ANTECEDENTS OF FUNDRAISING SUCCESS AND ENTREPRENEUR
IAL
 
PERFORMANCE IN CROWDFUNDING PLATFORMS
 
 
By
 
 
Eun Ju Jung
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
A DISSERTATION
 
 
Submitted to
 
Michigan State University
 
in 
partial fulfillment of the requirements
 
for the degree of
 
 
Business Administration
 
-
 
Business Information Systems
 
-
 
D
octor of Philosophy
 
 
201
5
 
 
 
 
 
ABSTRACT
 
 
ANTECEDENTS OF FUNDRAISING SUCCESS AND 
ENTREPRENEUR
IAL
 
PERFORMANCE IN CROWDFUNDING PLATFORMS
 
 
By
 
 
Eun Ju Jung
 
The emergence of crowdfunding could 
provide 
novel 
opportunities 
for startup
 
companies
 
and 
potentially 
transform the nature of entrepreneurship and new firm creation. 
The prevailing
 
modes 
of venture financing
 
limit the
 
types of business projects that get funded or limit the range 
of project choices availability to funder. In contrast, crowdfunding platforms present a diversified 
array of project choices and funds so that not just mai

prevail.
 
Crowdfunding platforms could connect 
entrepreneurs and funders
, and provide
 
new 
avenues for entrepreneu
rs to acquire legitimacy and 
engage with a 
broader 
pool of funders.
 
However, despite the growing volu
me of research, more insight is needed about (a) w
hat factors 
and
 
social d
ynamics explain
 
the ability 
to attract 
funding success
, and (b) how are the patterns of 
fundraising related to project success? 
Using the comprehensive theoretical lens of social net
work, 
electronic word
-
of
-
mouth, bandwagon effects, entrepreneurship and innovation theories, I address 
two research questions.
 

 
behaviors and online 
community 
engagement along with entre
preneur characteristics on fundraising success in 
crowdfunding platforms. We collect data about startup projects and funders from a reward
-
based 
crowdfunding platform in the U.S., as well as additional data from online social network sites and 
blogs. Our s
ample includes a total of 722 technology
-
related projects (March 2012 
-
 
January 2013) 
and more than 177,700 funders. The empirical results show that the networks of interactions 

nding community 
 
are key antecedents of funding success. Along with fundraising success in crowdfunding platforms, 
for start
-
ups both obtaining financial resources and creating innovation are important for their 
survival. However, there is lack of attention
 
as 
to whether successfully funded projects deliver 
outcomes. 
 
In my second essay, I investigate the existing dynamics in fundraising process, and how 
the fundraising patterns are related to crowdfunding projects performance. The current finding 
shows the 

contribute
 
to crowdfunding and entrepreneurship literature and
 
offer practical implications by 
providing a
 
comprehensive
 
theoretical framework and the supporting empiric
al evidence.
 
 
Keywords:
 
crowdfunding, network, community engagement, entrepreneurial performance, 
fundraising patterns
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Copyright by
 
EUN JU JUNG
 
2015
 
 
v
 
ACKNOWLEDGEMENTS
 
 
This dissertation would not have been completed without the generous 
support and 
encouragement of faculty, friends, and family members. I would like to express my 
g
ratitude to 
these individuals for their support and guidance throughout this long journey. 
 
 
First and foremost, I would like to express my deepest gratitude to 
my dissertation co
-
chairs
, Dr. Vallabh Sambamurthy and Dr. Anjana Su
s
arla for their invaluable guidance, unstinting 
support, encouragement, genuine caring and concern
 
throughout my entire
 
PhD 
journey
.
 
I have 
benefited from their unique perspectives and experiences in developing this dissertation and other 
research projects. Their enthusiasm and patience have made every step of my dissertation research 
an enjoyable learning process
. I am tr
uly blessed my PhD life to work with both of them and learn 
from them. 
Dr. Vallabh Sambamu
rthy, as a doctoral advisor, 
a mentor
,
 
and 
a 
friend
, he has made a 
tremendous impact on my academic and personal development. His rigorous supervision and 
insight
s
 
improved the quality o
f my dissertation significantly and his profound knowledge, 
experience, and professionalism have truly helped me build a solid foundation for my future as a 
researcher, a scholar, and 
a teacher.
 
