Content uploaded by Markus Weinmann
Author content
All content in this area was uploaded by Markus Weinmann on Jun 03, 2020
Content may be subject to copyright.
The Decoy Effect in Reward-Based Crowdfunding
Thirty Seventh International Conference on Information Systems, Dublin 2016 1
The Decoy Effect in Reward-Based
Crowdfunding: Preliminary Results
from an Online Experiment
Research-in-Progress
Matthias Tietz
University of Liechtenstein
Fuerst-Franz-Josef-Strasse,
9490 Vaduz, Liechtenstein
matthias.tietz@uni.li
Alexander Simons
University of Liechtenstein
Fuerst-Franz-Josef-Strasse,
9490 Vaduz, Liechtenstein
alexander.simons@uni.li
Markus Weinmann
University of Liechtenstein
Fuerst-Franz-Josef-Strasse,
9490 Vaduz, Liechtenstein
markus.weinmann@uni.li
Jan vom Brocke
University of Liechtenstein
Fuerst-Franz-Josef-Strasse,
9490 Vaduz, Liechtenstein
jan.vom.brocke@uni.li
Abstract
Rewards are one of the key mechanisms in crowdfunding, but we know little about how
they influence fundraising success. This research-in-progress takes a behavioral-science
perspective on crowdfunding and explores how to design reward menus to increase the
chances of reaching funding targets. We show that backers’ preferences can be influ-
enced with the help of “decoys”—asymmetrically dominated rewards that draw backers’
attention to more profitable rewards. We conducted an online experiment with forty
participants to pre-test the effect of decoy rewards on crowdfunding success for three
scenarios. Across all scenarios, the decoys increased the donations by approximately
11 percent. Mixed-effects logistic regression analysis confirmed the significance of the
decoy effect. We are currently developing a mock crowdfunding website that will pro-
vide a more realistic environment in which to test alternative decoy-placement strate-
gies in other crowdfunding scenarios with more participants and more rewards.
Keywords: Crowdfunding, decoy effect, rewards, behavioral science
The Decoy Effect in Reward-Based Crowdfunding
Thirty Seventh International Conference on Information Systems, Dublin 2016 2
Introduction
The Internet provides promising new opportunities for business funding, so increasing numbers of
startups use the Web to collect the money they need to turn their business ideas into reality―a funding
practice commonly referred to as “crowdfunding” (Bradford 2012; Schwienbacher and Larralde 2012).
Crowdfunding is also becoming increasingly important in the creative industries for funding movies, mu-
sic, design, and other arts (Agrawal et al. 2014). While conventional funding practices typically involve
only a few large investors, crowdfunding websites allow entrepreneurs and artists to raise contributions
from large numbers of small investors (Ahlers et al. 2015; Belleflamme et al. 2013; Wheat et al. 2013).
The crowdfunding industry has grown tremendously during the past few years. While $6.1 billion in
crowdfunding was raised in 2013, an estimated $34.4 billion was raised in 2015 (Massolution 2015).
Crowdfunding is expected to exceed even the venture-capital funding in 2016 (The Economist 2016).
However, in spite of crowdfunding’s global success, many projects and ventures have remained unfunded.
Only 35 percent of all campaigns launched on Kickstarter, one of the world’s oldest and largest crowd-
funding sites (Kuppuswamy and Bayus 2015), have been successfully funded (Kickstarter 2016a). While
many projects remained unfunded for good reason, some promising projects also failed (Kunz et al. 2016).
Against this background, researchers have started to explore the factors beyond project quality that de-
termine crowdfunding campaigns’ success, including funding goals (e.g., Koch and Siering 2015), the du-
ration of the funding campaign (e.g., Mollick 2014), project description like text and videos (e.g., Kunz et
al. 2016; Zhou et al. 2015), communication like updates and comments (e.g., Müllerleile and Joenssen
2015; Xu et al. 2014), and characteristics of the project’s ownership like gender and social network (e.g.,
Rhue 2015; Zvilichovsky et al. 2015).
Despite the increasing attention that the design and presentation of projects is receiving in research on
crowdfunding success, the rewards offered for donating have not yet been studied in sufficient depth. In
reward-based crowdfunding, backers are offered non-financial prizes (e.g., being credited in a movie, a
visit to the film set, or a DVD or Blu-ray) in exchange for their money (Mollick 2014). With the increasing
use of reward-based crowdfunding websites like Kickstarter or Indiegogo, rewards have become one of
the key mechanisms in crowdfunding, especially for creative projects (Agrawal et al. 2014; Belleflamme et
al. 2013). Some researchers have studied how the number of rewards influences fundraising success (e.g.,
Frydrych et al. 2014; Kunz et al. 2016), but the selection of rewards and the determination of the donation
amounts to which they are connected have largely been neglected. Xiao et al. (2014) provided first evi-
dence that higher maximum donations and fewer reward tiers lead to significantly higher success rates,
but there is a great deal more to learn about how to design reward menus—a considerable knowledge gap,
considering that rewards are the primary mechanism with which to incent backers to provide funding. To
contribute to filling this gap, our research takes a behavioral-science perspective on crowdfunding and
explores how to design reward menus that increase the chances of funding success.
