Influence of Soldout Products on Consumers’ Choice 1 Xin Ge a , Paul R. Messinger b , and Jin Li c a Assistant Professor, University of Northern British Columbia, b Associate Professor, University of Alberta, c Assistant Professor, North Dakota State University, ABSTRACT Previous research on consumer choice has largely focused on how consumers make purchase decisions given information about available products. However, consumers often face situations in which information about soldout products is still present in the decision context. We argue that soldout products convey special information about the decision environment, and that consumers assimilate this information when making their decisions. In particular, three experiments demonstrate what can be called an immediacy effect, whereby soldout products prompt consumers to expedite their purchases. Soldout products also influence consumers’ choice among available options. Consumers may even actively search for information about soldout products. 1 This research was supported by the Social Sciences and Humanities Research Council of Canada, and by faculty research grants from the University of Northern British Columbia and North Dakota State University. For helpful suggestions, the authors would like to thank participants of the session ―Important Consumer Behavior Issues in Retailing‖ in the 2007 Marketing Science Conference in Singapore.
33
Embed
Influence of Soldout Products on Consumers’ Choice · INTRODUCTION Much of the ... because of newspaper lead times or lags in changing ... decision environment can influence consumers’
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Influence of Soldout Products on Consumers’ Choice1
Xin Gea, Paul R. Messinger
b, and Jin Li
c
aAssistant Professor, University of Northern British Columbia,
bAssociate Professor, University of Alberta,
cAssistant Professor, North Dakota State University,
ABSTRACT
Previous research on consumer choice has largely focused on how consumers
make purchase decisions given information about available products. However,
consumers often face situations in which information about soldout products is still
present in the decision context. We argue that soldout products convey special
information about the decision environment, and that consumers assimilate this
information when making their decisions. In particular, three experiments demonstrate
what can be called an immediacy effect, whereby soldout products prompt consumers to
expedite their purchases. Soldout products also influence consumers’ choice among
available options. Consumers may even actively search for information about soldout
products.
1This research was supported by the Social Sciences and Humanities Research Council of Canada,
and by faculty research grants from the University of Northern British Columbia and North Dakota State
University. For helpful suggestions, the authors would like to thank participants of the session ―Important
Consumer Behavior Issues in Retailing‖ in the 2007 Marketing Science Conference in Singapore.
2
Influence of Soldout Products on Consumers’ Choice
Seinfeld, Episode 119,
“The Sponge”:
ELAINE: Well, Kramer was right. My friend Kim told me the sponge is off the market.
JERRY: So what are you gonna do?
ELAINE: I’ll tell you what I’m gonna do – I’m gonna do a hard-target search. Of every drug store, general
store, health store and grocery store in a 25-block radius.
ELAINE (after much futile search, with little hope): Do you have any Today sponges? I know they’re off
the market, but…
PHARMACIST: Actually, we have a case left.
ELAINE (excited): A case! A case of sponges? I mean, uh…a case. Huh. Uh…how many come in a case?
PHARMACIST: Sixty.
ELAINE: Sixty?! Uh…well, I’ll take three.
PHARMACIST: Three.
ELAINE: Make it ten.
(... Then ELAINE goes for twenty, twenty-five, and finally…)
ELAINE: Yeah. Just give me the whole case and I’ll be on my way.
INTRODUCTION
Much of the existing literature examines how consumers process information
about available offers on the market and make a choice from among these options.