I would also like to thank Dr. Anjana 
Susarla for her 
unwavering support and faith in me
,
 
and for her
 
genuine caring,
 
and
 
guidance
 
as an advisor, teacher, 
and friend. Without her 
help
 
and support, I would
 
not have come this far. I will
 
never forget the 
many wonderful lunches, research brainsto
rm meeting
s
, and the excitement of discovery after our 
long discussion
s at her office or her
 
house
.
 
 
I 
am 
heartily thankful to my committee members, Dr. Brian Pentland, Dr. Roger Calantone, 
and Dr. Yong Tan for providing 
insightful
 
comments, 
careful critiq
ue
s
, 
and 
suggestion
s
 
on my 
disserta
tion. Their academic support,
 
input and personal encouragement
 
are greatly appreciated. I 
vi
 
am
 
also thankful to my former college research advisor, Professor Chang
k
yo Suh, who guided me 
into the world of academia
.
 
T
hank
 
you
, too, to
 
my fellow doctor
al colleagues for their support
, es
p
ecially Yen Yao 
Wang, Kangkang Qi, Yu Huang, Dr. Jung 
Young Lee, 
Dr. ChangSeob Yeo, 
and 
Heyseong Koh
,
 
for 
their support and encouragement. 
I
 
am 
indebted to them
 
for providing a warm environment 
as
 
weathered
 
all the hard moment
s in the adventure
 
of getting a Ph.D degree.
 
T
his journey would not have been possible w
ithout the support and love of 
my family. I 
thank my parents, Cheon Sik Jung and Chae Yoon Lee for their unconditional love, perpetual f
aith 
and wholehearted support. I am forever grateful to them and 
for 
the opportunities they provided 
me. I also thank my sister, Yeon Woo Jung, and my brother Woo Cheo
l Jung for their love and 
inspiration, which helped
 
me 
through
 
all the
 
difficult times 
during my PhD years. I thank my 
parents
-
in
-
law, In Kyo Jung and Gum Bun Jung
,
 
for 
their trust in me, 
for 
cheering me with their 
warm hearts
,
 
and 
for 
their unwavering prayer
s
.
 
M
y 
special thanks go to my loving husband, Dr. EuiSung Jung, for being a best fri
end and 
true supporter. His 
unconditional love, genuine caring
, 
and never 
ending support
 
and 
encouragement throughout my PhD program
 
made this journey possible. 
 
 
 
vii
 
TABLE OF CONTENTS
 
 
LIST OF TABLES
 
................................
................................
................................
.........................
 
ix
 
LIST OF FIGURES
 
................................
................................
................................
.........................
x
 
CHAPTER 1. 
OVERVIEW
 
OF 
DISSERTATION
 
................................
................................
......
1
 
CHAPTER 2. 
 
ESSAY 1
: 
COMMUNITY ENGAGEMENT AND COLLECTIVE EVLAUATION IN 
CROWDFUNDING 
 
................................
................................
................................
.....................
2
 
 
2.1 
INTRODUCTION
 
................................
................................
................................
.................
2
 
 
2.2 
THEORETICAL BACK
GROUND AND RESEARCH HYPOTHESES
 
.............................
9
 
2.2.1 
Theo
retical Background
 
................................
................................
................................
9
 
2.2.2 
Information Asymmetry and Collecti
ve Evaluation in 
Crowdfunding
 
.......................
13
 
2.2.3 
Hypotheses Development
 
................................
................................
...........................
18
 
  

 
................................
................................
................................
...
18
 
  

th the Crowdfunding Community
 
................................
...
23
 
 
 

 
................................
................................
.........................
24
 