As a starting point, this research-in-progress evaluates the significance of the decoy effect (Huber et al.
1982) in reward-based crowdfunding. We present preliminary results from an online experiment with
forty participants and pre-test the hypothesis that backers’ preferences change when a reward is added
that creates an asymmetrically dominated choice, that is, a choice in which one of at least two rewards is
more valuable than a third reward in all relevant dimensions (i.e., quality and price) (Bateman et al.
2008). The third reward serves as a decoy; although virtually no backer ever chooses the decoy, it increas-
es the number of times the more valuable reward is chosen (Bateman et al. 2008; Huber et al. 1982). Be-
cause a more valuable reward can be designed to have a higher donation amount than the other rewards,
the use of decoys can help crowdfunding campaigns reach their funding targets.
First, we provide background on reward-based crowdfunding and then explain how decoys can be used to
increase fundraising success. Next, we outline the methods we used to collect and analyze the experi-
mental data and present the results. Finally, we discuss implications and limitations and provide an out-
look on our future research agenda.
Reward-Based Crowdfunding
The term “crowdfunding” refers to the collection of relatively small amounts of money from a large num-
ber of people, typically through Internet websites (Ahlers et al. 2015; Bradford 2012; Schwienbacher and
The Decoy Effect in Reward-Based Crowdfunding
Thirty Seventh International Conference on Information Systems, Dublin 2016 3
Larralde 2012). Crowdfunding is closely related to crowdsourcing (Howe 2006), but it collects money
instead of ideas and feedback. During the past few years, crowdfunding websites have emerged on the
Internet that follow a variety of funding models—donation-based, lending-based, equity-based, and re-
ward-based (Bradford 2012; Bretschneider et al. 2014; Frydrych et al. 2014; Haas et al. 2014). While do-
nation-based crowdfunding websites (e.g., ammado, GlobalGiving, and JustGiving) are typically used for
charity-related projects, lending-based crowdfunding (e.g., Kiva, Lending Club, and Prosper) has become
popular as a source of private credit, equity-based crowdfunding (e.g., Crowdcube, EquityNet, and
StartupValley) has emerged for small-business funding, and reward-based crowdfunding (e.g., Kickstart-
er, Indiegogo, and Crowdfunder) is used for creative projects (Sharp 2014). These four crowdfunding
models are distinguished based on what backers receive in return for their money: nothing (donation-
based), interest (lending-based), shares or dividends (equity-based), or a non-financial reward (reward-
based) (Bretschneider et al. 2014; Mollick 2014; Kunz et al. 2016).
Reward-based crowdfunding has become the most popular crowdfunding model on the Internet (Belle-
flamme et al. 2013; Mollick 2014). Most reward-based crowdfunding websites work as follows (Koch and
Siering 2015): Project creators post profiles regarding their professional backgrounds and crowdfunding
activities (e.g., previous projects, backing history) and use standardized templates to describe their pro-
jects (e.g., texts, images, videos). The funding-campaign’s duration is fixed, a funding goal is also set, and
backers are offered non-financial rewards in return for their money. Funding goals, campaign durations,
and project descriptions cannot usually be changed after project launch, but project creators can post
updates to inform actual and potential backers about a project’s progress. Rewards vary with the amount
of the donation and include such offers as copies of the resulting product (e.g., a DVD of the film), creative
collaboration (e.g., a role in the film), creative experiences (e.g., a visit to the film set), and creative me-
mentos (e.g., thanks in the film credits) (Kuppuswamy and Bayus 2015). Many reward-based crowdfund-
ing websites follow an “all-or-nothing” principle in which projects are realized only if a predefined funding
goal is reached. Another popular crowdfunding principle is “keep-it-all,” in which project creators keep
whatever they collect, even if they do not reach their fundraising goals (Cumming et al. 2014).
Researchers have explored how these design options influence crowdfunding success. In particular, pro-
ject characteristics like the funding goal and the campaign’s duration have received considerable atten-
tion. While higher funding goals lower the chance that a project will be funded successfully (e.g., Koch and
Siering 2015), longer campaign durations have been found to decrease funding success rates (e.g., Mollick
2014), but research has also delivered contradictory findings (e.g., Frydrych et al. 2014). Project descrip-
tions have also been studied in depth (e.g., Gao and Lin 2015; Koch and Siering 2015; Kunz et al. 2016;
Xiao et al. 2014). For example, Marom and Sade (2013) found that project descriptions that focus on the
personage or business idea are positively associated with fundraising success, a finding that Zhou et al.
(2015) confirmed in terms of argument quality and source credibility and that Koch and Siering (2015)
confirmed in terms of texts, images, and videos. In addition, fundraising success is a function of project
ownership, as Zvilichovsky et al. (2015) and Koch and Siering (2015) demonstrated, as being an active
backer of other projects increases one’s own fundraising success. Finally, communication in, for example,
the form of comments, updates, and blog posts has also been confirmed as a success factor (e.g., Kunz et
al. 2016; Xiao et al. 2014; Xu et al. 2014).
The literature review shows that researchers have studied several factors related to the design and presen-
tation of crowdfunding projects, but only a few have studied how rewards influence funding success, and
the results have been contradictory. For example, Frydrych et al. (2014) could not provide clear implica-
tions of the relationship between the number of rewards and crowdfunding success, while Kunz et al.