However, a phenomenon that has largely escaped research attention is that consumers
often face situations in which information about soldout products is still present in the
decision context. For example, advertisements of recently-sold houses or automobiles
often remain in newspapers for a period of time before being withdrawn, marked by a
soldout stamp in such a way that consumers can still examine the attribute information of
3
the alternatives if they desire to do so. Consumers browsing online clothing stores often
find out that clothes of certain styles, colors, or sizes of interest are sold out. Retailers in
consumer electronics sometimes even intentionally keep a model on display even after all
units of that model are sold out. Information about soldout products may, thus, be
available unavoidably (for instance, because of newspaper lead times or lags in changing
shelf facings or retail displays) or it may be intentionally kept available (as a form of
promotion). But in either case, such information often constitutes part of the decision
context, and it is incumbent on retailing practitioners to understand the impact of such
information on consumer choice. Towards this end, we examine in this paper some
features about how consumers make purchase decisions in the presence of soldout
products. In particular, we ask whether consumers ignore information about soldout
products in making choices, and, if not, how soldout products influence consumers’
choices.
A relevant question in this context is whether the presence of soldout products
reduces the incidence of the ―no-choice‖ option, i.e., the option to defer a decision
(Tversky and Shafir 1992; Dhar 1997). Rational choice theory remains silent with
respect to the influence of unavailable products on the incidence of the no-choice option.
Standard theory assumes that unavailable products do not affect the ratios of choice
probabilities associated with the available items (Luce 1959). By contrast, we suggest
that soldout products convey special information about the decision environment, and
that consumers actively process this information in a way that systematically influences
their preference for the no-choice option (as well as their choice from among the
available options).
4
The current research is motivated by two propositions. First, we propose that
soldout products make salient to consumers a time dimension in the decision environment
and cause consumers to pay attention to the flow of supply and demand, turning the
otherwise static decision scenario into a dynamic one. When consumers do not have a
reason to make an immediate purchase decision, they often decide to defer a choice by
choosing to search for information about additional options, or simply by choosing to
wait (Tversky and Shafir 1992; Dhar 1997). Research has also found that substantial
delay often exists between the recognition of need and the time of purchase (Greenleaf
and Lehmann 1995). As the Seinfeld episode at the beginning of this paper illustrates,
unavailability of a product provokes consumer’s attention to changes in product supply
and demand in the market, which in turn creates a sense of immediacy in making the
purchase decision, and even prompts the consumer to take extra effort to remove the
constraint on availability.
Second, we propose that soldout products provide a background context serving
as a reference point against which consumers evaluate the available options. An
abundance of work has demonstrated that consumers often form judgments of an option
not by its absolute value, but rather by its relative position compared with other market
offerings (Huber et al. 1982; Simonson 1989; Tversky and Simonson 1993; Wernerfelt
1995). Just as individuals rely on past prices as the reference price to make judgments
and choices of the observed offerings (Kalyanaram and Winer 1995), they also tend to
consolidate the information about the soldout product into their information set and use
the soldout products as the basis for considering the available options. We propose that
soldout products, thus, help consumers form a better understanding of the product
5
distribution and facilitate evaluation of the available options. The remainder of this paper
is organized as follows. We begin by reviewing the literature most relevant to
understanding and predicting the influence of soldout products on consumers’ choice.
We, next, propose a number of hypotheses, followed by the results of three experiments.
The first two experiments focus on the influence of a soldout product on consumers’
attitudes toward decision deferral, as measured by their choice of the no-choice option
versus the available options. The third experiment examines the shift in consumers’
choice among available options as a response to the soldout product. This experiment
also investigates consumers’ search behavior with respect to the information about
soldout products. We conclude with a discussion of the theoretical and practical
implications of the results.
SOLDOUT PRODUCTS AND CONSUMERS’ CHOICE
Whereas limited work has been directed at understanding the influence of soldout
products on subsequent consumer decision processes, the notion of unavailable options in
the decision environment has attracted previous research interest. Pratkanis and Farquhar
(1992) coined the term ―phantom alternatives‖ to refer to any choice options that are
unavailable at the time of decision making. A distinction can be made between known
phantoms, i.e., phantoms that are recognized as unavailable from the outset, and
unknown phantoms, i.e., those whose unavailability is initially unrecognized (Pratkanis
and Farquhar 1992; Farquhar and Pratkanis 1993). Examples of the former include
preannounced new products (Eliashberg and Robertson 1988) and past prices
(Kalyanaram and Winer 1995). Examples of the latter include fully booked hotels or
6
restaurants that consumers include in their consideration sets before the unavailability of
these products becomes clear. In a retail context, unknown phantoms can occur as a result
of stores’ conscious manipulation of the selling environment as is done with ―bait and
switch‖ tactics, in which stores advertise a product at extremely low price, and then
reveal to interested customers that the advertised product (the bait) is not available but a
substitute (the switch) is.