 
2.3 
DATA AND METHODS
 
................................
................................
................................
......
2
6
 
2.3.1 Data Collection
 
................................
................................
................................
...........
2
6
 
2.3.2 Network Analysis
 
................................
................................
................................
........
28
 
  
2.4 EM
PIRICAL MODEL
 
................................
................................
................................
.........
29
 
2.4.1 Funding Success Model
 
................................
................................
..............................
29
 
  
2.5 RESUTLS AND DISCUSSION
 
................................
................................
..........................
3
1
 
2.5.1 Results
 
................................
................................
................................
.........................
3
1
 
2.5.2 Robustness Checks
................................
................................
................................
......
39
 
  
Endo
geneity 
-
 
Instrument Variable
 
................................
................................
..................
39
 
  
Selection Bias 
-
 
Heckman Selection Model
 
................................
................................
.....
4
3
 
  
Sentiment Analysi
s of Electronic Word of Mout
h
 
................................
...........................
46
 
  
Heterogeneity of Projects
 
................................
................................
................................
.
47
 
  
2.6 CONCLUSION AND IMPLICATIONS
................................
................................
..............
49
 
CHAPTER 3. 
 
ESSAY 2
: 
DYNAMIC FUNDRAISING PATTERNS AND ENTREPR
E
NEURIAL 
PERFORMANCE IN CROWD
FUNDING
 
PLATFORMS
 
 
................................
...................
5
3
 
 
3.1 INTRODUCTION
 
................................
................................
................................
...............
5
3
 
 
3.2 
THEORETICAL BACKGROUND AND RESEAR
CH 
HYPOTHESES
 
...........................
6
0
 
3.2.1 
Literature Review
................................
................................
................................
........
6
0
 
  

 
................................
................................
................................
........
6
0
 
 
 

 
................................
................................
...........................
69
 
  
Project Characteristics
 
................................
................................
................................
.....
7
1
 
  
Moving Toward Conceptualization of 
Fundraising Patterns
 
in Crowdfunding
 
...............
7
2
 
     
3.2.2 
Research Model
 
................................
................................
................................
..........
7
6
 
viii
 
     
3.2.3 Hypotheses
 
Development
 
................................
................................
...........................
7
7
 
 
3.3 DATA
 
................................
................................
................................
................................
...
8
0
 
3.3.1 Data Collection
 
................................
................................
................................
...........
8
0
 
3.4 
METHODOLOGY
 
AND EMPIRICAL MODEL
 
................................
...............................
8
1
 
3.4.1 
Functional Data Analysis
 
................................
................................
............................
8
1
 
3.4.2 
Functional
 
Principal Component Analysis
................................
................................
..
8
3
 
     
3.4.3 Empirical Model
 
(Functional Regression Model)
 
................................
......................
8
4
 
  
Control variables
 
................................
................................
................................
..............
8
5
 
  
Measurement
 
................................
................................
................................
....................
8
6
 
  
3.5 
EMPIRICAL ANALYSIS AND RESULTS
 
................................
................................
.........
9
0
 
3.5.1 
Exploring the Fundraising Patterns
 
................................
................................
.............
9
0
 
3.5.
2
 
Results
 
................................
................................
................................
.........................
9
2
 
3.5.
3
 
Robustness Check
s
................................
................................
................................
......
99
 
  
Potential Endogeneity: Latent Instrumental Variable Approach
 
................................
......
99
 
  
H
eterogeneity of Projects: Cro
wd
-
 
vs. 
Expert
-
Based Funding
 
................................
......
104
 
  
3.6 CONCLUSION
 
AND IMPLICATIONS
................................
................................
............
109
 
REFERENCES
 
................................
................................
................................
...........................
 
11
2
 
 
 
 
ix
 
LIST OF TABLES
 
 
Table 2.
1
 
Current Crowdfunding Studies
 
................................
................................
......................
 