(2016) provided evidence that the number of rewards increases the probability of success, and Xiao et al.
(2014) concluded that fewer reward tiers lead to significantly higher success rates. Against this back-
ground, the selection of rewards and the determination of the donation amounts to which they are con-
nected have not been studied in sufficient depth. As the next section explains, our research takes a behav-
ioral-science perspective on crowdfunding to explore how to design rewards menus that increase the
chances of reaching funding targets.
The Decoy Effect
The objective of crowdfunding campaigns is to collect enough money to fund a particular project, so an
important question for project initiators concerns how to encourage backers to select the rewards that
The Decoy Effect in Reward-Based Crowdfunding
Thirty Seventh International Conference on Information Systems, Dublin 2016 4
come with the highest donations. Research in the field of behavioral science has demonstrated that peo-
ple’s choices are influenced by how the choices are presented (Johnson et al. 2012). Even simple modifica-
tions of the choice set can influence people’s decisions, so presenting choices in certain ways―modifying
the “choice architecture” (Thaler et al. 2013)―can “nudge” people and alter their behavior in a predictable
way (Thaler and Sunstein 2008), a concept that also applies to digital environments (Weinmann et al.
2015). In reward-based crowdfunding, the decoy effect—the effects of asymmetrical dominance in particu-
lar—may be used as a nudge to draw backers’ attention to the high-priced rewards.
Quality(rewarditems)
Price(US$)
Download Blu-ray Blu-ray and
download
10.00
20.00
30.00
40.00
Target
Competitor
1
2
3
Decoys
Fi
g
ure 1. Exem
p
lar
y
choice set for the
K
un
g
Fur
y
movie
(adapted from Huber et al. 1982)
Decoys can shift consumers’ preferences by adding an asymmetrically dominated alternative to the choice
set (Josiam and Hobson 1995). Most traditional choice models that explain how the introduction of a new
product changes existing products’ market shares incorporate the two fundamental hypotheses of similar-
ity, where the new product takes disproportionally more market share from similar products than it does
from dissimilar products, and regularity, where the new product does not increase the market share of
other available products (see Huber et al. 1982 for a discussion). However, Huber et al. (1982) argued that
these two hypotheses fail when an asymmetric alternative is offered. An asymmetrically dominated choice
refers to situations in which one of at least two options (that do not dominate each other) dominates a
third option in all relevant dimensions (e.g., price and quality) (Bateman et al. 2008). Adding an asym-
metric alternative like the third option to a choice set increases the number of times the dominant option
will be chosen, which violates the regularity assumption and reverses the similarity assumption, so an
asymmetric alternative serves as a “decoy” that increases the share of the dominant, more profitable op-
tion (Huber et al. 1982).
Research has confirmed the significance of the decoy effect in several consumer-behavior situations, in-
cluding the selection of restaurants, beer, cars, and tour packages (e.g., Heath and Chatterjee 1995;
Hedgcock et al. 2009; Huber et al. 1982; Josiam and Hobson 1995). Some studies have also delivered
contradictory findings (e.g., Kim et al. 2006), but the applicability of decoys in reward-based crowdfund-
ing remains to be assessed. In reward-based crowdfunding, rewards vary regarding quality and price,
which are the two dimensions that determine backers’ choices. The martial arts comedy movie Kung Fu-
ry, which was crowdfunded through Kickstarter, is an example of a crowdfunding campaign that can be
used to illustrate how decoys can work in reward-based crowdfunding.
The movie project offered more than thirty rewards to backers, from general project support for US$ 1.00
to a major film role for US$ 10,000 (Kickstarter 2016b). For the purposes of this illustration, and ground-
ed in marketing research on hybrid products and bundling (e.g., Bakos and Brynjolfsson 1999; Koukova et
al. 2008; Venkatesh and Chatterjee 2006), let us assume that the project involved only two basic rewards:
a download option at US$ 20.00 and a Blu-ray-and-download option at US$ 30.00. (Most of Kickstarter’s
movie projects offer at least these two types of rewards, with higher-tier rewards usually designed to in-
clude some lower-tier rewards.) As the Blu-ray-and-download option is more expensive, it is also the pref-
erable option from the filmmakers’ viewpoint, so it is the higher-priced, higher-quality “target,” using
The Decoy Effect in Reward-Based Crowdfunding
Thirty Seventh International Conference on Information Systems, Dublin 2016 5
decoy language, while the download option is the lower-priced, lower-quality “competitor” (Huber et al.
1982). As Figure 1 shows, neither the competitor nor the target dominate the other in terms of quality or
price, so the choice set could benefit from adding a decoy within the shaded area.
Table 1 shows three decoy-placement strategies for the Kung Fury case. The lower-value decoy may offer
the same reward at a higher price (e.g., US$ 40.00 instead of US$ 30.00), a worse reward at the same
price (e.g., only the Blu-ray, no download option), and a combination of these two strategies, a worse re-
ward at a higher price (e.g., only the Blu-ray for US$ 40.00). Decoys are rewards that no one would logi-
cally choose but that make the Blu-ray-and-download reward look better (Huber et al. 1982). There are
two reasons that backers may prefer this reward when a decoy is added to the choice set: the perceptual
framing of the decision problem—such as when adding a higher-priced decoy makes the Blu-ray appear
less expensive— and the evaluation processes used—such as when the easy choice between the target and
the decoy makes it less likely that the download-only option will be considered (Bateman et al. 2008; Hu-
ber et al. 1982).