Previous research has examined different responses elicited by unknown
phantoms and known phantoms. If a phantom is not clearly identified as unavailable at
the outset, consumers may feel entitled to availability of the phantom as they go through
the decision process. As the unavailability is revealed at a later stage of decision making,
it can result in reactance, i.e., a feeling of deprivation and dissatisfaction (Fitzsimons
2000). In addition, when consumers perceive that availability of the unknown phantom
can be restored, reactance produces a motivation directed toward removing constraints on
obtaining the product (Brehm 1966). However, when consumers perceive that
unavailability of the unknown phantom can not be reversed, they tend to derogate the
attractiveness of the option or reduce the importance weightings given to the option’s
best attributes (Hammock and Brehm 1966; Clee and Wichlund 1980; Potter and Beach
1994). Furthermore, in the context of compliance technique, such as low-balling
procedure through which the influencer first recruits participants to perform a task by
indicating only a low level of cost, but then reveals much higher costs actually required to
complete the task (thus, participants’ original commitment is no longer a valid option and
becomes an unknown phantom), it has been demonstrated that participants tend to
7
escalate their commitment to performing the onerous task by complying to costs higher
than their expectation (Cialdini et al. 1978).
On the other hand, a known phantom has been shown to affect decision makers’
preferential judgments. In the same manner as consumers’ preference for an option is
contingent on the structure, or composition, of the offered set, unavailable options in the
decision environment can influence consumers’ relative preference for the available
options. Doyle et al. (1999) and Fitzsimons (2000) established the attraction effect in the
context of unavailable options. These researchers showed that the presence of a third
unavailable option that is clearly inferior to one of the existing options (but not the other)
leads to an increased choice incidence of the option that dominates the unavailable
option. In addition, Pratkanis and Farquhar (1992) demonstrated that, in an offered set of
two options (referred to as the target and the referent) that involve a tradeoff between two
attributes such that the target is superior to the referent on attribute 1, then the presence of
a phantom that is preferred to the target on attribute 1 leads to an increased weight
assigned to attribute 1, which in turn has the potential to result in an increased likelihood
of choosing the target. However, at the same time a comparison with the phantom makes
the target less attractive, a result in line with the attraction effect. The probability of
choosing the target is a function of the magnitude of the two competing processes.
In the current research, the soldout products are conceptualized as recognized
phantoms that are known from the outset to be unavailable (because, for example, the
product may be marked by a soldout stamp). In principle, the known unavailability may
reduce consumers’ feelings of commitment to these products in the decision process.
8
Even so, we propose that soldout products can produce significant changes in consumers’
choice.
The provision of information about soldout product can be viewed as a form of
market signaling. Market signals arise from information asymmetry between two parties
to a transaction about the characteristics of the marketplace (Kirmani and Rao 2000;
Spence 1974, 2002). Market signals, according to signaling theory, are often used by
sellers to convey information to consumers that the latter do not have easy access to but
desire to possess. For example, retailers might provide a low price guarantee to signal to
consumers the proximity of retailer’s offer price to the lowest market price (Dutta and
Biswas 2005; Biswas et al. 2006).