11
 
Table 2.2
 
Variable Definitions and Descriptive Statistics
 
................................
..............................
3
2
 
Table 2.3
 
The 
Correlation
s
 
of Variables
 
................................
................................
.........................
3
4
 
Table 2.4
 
Results
 
................................
................................
................................
............................
37
 
Table 2.5
 
Instrumental Variable Models
 
................................
................................
........................
40
 
Table 2.6
 

 
St
age
 
Selection Model
................................
................................
.........
4
5
 
Table 3.
1 
Literature Review
 
................................
................................
................................
...........
6
5
 
Table 3.2
 
Investors Types
 
................................
................................
................................
...............
69
 
Table 3.3
 
Fundraising Patterns and Characteristics
 
................................
................................
.......
7
5
 
Table 3.4
 
Frequency Table for Dependent Variables
 
................................
................................
.....
8
8
 
Table 3.5
 
Variable Definitions and Descriptive Statistics
 
................................
..............................
89
 
Table 3.6
 
Results of Initial Project Performance
 
................................
................................
...........
9
5
 
Table 3.7
 
Results of New Product into Markets
 
................................
................................
............
9
7
 
Table 3.8
 
Results of Growth/Financial Resource
 
................................
................................
..........
9
8
 
Table 3.9
 
MCMC Bayesian Results of Latent Instrument Model (Posterior
)
 
.............................
10
2
 
 
 
 
 
x
 
LIST OF FIGURES
 
 
Figure 1
.1 
Overview of Dissertation
 
................................
................................
................................
1
 
Figure 2
.1 Web Search Volume
 
................................
................................
................................
........
3
 
Figure 2
.2 The Ladder Model of Co
-
V
alue Creation in Crowdfunding
 
................................
........
15
 
Figur
e 2
.3 Two
-
sided 
Interact
ion in Crowdfunding Platforms
 
................................
......................
16
 
Figure 2.4 Research Model
 
................................
................................
................................
............
18
 
Figure 2
.5 

 
................................
...........................
27
 
Figure 2
.6 
Transpose 
T
wo
-
Mode Network to Two One
-
Mode
 
................................
......................
28
 
Figure 2
.7 
Distribution: Pledged Percentage and Project Goal
 
................................
.....................
33
 
Fig
ure 2
.8 
Geographical Distribution (Technology Projects)
 
................................
........................
38
 
Figure 2
.9 
% of Cumulative Amounts and Number of Projects
 
................................
....................
4
2
 
Figure 2
.10 
%
 
of Projects and Backers by Reward Types
 
................................
.............................
48
 
Figure 3
.1 
% of Projects by Reward Types
 
................................
................................
....................
7
2
 
Figure 3
.2 
Research Model
 
................................
................................
................................
............
7
6
 
Figure
 
3.
3
 
Smoothing Parameter Lambda
 

 
................................
.................
8
3
 
Figure 3.
4
 
Smoothing Sp
l
ines of Crowdfunding Projects
 
................................
.............................
9
0
 
Figure 3.
5
 
Examples of Funding Trajectory for Each Project
 
................................
.......................
9
1
 
Figure 3.
6
 
Examples of Different Fundraising Patterns  
 
................................
............................
9
2
 
Figure 3.7
 
Funding 
Principal Component Analysis
 
................................
................................
.......
9
3
 
Figure 3.8
 
Harmonics of FPCA Patterns
................................
................................
........................
94
 
Figure 3.9
 
Eigenvalue Number
 
................................
................................
................................
......
94
 
Figure 3
.
10 
Trace Plots and Density Plots
 
................................
................................
...................
103
 
xi
 
Figure 3
.
11
 
Plots of Crowd Vs. Expert
-
based Model
 
 
................................
................................
105
 
Figure 3
.
12 
Mean Curves for Crowd and Expert Types
 
................................
..............................
106
 
Figure 3
.
13 
Plots of estimated beta curves with functional confidence intervals
 