Price (US$)
Quality (reward items)
Strategy 1
Target 30.00 Blu-ray and download
Competitor 20.00 Download
Added decoy 40.00 Blu-ray and download
Strategy 2
Target 30.00 Blu-ray and download
Competitor 20.00 Download
Added decoy 30.00 Blu-ray
Strategy 3
Target 30.00 Blu-ray and download
Competitor 20.00 Download
Added decoy 40.00 Blu-ray
Table 1. Decoy-placement strategies for the Kung Fury movie
(adapted from Huber et al. 1982)
The next section explains how we tested placing decoys in crowdfunding projects using the second strate-
gy of a less valuable reward at the same price as that of a more valuable target reward.
Research Method
Participants and experimental design. As our study investigates crowdfunding, an online activity,
we conducted an online experiment. We did not collect data about real crowdfunding campaigns from the
Internet (Huhtamäki et al. 2015), as most previous studies have done, because the use of decoy rewards is
still an uncommon crowdfunding practice. We recruited forty native English speakers of at least eighteen
years of age from prolific.ac1 to ensure sufficient comprehension skills. (Prolific allows researchers to filter
for participants who were born in the English-speaking countries of the United Kingdom, the United
States, Ireland, Australia, Canada, or New Zealand.) The mean age of all participants was 33.9 years, and
65.0 percent of them were men. We conducted a single-factor repeated-measure experimental design with
two conditions: a baseline condition and a decoy condition. We randomly assigned the forty participants
to the baseline condition and the decoy condition, with each condition consisting of the same three crowd-
funding scenarios, resulting in 120 observations (60 observations per condition). Each session lasted an
average of six minutes. We excluded participants from the dataset who needed less than four minutes to
complete the experiment (which made it unlikely that they have thoroughly read and understood the in-
structions provided), which resulted in a final dataset of 96 observations (48 observations per condition).
Subjects received £1 for participation, the approximate average hourly wage in the UK of £10.
1 Online recruitment platforms have been found to be appropriate for random-sample populations (Berinsky et al. 2012). For exam-
ple, Mason and Suri (2012) found that the behavior of respondents on an online recruiting platform closely resembled that of par-
ticipants in traditional laboratory experiments.
The Decoy Effect in Reward-Based Crowdfunding
Thirty Seventh International Conference on Information Systems, Dublin 2016 6
Materials and procedure. We used three Kickstarter projects to design the crowdfunding scenarios
for the experiment and created project descriptions of similar length (Table 2). These scenarios incorpo-
rated a book, a video game, and a movie, so they were designed to represent some of the most common
project categories on creativity-oriented crowdfunding websites like Kickstarter. We used Qualtrics.com,
an online survey website, for presentation purposes. Participants read the experiment’s instructions and
then went through the three scenarios, which were presented in random order. Our experiment was a
hypothetical thought experiment, so participants were asked to select the funding options they would
chose in real life. After they completed the scenarios, they were surveyed to collect demographic data.
Project Description
Book2
The children’s book Allen and the Wolf Pack tells the story of Allen, a nine-year old boy who wants
to discover the reason for an unnaturally long winter. Although it is end of May, spring has not
come, winter storms continue to bluster, and the winds carry the voices of howling wolves. When
Allen goes to the forest to find the reason for this confused state, he gets lost, but he survives with
the help of wolves that care for him until he is rescued. The book is full of adventure, mystery,
excitement, and great characters.
Video game3
Heading to Armageddon is a third-person-shooter PC game that takes place in the twenty-second
century. The world is separated into three unevenly powerful blocs that have competed for years:
the Western Alliance, the Russian Regime, and the Restored Caliphate. A nuclear disaster is immi-
nent and deemed unavoidable, and the future of humankind is in the player’s hands. The player is
the leader of an underground movement whose mission is to rescue the world from the apocalypse
by locating all of the world’s nuclear devices and destroying them.
Movie4
Ambulance Affairs is a thriller movie about Dave, a young ambulance technician in New York, on
his first shift with his partner Michelle. When they return to the station in their ambulance after
their shift, they encounter an injured person lying in the road. They offer assistance but are taken
hostage at gunpoint by a group of hooded men, who take them and the injured person to a nearby
building and force them to keep the injured patient alive. It turns out that the men are wanted for
robbing the Federal Reserve Bank in Manhattan.
Table 2. Descriptions of the three crowdfunding scenarios
Treatment. The reward-choice alternatives in each project category consisted of a target, a competitor,
and a decoy. In each scenario, the competitor option was a comparatively cheap reward—a digital-
download option for the book, video game, or movie. The more expensive target options included the digi-
tal-download option as well as physical versions of the product. (To determine realistic prices for these
two options, we reviewed various project descriptions available on Kickstarter.) The decoy option had the
same price as the target option in all scenarios but did not include the digital-download reward, so we
followed the second of the decoy-placement strategies described above. Participants in the baseline condi-
tion decided between the competitor and the target options, while we added the decoy as a third option to
the decoy condition. Table 3 provides an overview of the choice sets for the book scenario.