Whereas a seller knows that the demand for a product is relatively high compared
to the supply, consumers might lack that knowledge. To eliminate such information
asymmetry, a seller can present soldout products in order to signal to consumers the
otherwise unobservable information about market demand. This information allows
consumers to distinguish between a vibrant and a stagnant market. In a decision scenario
in which consumers fail to consider the dynamics of product demand and supply, they
often decide to take the ―no-choice‖ option, i.e., option to defer a purchase decision by
choosing to search for information about additional options, or simply by choosing to
wait (Tversky and Shafir 1992; Dhar 1997). With the soldout products signaling that an
offer might not last long on the market, consumers are more inclined to expedite their
purchase decisions by choosing an available option.
It is noteworthy that while perceived unavailablity can imply scarcity, which
might increase the desirability of an object (Lynn 1992), we argue that preference
9
elicitation by soldout products does not have to be an indispensable step toward the
choice of an object. With the soldout products serving as a cue to the relationship
between demand and supply that might be otherwise neglected in the decision process,
consumers tend to choose an available option without necessarily adjusting their
evaluation of that option. On the basis of this discussion, we offer the following
hypothesis:
H1: The soldout products will create an immediacy effect according to which the
presence of soldout products will prompt consumers to choose an available option
rather than defer their purchase decisions.
A considerable body of research focuses on the influence of the composition of an
offered set on people’s choice between the options in the set (Huber et al. 1982;
Simonson 1989; Tversky and Simonson 1993). A subset of this research examined the
attraction effect. As discussed earlier, recent advances in research have demonstrated the
robustness of the attraction effect by extending this effect to the context of unavailable
options (Doyle et al. 1999; Fitzsimons 2000; Pratkanis and Farquhar 1992). In this
research, we propose that the soldout products will influence consumers’ choice between
the available options through a related process—the compromise effect, i.e., an
alternative’s choice probability will increase when it becomes a middle option in the
product distribution.
Under preference uncertainty, such as when consumers are unfamiliar with the
decision situation, they might not know how to evaluate different market offerings. Under
such circumstances, a safe option for the consumer to choose appears to be the one that
the average consumer will choose. Wernerfelt (1995) argued that consumers who do not
10
know their absolute but only their relative tastes can infer the correct choice from market
offerings, based on an intuitive assumption that these reflect the distribution of tastes in
the population. Thus, consumers will consider the option in the middle range of the
product distribution a reasonable choice. However, consumers often only know that the
distribution of products comes from a certain family of distributions (e.g. normal
distribution), but may not know the exact shape of the distribution (e.g. mean and
dispersion). As the soldout products represent options that did exist in the marketplace,
they convey relevant information about more prices and assortment of the product
continuum. For example, when buying a house in an unfamiliar market, information
about the real estates that have been recently sold out helps to provide a more complete
picture of the market. Combining the information about the soldout products with that
about available options enables consumers to form a better understanding of the product
distribution. Specifically, the presence of extreme soldout products helps consumers to
estimate the precise location of the available options in the product continuum, serves as
the reference point to identify the available option that falls in the middle of the product
distribution, and facilitates consumers’ choice between the available options. In
conclusion, we propose that the compromise effect can be extended to the context of
soldout products, as follows:
H2: The presence of extreme soldout products enables consumers to identify the
available option that falls in the middle of product distribution and increases the choice
incidence of the middle option.
11
EXPERIMENT 1
The goal of Experiment 1 is to examine whether an immediacy effect is triggered
by a soldout product in the decision environment (as predicted by Hypothesis 1).
Method
Subjects. Seventy-eight undergraduate students in a subject pool participated for
partial course credit. They did so in a university research lab, and in groups of less than
fifteen per session.
Materials. All stimuli were presented on a computer screen, and participants
entered their responses using a computer mouse and keyboard. Participants were asked to
consider whether or not they would like to take advantage of a special offer of a ski pass.
Two ski pass alternatives were used in this experiment, one sold at $20 for 5 hours and
the other sold at $40 for 10 hours. One of the two ski passes was presented as the soldout
product, and the other as the available option. The selection of the soldout product was
counterbalanced. Each of the two ski passes was indicated as the soldout product 50
percent of the time (See Table 1).