........................
107
 
Figure 3
.
14 
Plots of Permutation F
-
Test and P
-
Value
 
................................
................................
..
108
 
 
1
 
CHAPTER
 
1. OVERVIEW OF DISSERTATION
 
 
 
Using a two
-
essay 
format, I would like to investigate how entrepreneurs successfully raise 
money in a crowdfunding platform, and how dynamic fundraising patterns influence 
entrepreneurial performance. Essay 1 examines the im
pact of community engagement 
of
 
funders 
and entrep
reneurs, and an electronic word
-
of
-
mouth 
effect
, as well as the en
trepreneurial 
characteristics o
n
 
the probability of fundraising success. Essay 2 explores different dynamic 

rial improvisation, 
a
nd project outcome type, and
 
influence
 
entrepreneurial project performance and innovation 
performance. An ove
rview of the two essays 
is given below.
 
 
 
 
 
Figure 1.1 
Overview of 
Dissertation
 
 
 
 
2
 
CHAPTER 2.
 
ESSAY 1:
 
COMMUNITY
 
ENGAGEMENT
 
AND COLLECTIVE
 
EVALUATION 
IN CROWDFUNDING
 
 

Innovation is not led by lone inventors in their garrets but is the product of a collaborative 
process

 
--
Samuel J. Palmisano, 
former 
CEO
 
of
 
IBM
 
 
 
2.1 
INTRODUCTION
 
 
 
D
igitally mediated social networks have 
transformed social, economic, and business 
activities (Agarwal, Gupta, and Kr
a
ut, 2008; Aral, Dellarocas, and Godes, 2013). These networks 
foster social 
interactions and create new types of interactions, such as electronic ties with partners, 
interactions 
with customers, and collaboration within communities (Agarwal et al., 2008; Singh, 
Tan, and Mookerjee, 2011). Along with social and economic transformation, 
the
se networks 
engender new business models whereby companies can tap into collaborative platforms 
as a way 
to acquire resources and new ideas.
 
One such novel model of digital value creation is crowdfunding, 
whereby project creators from a variety of arenas can tap into a community to raise funding via 
the Internet (Schwienbacher and Larralde, 2010; 
Mollick, 2012).
 
Based on prior research, we 

allowing project creators (e.g., entrepreneurs) from various areas to request funding from many 
individuals via the I

or entrepreneur can create or fund projects in crowdfunding platforms. According to recent reports, 
there are around 452 crowdfunding platforms active worldwide. In 201
2
, these rais
ed almost $
2
.
7
 
3
 
billion and successfully funded more than one million projects (crowdsour
c
ing.org, 2012).
 
Kickstarter, one of the most popular crowdfunding platforms, had received about $450 million in 
funding in 
2012
1
. 
Figure 2
.1
 
shows the web search volum

from 2009 to 2014. The trend graph (of Kickstarter) shows a steep growth of web search volume. 
The 
Jumpstart Our Business Startups (
JOBS
)
 
A
ct of 2012 
in the U.S. 
allows 
startup companies
 
to 
initiate equity
-
based projec
ts in crowdfunding platforms, a move that could significantly ease the 
burden of raising funds for startup companies
 
as well as reduce barriers for investors to participate 
in 
entrepreneurial investment.
 
The Securities and Exchange Commission (SEC) has sim
ilarly 
moved to make regulatory changes that will enable equity participation through crowdfunding. 
 
 
Figure 
2
.
1
 
Web Search Volume
 
(Kickstar
ter: top & crowdfunding: bottom
, Google Trends , 2014
)
 
 
The emergence of c
rowdfundin
g
 
could unleash substantial 
changes in the
 
nature of 
entrepreneurship and new firm creation. 
Under traditional funding mechanisms, start
-
up 
companies have experienced considerable 
difficulties
 
in obtaining external funding (Gompers 
and
 
                                        
        
 
1
 
www.kickstarter.com/help/stats
 
4
 
Lerner, 2001; Hsu, 2007)
, and in creating
 
visibi
lity and establish
ing
 
legitimacy (Zimmerman and 
Zeitz, 2002
; 
Zott and Huy, 2007). Crowdfunding 
allows
 
entrepreneurs to reach crowds
 
of 
unprecedented scale at a low cost
 
an
d 
provides a variety of avenues to 
enhance legitimacy for 
necessary resources. For 
example,
 
Pebble
, one of most popular Kickstarter projects, raised $10.2 
million from 68,929 people within 37 days in 2012.
 