Option Baseline condition Decoy condition
Competitor PAY $10 – GET an eBook PAY $10 – GET an eBook
Decoy — PAY $20 – GET a hardcover book
Target PAY $20 – GET an eBook and a hardcover book PAY $20 – GET an eBook and a hardcover book
Table 3. Choice sets for the book scenario
Measures. Our outcome variable was binary—participants in the baseline condition could choose either
the cheap reward (0 = competitor) or the expensive reward (1 = target). Likewise, participants in the de-
coy condition could choose between the competitor (cheap) and the target (expensive), or, though unlike-
2 The scenario is based on the project Ellen and the Winter Wolves: https://www.kickstarter.com/projects/567406064/ellen-and-
the-winter-wolves-by-jamin-still
3 The scenario is based on the project Road to Armageddon – A Modern Military Role Playing Game:
https://www.kickstarter.com/projects/1009649146/road-to-armageddon-a-modern-military-role-playing?ref=category
4 The scenario is based on the project Ambulance—A Short Film: https://www.kickstarter.com/projects/1178516301/ambulance-a-
short-film?ref=category
The Decoy Effect in Reward-Based Crowdfunding
Thirty Seventh International Conference on Information Systems, Dublin 2016 7
ly, they could decide to take the decoy option (which was also expensive). In line with previous research
(Huber et al. 1982), we merged the decoy and target choices into an expensive choice.
Results
In all three scenarios, a higher number of participants selected the expensive option when the decoy was
added to the choice set (although four participants selected the decoy). Figure 2 presents histograms for
the three funding scenarios, and Table 4 compares the revenues for the three scenarios. In total, the de-
coys generated additional revenues of US$ 120.00, an increase of approximately 11 percent.
Observations
Figure 2. Choices in the three scenarios
We conducted a logistic regression to test the effect of the treatment (i.e., the decoy) on the likelihood that
participants would choose the expensive option (Decision = 1). Because we tested three scenarios per
participant, we had multiple observations for each participant, so we had to assume that the residuals are
not independent but clustered within participants. Since observations from the same participant might be
correlated (Gelman and Hill 2006), we used mixed-effects logistic regression analysis to account for
individual effects (i.e., by allowing the intercept of each participant to vary):
Pr(Decisions = 1) = logit–1 (αi[s] + β1 · Decoys + ui + γ’ · Controlsi),
where i[s] indexes the individuals i that correspond to each scenario’s observations s (i.e., book, video
game, movie), αi[s] represents the individual intercept, β1 is the effect of the Decoys in each scenario, Con-
trolsi are the control variables Age, Gender, and Scenario, and ui is a random effect designed to capture
the correlation between the decision (i.e., the outcome variable) observed for the same subject i across the
scenarios s.
Total revenues
Scenario Condition
Baseline Decoy
Book $210 $270
Video game $460 $490
Movie $395 $425
Total $1,065 $1,185
Table 4. Revenues across scenarios
Table 5 presents the results of three mixed-effects logistic regression models. Model 1 contains only the
decoy variable, while Model 2 contains the decoy, the control variables (Age and Gender), and an interac-
tion term between Decoy and Scenario.
In another specification (Model 3), we added Scenario as a random factor5 and allowed both the intercept
of individuals and the coefficient of Scenario (i.e., the slopes) to vary (i.e., varying-intercept, varying-slope
model):
Pr(Decisions = 1) = logit–1 (αi[s] + βi[s] · Decoys + ui + γ’ · Controlsi),
where βi[s] refers to the individual slopes for Scenario.
5 An effect is said to be random if the study contains only a random sample of possible conditions (Field 2013, p. 862). For example,
the variable Scenario can be considered random because we could have used other/more scenarios than book, video game, and
movie.
The Decoy Effect in Reward-Based Crowdfunding
Thirty Seventh International Conference on Information Systems, Dublin 2016 8
Dependent variable: Decision (binary)
(1) (2) (3)
Treatment (decoy) 2.43 * (1.04) 5.31 * (4.12) 5.30 * (4.12)
Intercept .41 ** (.13) .29 (.30) .29 * (.30)
Age — 1.03 (.02) 1.03 (.02)
Gender — .69 (.33) .69 (.33)
Interaction (decoy * scenario) NO YES (n.s.) YES (n.s.)
Random intercept (participant) YES YES YES
Random slope (scenario) NO NO YES
N 96 96 96
-2LL 124.5 118.3 118.3
AIC 128.5 134.3 134.3
BIC 133.6 154.8 154.8
Notes: * p < .05, ** p < .01, *** p < .001; standard errors are in parentheses.
Table 5. Mixed-effects logistic regression results
Other than the treatment variable (Decoy), none of the predictor variables (controls and interaction ef-
fects) were statistically significant. We chose Model 1 after performing a likelihood ratio test.6 The results
show that, in Model 1, participants in the decoy condition were 2.43 times more likely to choose the ex-
pensive option than were participants in the baseline condition. (In Models 2 and 3, participants were 5.3
times more likely to choose the expensive option). Accordingly, our early-stage experiment confirmed the
applicability of decoys in reward-based crowdfunding.