Design and procedure. We used a one-factor, two-level, between-subjects design.
The key experimental manipulation consisted of the presence versus absence of a soldout
product. In the soldout condition, participants were presented with two ski passes, one of
which had a tag attached to it that read ―Sorry, Sold Out.‖ (See Table 1) Participants
were then asked to decide whether they would like to purchase the available option
(―Yes, I will purchase‖ vs. ―No, I will not purchase‖). In addition, participants were
12
asked to evaluate the attractiveness of the two options on a seven-point scale (where 1 =
not at all attractive and 7 = very attractive).
The absence-of-soldout condition was the control condition. Under this
condition, participants were only presented with the available option, which was the same
available product as in the soldout condition (See Table 1). Participants were asked to
decide whether they would like to purchase the product. In addition, participants were
asked to evaluate the attractiveness of this product.
Table 1: Experimental Stimuli in Experiment 1
Soldout condition Absence-of-soldout condition
Ski Pass 1: $20 (for 5 hrs)
Ski Pass 2: $40 (for 10 hrs)
– Sorry, Sold out
Ski Pass : $20 (for 5 hrs)
Or (counterbalancing case)
Soldout condition Absence-of-soldout condition
Ski Pass 1: $40 (for 10 hrs)
Ski Pass 2: $20 (for 5 hrs)
– Sorry, Sold out
Ski Pass : $40 (for 10 hrs)
Results
First, we compared the attractiveness rating data of the two ski pass alternatives in
the soldout condition. The mean rating of the alternative consisting of a 5 hour pass for
$20 was 4.00, and the mean rating of the alternative consisting of a 10 hour pass for $40
was 4.18, which is not significantly different (t (38) = -0.61, p = .55, two-tailed). Since
participants perceived the two ski pass alternatives as equally attractive, we conclude that
any differences in the tendency to expedite choice (i.e., select the available option) are
not due to perceived differences in attractiveness of the two alternatives.
13
Next, we tested whether the sample choice share of the no-choice option (i.e.,
option to defer the purchase decision) is different in the two experimental conditions
using a logistic regression within GLM analysis (generalized linear model) in R. The
share of the no-choice option vs. the available option as a function of whether the soldout
product is absent or present in the decision environment is presented in Figure 1. In the
absence-of-soldout condition, 64.1 percent (25 out of 39) of the participants selected the
no-choice option. In the soldout condition, 38.5 percent (15 out of 39) of the participants
selected the no-choice option, which is significantly smaller (z = -2.24, p = .01, one-
tailed). These results demonstrate that the presence of a soldout product creates a sense of
urgency which prompts consumers to make a purchase of the available options rather
than defer the purchase decision, a phenomenon we refer to as the immediacy effect.
Figure 1
Exp. 1: Choice Share of No-choice Option vs. Available Option
38.5%
64.1%61.5%
35.9%
0%
10%
20%
30%
40%
50%
60%
70%
Soldout Absence-of-soldout
Ch
oic
e S
har
e
No-choice Option
Available Option
Discussion
Unlike prior research that deliberately constructed the unavailable option to be
clearly inferior or superior to one of the available options so that it caused a preference
shift for the latter (Walster and Festinger 1964; Farquhar and Pratkanis 1993; Fitzsimons
14
2000, Study 3), our experiment involved a soldout product that was equally attractive as
the available option (by introducing the same unit price of an activity, i.e., $4/hour). The
purpose of introducing an equally attractive soldout product was to rule out the
possibility that a comparison with an inferior soldout product would provide a reason for
participants to choose an available option (as in Simonson 1989; Tversky and Simonson
1993). Rather, in this experiment we aimed to demonstrate the immediacy effect—i.e., a
tendency to choose an available option simply due to felt urgency triggered by the
soldout product, an effect independent of perceived value of the available option relative
to the soldout product.