 
 
When investors lend, there is a problem of verifiability of effort or moral hazard. In 
traditional financing, lenders have an incen
tive to screen borrowers given the risk of default in 
lending to an inexperienced lender.
 
Professional funding firms such as venture capitalists also have 
developed expertise and managerial routines to evaluate potential entrepreneurs (Kirsch et al. 
2009).
 
Crowdfunding marketplaces lack features of institutional lending such as certification. In 
traditional funding markets, project creators invest in time and effort in building relationships with 
large net worth individuals or angel investors. While individ
ual investors make decisions on which 
projects to fund, in most crowdfunding platforms these decisions can only be executed once the 
projects reach a funding goal. That also means that for project creators, there is no upfront 
investment in building their 
company or creating a product; rather, they need to deliver the 

ensure that there is no moral hazard problem? How do funders mitigate information asymmetry?
 
C
row
dfunding
 
exhibits features of asymmetric information characteristic of Peer
-
to
-
peer 
lending since the lenders are mostly small investors and not as informed as institutional lenders in 
evaluating project quality and the capabilities of funders. According t
o Kickstarter Statistics
2
, 
fewer 
than 50% of projects reached their funding goals in 2012
. 
Further, there is also a lack of 
regulation relating to consumer protections, so investors have to rely on their judgment and the 
                                        
        
 
2
 
www.kickstarter.com/year/2012
 
5
 
collective evaluation of the crowd.
 
Question is also what underlying value can be added by the 
crowd? So
,
 
how do they choose investments? 
 
Recently, researchers have emphasized the critical role of co
-
value creation, a collaborative 
process
 
between customers and companies or among actors in
 
a various networks in innovations 
(Prahald and Ramaswamy, 2004; Nambisan and Sawhney, 2011; Lusch and Nambisan, 2015). Co
-
value creation not only facilitates relationship building but also creates shared 
knowledge
 
and 
experience
. Drawing on this perspecti
ve, customers are not passive recipients but are actively 
involved in innovation process (Nambisan and Sawhney, 2011). Information technology is a 
critical enabler of a co
-
value creation. Online communities and interactive web functions (web2.0) 
enable 
spo
ntaneous
 
conversation and richness of interactions among online community 
participants, and increase the 
reaches
 
of audience (Prahald and Ramaswamy, 2004; Ray, Kim, and 
Morris, 2014). 
 
A 
crowdfunding platform
 
supports two
-
sided social interaction in that t
he
 
digital platform 
facilitates interactions
 
both
 
between entrepreneurs and 
funders and 
interactions 
among
st
 
funders. 
Such interactions allow entrepreneurs to convey more information about projects as well as 
allowing 
funders 
to
 
share information about pro
jects. 
Crowdfunding platforms also provide 
visibility of projects and entrepreneurs

. Funders can easily 
obtain 
detailed project 
information such as a goal, duration of project, descriptions, a video, total 
pledged
 
amounts, 
entrepreneurs (pro
ject creators) and other funders

 
information, and p
roject cre
ators

embeddedness in the two
-
mode network of funders and creators. Entrepreneurs may be able to 
share both soft and hard information. Further, funders may persuade other funders to contribute 
money to entrepreneurs.
 
Thus, crowdfunding platforms provide with an online social space which 
enables 

connection

, interactions among funders and entrepreneurs or among funders, 
6
 

collaboration

, more than simple involvement (i.e., share experience and 
kn
owledge
 
about 
funding projects), and 

co
-
value creation

, evaluating and supporting entrepreneurs. 
We
 
suggest 
that the problems of information asymmetry in such settings can be mitigated by the collective 
evaluation mechanisms enabled by the online 
community aspects of crowdfunding
. 
 