Discussion
Our results suggest that the use of decoy rewards can help crowdfunding campaigns increase the size of
their average donations. No one is expected to choose a decoy, as they offer the same reward as another
funding option at a higher price, a less valuable reward at the same price, or even a less valuable reward at
a higher price. Adding a decoy to the set of funding options can make an expensive reward appear more
attractive and help crowdfunding campaigns reach their funding targets—and reach them more quickly.
(We do not advocate their use for ethically questionable behavior; see Sunstein (2015) for a discussion.)
Our early-stage research on using decoys in crowdfunding has several limitations. First, with forty partici-
pants, the sample size was small. Second, we recruited our participants through an online survey platform
and did not collect data from real crowdfunding campaigns. Third, we tested the decoy effect using three
fictitious and simplified crowdfunding campaigns. As a result, our findings provide only preliminary evi-
dence for the usefulness of decoys in crowdfunding, as 1) there are several strategies for decoy placement,
and we tested only one of them; 2) the applicability of decoys in other crowdfunding categories than
books, video games, and movies remains to be evaluated; 3) participants did not self-select into support-
ing a crowdfunding campaign and did not invest real money; 4) we tested the possibility of decoy place-
ment for a limited scheme of only three rewards; 5) crowdfunding sites offer various features related to
project design that have been found to influence crowdfunding success, but our experiment accounted for
only some of them; and 6) we excluded several individual-level variables from our study that influence
backing behavior, including personal preferences and level of income.
6 We used the likelihood ratio test to compare the model specifications against an intercept-only model. Model 1 was statistically
significant (χ2(1) = 4.27, p < .039) compared to an intercept-only model. Adding control variables and the interaction effect did not
significantly improve the model fit (Model 2: χ2(7) = 9.36, p < .228), nor did adding Scenario as random effect increase the model
fit significantly (Model 3: χ2(7) = 9.36, p < .228).
The Decoy Effect in Reward-Based Crowdfunding
Thirty Seventh International Conference on Information Systems, Dublin 2016 9
The goal of this early-stage research was to focus on internal validity by isolating the decoy effect in a sim-
ple, controlled experiment. To establish ecological validity, we are currently developing a mock crowd-
funding website that provides a more realistic environment in terms of look and feel and functionality in
order to test alternative decoy-placement strategies in other crowdfunding scenarios with more partici-
pants, a higher number of rewards, and a variety of design features. If successful, our study will have im-
portant implications for crowdfunding practice and research, as project creators can use our results to
select and offer rewards in pursuit of funding targets, while researchers can use our results to theorize on
reward-based crowdfunding, especially from a behavioral-science perspective. For information systems
research, our results can inform the design of reward menus and crowdfunding projects as well as the
design of crowdfunding websites. For example, the use of individual, adaptive rewards (e.g., related to
level of income and interest) is a promising direction for future design research.
This research-in-progress is part of a larger research endeavor that will identify and test additional mech-
anisms that may influence choice behavior in crowdfunding settings, including the middle-option bias
(Simons et al. 2017). Kamenica (2008) showed that people tend to take the “middle” option, so designing
a choice set with a desirable middle target reward may increase donations. In addition, Kahneman et al.
(1991) demonstrated that people tend to stick with the default option, so preselecting target options in
crowdfunding contexts may raise funding as well. Finally, the prospect theory suggests that people tend to
weigh losses more heavily than gains (Kahneman and Tversky 1979), so they may prefer to choose more
expensive rewards if they are offered in limited number. The mock crowdfunding website we are currently
developing will provide the technical infrastructure required to evaluate the significance of these and re-
lated effects in reward-based crowdfunding.
References
Agrawal, A., Catalini, C., and Goldfarb, A. 2014. “Some Simple Economics of Crowdfunding,” Innovation
Policy and the Economy (14:1), pp. 63-97.
Ahlers, G. K. C., Cumming, D., Günther, C., and Schweizer, D. 2015. “Signaling in Equity Crowdfund-
ing,” Entrepreneurship Theory and Practice (39:4), pp. 955-980.
Bakos, Y., and Brynjolfsson, E. 1999. “Bundling Information Goods: Pricing, Profits, and Efficiency,”
Management Science (45:12), pp. 1613-1630.
Bateman, I. J., Munro, A., and Poe, G. L. 2008. “Decoy Effects in Choice Experiments and Contingent
Valuation: Asymmetric Dominance,” Land Economics (84:1), pp. 115-127.
Belleflamme, P., Lambert, T., and Schwienbacher, A. 2013. “Individual Crowdfunding Practices,” Venture
Capital (15:4), pp. 313-333.
Berinsky, A. J., Huber, G. A., and Lenz, G. S. 2012. “Evaluating Online Labor Markets for Experimental
Research: Amazon.com’s Mechanical Turk,” Political Analysis (20:3), pp. 351-368.
Bradford, C. S. 2012. “Crowdfunding and the Federal Securities Laws,” Columbia Business Law Re-
view (1), pp. 1-150.
Bretschneider, U., Knaub, K., and Wieck, E. 2014. “Motivations for Crowdfunding: What Drives the
Crowd to Invest in Start-Ups?” in Proceedings of the 22nd European Conference on Information Sys-
tems (ECIS), Tel Aviv, Israel.