We realize, however, that introducing an equally attractive soldout product leaves
open an alternative explanation, according to which consumers are following herd
behavior. That is, in Experiment 1, participants might decide to choose the available
option with greater frequency simply because they realize that other people have chosen a
very similar alternative, i.e., the soldout product.
Banerjee (1992) and Bikhchandani et al. (1992) independently proposed the
―informational cascades‖ framework to interpret herd behavior. They showed that when
individuals need to make decisions using incomplete or ambiguous information, they tend
to assume that other people hold more valuable information than themselves, and thus
infer product quality from other people’s choices. Thus, when participants felt uncertain
about the value of the available option, other people’s choice of a highly similar
alternative can become informative to the participants and increase their preference for
the available option. If this ―informational cascades‖ account of herd behavior holds, it is
15
expected that the presence of a soldout product should increase people’s evaluation of the
available option.
To examine the possibility of the information cascade argument, we compared the
attractiveness rating data of the available option in the two experimental conditions.
(Specifically, we compared the mean of the attractiveness ratings of the available options,
i.e., the 5 hour pass for $20 and the 10 hour pass for $40, in the soldout condition with
that in the absence-of-soldout condition. This differs from our earlier comparison of the
attractiveness ratings between the 5 hour pass for $20 and the 10 hour pass for $40, under
the soldout condition.) The mean rating of the available option in the soldout condition
was 4.18, and that of the available option in the absence-of-soldout condition was 3.51.
The difference between the two numbers is marginally significant (t (76) = 1.72, p = .09,
two-tailed). This analysis suggests that the presence of a highly similar soldout product
tends to increase people’s evaluations of the available option to some extent. Based on
these results, herd behavior, in principle, could account for the pattern of results in this
experiment, rather than (or in addition to) an immediacy effect, as suggested earlier.
However, the explanation of information cascades is more of an experimental
artifact as it is contingent on the condition that the soldout product and the available
option were highly similar, if not interchangeable alternatives, so that participants were
able to make inferences about the value of the available option by observing other
people’s choice of the soldout product. In the next experiment, we investigate whether
the soldout product can still generate a higher choice share of the available option even
when the soldout product is from a different product category than the available option.
16
EXPERIMENT 2
Experiment 1 showed greater choice share for the available option when a soldout
option was present, which could arise either from an immediacy effect or from an
information cascades effect, or from both. The goal of Experiment 2 is to examine
whether the immediacy effect by itself is sufficient to expedite consumers’ purchase
decisions. To isolate the immediacy effect from an information cascades effect, we
introduce a soldout product from a different product category than the available option.
The rationale behind this is that participants can not transfer other people’s preference for
a product to their evaluations of a product in a totally different category. Nonetheless, the
presence of the soldout product in the decision environment can instill a sense of urgency
by causing participants to pay attention to a potentially high demand for the available
option.
Method
Subjects. Seventy-five undergraduate students in a subject pool participated for
partial course credit. They did so in a university research lab, and in groups of less than
fifteen per session.
Materials. All stimuli were presented on a computer screen, and participants
entered their responses using a computer mouse and keyboard. Participants were asked to
consider whether or not they would like to take advantage of a special offer from a store
on campus. Two promotion items were used in this experiment, a $20 value gift card sold
at $16.50, and an $18 value bus tickets available at $15. One of the two products was
presented as the soldout product, and the other as the available option. The selection of
17
the soldout product was counterbalanced. Each of the two products was indicated as the
soldout product 50 percent of the time (see Table 2).
In this experiment, the soldout product and the available option were from two
different product categories. In this case, the soldout product conveyed no information
about other consumers’ preferences for an available option in a different category.
Therefore, any observed effect of the soldout product can not be attributed to the
influence of information cascades.
Design and procedure. The procedure was similar to that employed in Experiment
1. We used a one-factor, two-level, between-subjects design. The experimental
manipulation consisted of the presence versus absence of the soldout product.
