects of Impulse Purchases on Consumers’ Affective States," Advances in Consumer Research, 15, 127-130.

Stern, Hawkins (1962), "The Significance of Impulse Buying Today," Journal of Marketing, 26 (April), 59-62.

Tucker, W. T. (1964), "The Development of Brand Loyalty," Journal of Marketing Research (August), 32-35.

Williams, Patti, and Jennifer L. Aaker (2002), "Can Mixed Emotions Peacefully Coexist?" Journal of Consumer Research, 28 (March), 636-649.



Anirban Mukhopadhyay, Hong Kong University of Science and Technology


NA - Advances in Consumer Research Volume 32 | 2005

Share Proceeding

Featured papers

See More


Perceptions of Epistemic vs. Aleatory Uncertainty Affect Stock Investment

Daniel Walters, INSEAD, France
Gulden Ulkumen, University of Southern California, USA
Carsten Erner, FS Card
David Tannebaum, University of Utah, USA
Craig Fox, University of California Los Angeles, USA

Read More


Using a Meta-Analysis to Unravel Relative Importance of Postulated Explanations for the Endowment Effect

Peter Nguyen, Ivey Business School
Xin (Shane) Wang, Western University, Canada
David J. Curry, University of Cincinnati, USA

Read More


Penny for Your Preferences: Leveraging Self-Expression to Increase Prosocial Giving

Jacqueline R. Rifkin, Duke University, USA
Katherine Crain, Duke University, USA
Jonah Berger, University of Pennsylvania, USA

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.