Cumming, D. J., Leboeuf, G., and Schwienbacher, A. 2014. “Crowdfunding Models: Keep-It-All Vs. All-Or-
Nothing,” retrieved September 7, 2016, from https://www.eurofidai.org/Schwienbacher_2014.pdf.
Field, A. 2013. Discovering Statistics Using IBM SPSS Statistics, 4th ed., London, England: Sage Publica-
tions.
Frydrych, D., Bock, A. J., Kinder, T., and Koeck, B. 2014. “Exploring Entrepreneurial Legitimacy in Re-
ward-Based Crowdfunding,” Venture Capital (16:3), pp. 247-269.
Gao, Q., and Lin, M. 2015. “Lemon or Cherry? The Value of Texts in Debt Crowdfunding,” Social Science
Research Network (SSRN), retrieved September 7, 2016, from http://ssrn.com/abstract=2446114.
Gelman, A., and Hill, J. 2006. Data Analysis Using Regression and Multilevel/Hierarchical Models, New
York, NY: Cambridge University Press.
Haas, P., Blohm, I., and Leimeister, J. M. 2014. “An Empirical Taxonomy of Crowdfunding Intermediar-
ies,“ in Proceedings of the 35th International Conference on Information Systems (ICIS), Auckland,
New Zealand.
The Decoy Effect in Reward-Based Crowdfunding
Thirty Seventh International Conference on Information Systems, Dublin 2016 10
Heath, T. B., and Chatterjee, S. 1995. “Asymmetric Decoy Effects on Lower-Quality Versus Higher-Quality
Brands: Meta-Analytic and Experimental Evidence,” Journal of Consumer Research (22:3), pp. 268-
284.
Hedgcock, W., Rao, A. R., and Chen, H. 2009. “Could Ralph Nader’s Entrance and Exit Have Helped Al
Gore? The Impact of Decoy Dynamics on Consumer Choice,” Journal of Marketing Research (46:3),
pp. 330-343.
Howe, J. 2006. “Crowdsourcing: A Definition,” retrieved September 7, 2016, from
http://crowdsourcing.typepad.com/cs/2006/06/crowdsourcing_a.html.
Huber, J., Payne, J. W., and Puto, C. 1982. “Adding Asymmetrically Dominated Alternatives: Violations of
Regularity and the Similarity Hypothesis,” Journal of Consumer Research (9:1), pp. 90-98.
Huhtamäki, J., Lasrado, L., Menon, K., Kärkkäinen, H., and Jussila, J. 2015. “Approach For Investigating
Crowdfunding Campaigns With Platform Data: Case Indiegogo,” in Proceedings of the 19th Interna-
tional Academic Mindtrek Conference, Tampere, Finland.
Johnson, E. J., Shu, S. B., Dellaert, B. G. C., Fox, C., Goldstein, D. G., Häubl, G., Larrick, R. P., Payne, J.
W., Peters, E., Schkade, D., Wansink, B., and Weber, E. U. 2012. “Beyond Nudges: Tools of a Choice
Architecture,” Marketing Letters (23:2), pp. 487-504.
Josiam, B. M., and Hobson, J. S. P. 1995. “Consumer Choice in Context: The Decoy Effect in Travel and
Tourism,” Journal of Travel Research (34:1), pp. 45-50.
Kahneman, D., Knetsch, J. L., and Thaler, R. H. 1991. “Anomalies: The Endowment Effect, Loss Aversion,
and Status Quo Bias,” The Journal of Economic Perspectives (5:1), pp. 193-206.
Kahneman, D., and Tversky, A. 1979. “Prospect Theory: An Analysis of Decision Under
Risk,” Econometrica (47:2), pp. 263-291.
Kamenica, E. 2008. “Contextual Inference in Markets: On the Informational Content of Product Lines,”
American Economic Review (98:5), pp. 2127-2149.
Kickstarter. 2016a. “Stats,” retrieved September 7, 2016, from https://www.kickstarter.com/help/stats.
Kickstarter. 2016b. “Kung Fury,” Kickstarter, retrieved September 7, 2016, from
https://www.kickstarter.com/projects/kungfury/kung-fury/description.
Kim, J., Park, J., and Ryu, G. 2006. “Decoy Effects and Brands,” Advances in Consumer Research (33:1),
pp. 683-687.
Koch, J.-A., and Siering, M. 2015. “Crowdfunding Success Factors: The Characteristics of Successfully
Funded Projects on Crowdfunding Platforms,” in Proceedings of the 23rd European Conference on
Information Systems (ECIS), Münster, Germany.
Koukova, N. T., Kannan, P. K., and Ratchford, B. T. 2008. “Product Form Bundling: Implications for
Marketing Digital Products,” Journal of Retailing (84:2), pp. 181-194.
Kunz, M. M., Englisch, O., Beck, J., and Bretschneider, U. 2016. “Sometimes You Win, Sometimes You
Learn—Success Factors in Reward-Based Crowdfunding,” in Proceedings of the Multikonferenz
Wirtschaftsinformatik (MKWI), Ilmenau, Germany.