In the soldout condition, participants were presented with two promotion items (a
$20 value gift card sold at $16.5, and a $18 value bus tickets sold at $15), one of which
had a tag attached to it which read ―Sorry, Sold Out‖ (See Table 2). Participants were
then asked to decide whether they would like to purchase the available option (―Yes, I
will purchase‖ vs. ―No, I will not purchase‖). In addition, participants were asked to
evaluate the attractiveness of the products on a seven-point scale (where 1 = not at all
attractive and 7 = very attractive).
The absence-of-soldout condition was the control condition, under which
participants were only presented with the same available option as in the soldout
condition (See Table 2). Participants were asked to decide whether they would like to
purchase this option. In addition, participants were asked to evaluate the attractiveness of
this option.
18
Table 2: Experimental Stimuli in Experiment 2
Soldout condition Absence-of-soldout condition
Promotion Item 1: $20 value gift card for
$16.50
Promotion Item 2: Ten bus tickets for a
total of $15 (Original price $18) – Sorry,
Sold out
Promotion Item: $20 value gift card for
$16.50
Or (counterbalancing case)
Soldout condition Absence-of-soldout condition
Promotion Item 1: Ten bus tickets for a
total of $15 (Original price $18)
Promotion Item 2: $20 value gift card for
$16.50 – Sorry, Sold out
Promotion Item: Ten bus tickets for a
total of $15 (Original price $18)
Results and Discussion
First, we compared the attractiveness rating data of the two promotion items in the
soldout condition. The mean rating of the gift card was 3.16, which was significantly
different from that of 3.84 of the bus tickets (t (36) = -2.73, p = .01, two-tailed),
indicating that participants assigned different evaluations to the two items that pertained
to different product categories.
In addition, we compared the attractiveness rating data of the available option in
the two experimental conditions. (Specifically, we compared the mean of the
attractiveness ratings of the available options, i.e., the $16.50 gift card and the $15 bus
tickets, in the soldout condition with that in the in the absence-of-soldout condition.) The
mean rating of the available option in the soldout condition was 3.57, and that of the
available option in the absence-of-soldout condition was 3.43. The difference between
the two numbers is not significant (t (73) = 0.32, p = .75, two-tailed). As expected, these
19
results demonstrate that the soldout product did not have an influence on participants’
evaluation of the available option.
Next, and most importantly, we tested whether the sample choice shares of the
no-choice option, i.e., option to defer the purchase decision, is different in the two
experimental conditions using a logistic regression within GLM analysis (generalized
linear model) in R. As shown in Figure 2, in the absence-of-soldout condition, 50.0
percent (19 out of 38) of the participants selected the no-choice option. In the soldout
condition, 27.0 percent (10 out of 37) of the participants selected the no-choice option,
which was significantly smaller than when a soldout product was absent (z = -2.02, p =
.02, one-tailed). Therefore, even when the soldout product pertained to a different product
category, it still prompted participants to expedite their purchase decision by accepting
the available option and rejecting the no-choice option.
These results suggest that the mere presence of a soldout product in the decision
environment tends to lead people to choose an available alternative. Whereas the
information cascades argument can provide an alternative explanation for Experiment 1,
it can not provide an alternative explanation for Experiment 2. In the latter experiment,
the soldout product is in a different product category than the available option, so the
soldout product is not informative of other people’s attitudes toward the available option
for participants to follow. But the soldout product can still create a sense of immediacy
and reduce the tendency to defer a purchase decision. Together, Experiments 1 and
Experiment 2 demonstrate that an immediacy effect is triggered by the presence of a
soldout product in consumers’ purchase environment. Thus, Hypothesis 1 is supported.
20
Figure 2
Exp. 2: Choice Share of No-choice Option vs. Available Option
27.0%
50.0%
73.0%
50.0%
0%
10%
20%
30%
40%
50%
60%
70%
80%
Soldout Absence-of-soldout
Ch
oic
e S
ha
re
No-choice Option
Available Option
Experiment 3
The first two experiments demonstrated the immediacy effect of the soldout
product on consumers’ purchase decisions, whereby consumers are less inclined to defer
a purchase decision in the presence of a soldout product. The goal of Experiment 3 is to
investigate another important question: how the soldout product influences consumers’
choice among the available options.
An abundant body of research has been devoted to understanding the influence of
the composition of an offered set on consumers’ choice among the alternatives in the set,
i.e., how the preference for an alternative is dependent on the relative position of the
alternative compared with other alternatives in the offered set (Huber et al. 1982;
Simonson 1989; Tversky and Simonson 1993; Wernerfelt 1995). A subset of this
contextual influence is referred to as the compromise effect, which describes the
phenomenon that an alternative’s choice share will increase when it becomes a
compromise, or the middle alternative, following the introduction of an extreme
alternative into the offered set.
21
We propose that these findings can be extended to the decision environment with
a soldout product. When consumers face uncertainty in determining the absolute values
of available options, probably due to lack of experience or product knowledge, a safe
strategy to take is to choose a middle option from the product continuum that reflects
preferences of an average consumer (Wernerfelt 1995). Just as a more extreme option in
the offered set, a soldout product at a far end of the product distribution helps consumers
learn about prices and assortment of the market offerings, as well as learn the locations of
the available options in the product distribution. As a result, an extreme soldout product
helps consumers identify in the offered set the option that falls in the middle range of the
product distribution, and that corresponds to average preference. In this Experiment, we
test whether participants are more inclined to choose the middle option with the presence
of an extreme soldout product in the decision environment (as predicted by Hypothesis
2).
Method
Subjects. Two hundred and thirty one undergraduate students in a subject pool
participated for partial course credit. They did so in a university research lab, and in
groups of less than fifteen per session.
Materials. All stimuli were presented on a computer screen, and participants
entered their responses using a computer mouse and keyboard. Participants were asked to
consider purchasing a wine for a party. Three different wine alternatives were used in this
experiment: (1) 2004 Burgess Napa Merlot available at $5.99, (2) 2002 St. Jean Napa
Merlot available at $13.99, and (3) 2000 Chappellet Napa Merlot available at $21.99. The
22
first two wines were presented as available options, and the third, most expensive wine
was indicated as the soldout product (See Table 3).
Design and procedure. We used a one-factor, three-level, between-subjects
design. The key experimental manipulation consisted of the presence of a soldout
product, the absence of a soldout product, and the presence of a soldout product that is
initially not described explicitly, but concerning which participants can choose to search
for further information.
In the soldout condition, participants were presented with three wine options:
2004 Merlot available at $5.99, 2002 Merlot available at $13.99, and 2000 Merlot
available at $21.99. The third, highest priced alternative had a tag attached to it which
read ―Sorry, Sold Out‖ (See Table 3). Participants were then asked to decide which of the
two available options they would like to purchase (Wine 1 versus Wine 2). In the search-
of-soldout condition, participants were presented with the same two available wine
options as in the soldout condition. The third alternative had a tag attached to it which
read ―Sorry, Sold Out‖. However, the product description was hidden. In order to search
for information about the soldout product, participants need to click on a button that read
―Click here to view Wine 3‖ (See Table 3). If participants decided not to pursue this
information, they proceeded to make a choice between the two available options right
away. Otherwise, they clicked on the button, the information about the soldout product
was returned – yielding the same screen that those in the soldout condition were
presented with, and participants then made a choice between the two available options
given the description of the soldout product. The absence-of-soldout condition was the
control condition, under which participants were only presented with the two available
23
options, and were asked to decide which one they would like to purchase (See Table 3).
In all three conditions, all participants were asked to rate how knowledgeable they
considered themselves to be with wine after making a purchase decision (on a seven-
point scale where 1 = not at all knowledgeable and 7 = very knowledgeable).