Kuppuswamy, V., and Bayus, B. L. 2015. “Crowdfunding Creative Ideas: The Dynamics of Project Backers
in Kickstarter,” Social Science Research Network (SSRN), retrieved September 7, 2016, from
http://ssrn.com/abstract=2234765.
Marom, D., and Sade, O. 2013. “Are the Life and Death of an Early Stage Venture Indeed in the Power of
the Tongue? Lessons from Online Crowdfunding Pitches,” Social Science Research Network (SSRN),
retrieved September 7, 2016, from http://ssrn.com/abstract=2255707.
Mason, W., and Suri, S. 2012. “Conducting Behavioral Research on Amazon’s Mechanical Turk,” Behavior
Research Methods (44:1), pp. 1-23.
Massolution. 2015. “2015CF: The Crowdfunding Industry Report,” retrieved September 7, 2016, from
http://reports.crowdsourcing.org/?route=product/product&product_id=54.
Mollick, E. 2014. “The Dynamics of Crowdfunding: An Exploratory Study,” Journal of Business Ventur-
ing (29:1), pp. 1-16.
Müllerleile, T., and Joenssen, D. W. 2015. “Key Success-Determinants of Crowdfunded Projects: An Ex-
ploratory Analysis,” Data Science, Learning by Latent Structures, and Knowledge Discovery, B.
Lausen, S. Krolak-Schwerdt, and M. Böhmer (eds.), Heidelberg, Germany: Springer, pp. 271-281.
Rhue, L. 2015. “Who Gets Started on Kickstarter? Demographic Variations in Fundraising Success,” in
Proceedings of the 36th International Conference on Information Systems (ICIS), Fort Worth, TX.
Schwienbacher, A., and Larralde, B. 2012. “Crowdfunding of Entrepreneurial Ventures,” The Oxford
Handbook of Entrepreneurial Finance, D. Cumming (ed.), New York, NY: Oxford University Press,
pp. 369-391.
The Decoy Effect in Reward-Based Crowdfunding
Thirty Seventh International Conference on Information Systems, Dublin 2016 11
Sharp, A. M. 2014. “Crowdfunding Success Factors,” International Research Journal of Applied Finance
(5:7), pp. 822-832.
Simons, A., Weinmann, M., Tietz, M., and vom Brocke, J. 2017. “ Which Reward Should I Choose? Prelim-
inary Evidence for the Middle-Option Bias in Reward-Based Crowdfunding,” in Proceedings of the
50th Hawaii International Conference on System Sciences (HICSS), Big Island, HI.
Sunstein, C. R. 2015. “Nudging and Choice Architecture: Ethical Considerations,” Yale Journal on Regu-
lation, forthcoming, Social Science Research Network (SSRN), retrieved September 7, 2016, from
http://ssrn.com/abstract=2551264.
Thaler, R. H., and Sunstein, C. R. 2008. Nudge: Improving Decisions About Health, Wealth, and Happi-
ness, London, England: Penguin Books.
Thaler, R. H., Sunstein, C. R., and Balz, J. P. 2013. “Choice Architecture,” The Behavioral Foundations of
Public Policy, E. Shafir (ed.), Princeton, NJ: Princeton University Press, pp. 428-439.
The Economist. 2016. “Economic Milestones of the Year Ahead,” retrieved September 7, 2016, from
http://www.economist.com/blogs/graphicdetail/2016/01/daily-chart.
Venkatesh, R., and Chatterjee, R. 2006. “Bundling, Unbundling, and Pricing of Multiform Products: The
Case of Magazine Content,” Journal of Interactive Marketing (20:2), pp. 21-40.
Weinmann, M., Schneider, C., and vom Brocke, J. 2015. “Digital Nudging,“ Social Science Research Net-
work (SSRN), retrieved September 7, 2016, from http://ssrn.com/abstract=2708250.
Wheat, R. E., Wang, Y., Byrnes, J. E., and Ranganathan, J. 2013. “Raising Money for Scientific Research
Through Crowdfunding,” Trends in Ecology & Evolution (28:2), pp. 71-72.
Xiao, S., Tan, X., Dong, M., and Qi, J. 2014. “How to Design Your Project in the Online Crowdfunding
Market? Evidence from Kickstarter,” in Proceedings of the 35th International Conference on Infor-
mation Systems (ICIS), Auckland, New Zealand.
Xu, A., Yang, X., Rao, H., Fu, W.-T., Huang, S.-W., and Bailey, B. P. 2014. “Show Me the Money! An Anal-
ysis of Project Updates During Crowdfunding Campaigns,” in Proceedings of the SIGCHI Conference
on Human Factors in Computing Systems (CHI), Toronto, Canada.
Zhou, M., Du, Q., Zhang, X., Qiao, Z., Wang, A. G., and Fan, W. 2015. “Money Talks: A Predictive Model
on Crowdfunding Success Using Project Description,” in Proceedings of the 21st Americas Conference
on Information Systems (AMCIS), Puerto Rico.
Zvilichovsky, D., Inbar, Y., and Barzilay, O. 2015. “Playing Both Sides of the Market: Success and Reci-
procity on Crowdfunding Platforms,” Social Science Research Network (SSRN), retrieved September
7, 2016, from http://ssrn.com/abstract=2304101.