P
rior research 
examin
ing social influence in crowdfunding has focused on the history of 
fundraising
 
process (Agarwal, Catalini and Gol
dfarb, 2011; Burtch et al., 2013
; Mollick, 2012) 
rather than 
the 
participants

 
social 
network 
and social interactions
.
 
Prior studies also 
pay less 
attention 

 
characteristics
, though they
 
have 
been important factors in venture 
capital investment. A theoretical explanation of the underlying phenomena will not only contrib
ute 
to related literatures, but will 
help
 
entrepreneurs who 
want
 
to 
initiate
 
projects and 
obtain
 
legitimacy,
 
as well as
 
in 
assist
ing
 
potential funders and platform designers. 
We draw upon multiple theoretical 
perspectives such as theories of social network
s, theories of online communities and theories of 

 
crowdfunding 
platform and 

The main research questions we
 
will
 
address are as foll
ows:
 
 
(1)
 

-
WOM) 
influence the likelihood of success in fundraising?
 
(2)
 

fundraising?
 
 
We examine our research questions
 
using data
 
collected
 
about startup projects and 
their 
funders
 
from multiple sources. We collected 
startup
 
projects 
and funder data
 
from 
Kickstarter, one 
of the most popular 
reward
-
based crowdfunding platform
s
 
in the U.S. Additional data were 
gathered from
 
online social network sites and blogs. 
Our sample includes a total of 
7
2
2
 
technology
-
7
 
related projects (
Mar
ch
 
2012 
-
 
January
 
2013)
 
and more than 177,700 funders. 
The empirical results
 
show that the 
embeddedness of funders contributing to a specific project, the external electronic 
wor
d of mouth (from outside the crowdfunding community) for a project and the 
engagement of 
entrepreneurs 
into the community of funders 
are key antecedents 
that a project i
s successfully 
funded
. Interestingly, these factors dominate the effects of the entrepreneurs

 
human capital
. 
 
 
Identifying the impacts of 
online
 
community 
engagement
 
on the probability of 
fundraising
 
success poses several econometric challenges. 
First, 
a 
critical issue in the estimation process is 
reverse causality between entrepreneurial 
engagement
 
(i.e., updating information about projects) 
and fundraising success. Not only the increase of updating actions influence fundraising, but also 
entrepreneurs wh
o get higher funding amounts may engage in updating more often. To deal with 
this issue, we used three sets of instruments which are correlated with entrepreneurial 
engagement 
actions
 
but are not correlated with fundraising performance.
 
S
econd, selection
-
b
ias
 
problem may 
exist in the research context. The unobserved characteristics of successfully funded projects may 
create estimation bias in the model of pledged percentage amounts model. We employ 

two stage
 
selection model to deal with selection 
bias problem. Additionally, to consider the 
potential heterogeneity in projects and e
-
WOM, we analyze the sentiment of e
-
WOM. 
 
This paper can make 
several contributions to the literature. First, we 
provide 
a 
comprehensive 
theoretical 
framework to understan
d fundraising mechanisms in a novel funding 
market. Second, 
although there are several studies on crowdfunding, the
 
extant
 
literature
 
mainly 
emphasizes the role of project attributes (i.e., goal, duration, 
demo materials such as 
video
s etc.
),
 
impacts of en

 
s
ocial network
s
, 
geographical
 
proximity, and 
contribution history in crowdfunding success
. In this paper, we explicitly focus on the factors 

network 
embeddedness
, 
f

8
 

of mouth effects and 
entrepreneurs

 
human capital
. Therefore, 
we 
contribute to 
multiple streams 
of research on 
crowdfunding, crowdsourcing, and entrepre
neurship literature
s
.
 
The 
rest of the 
paper proceeds as follows. Section 2 provides theoretical background and research hypotheses. In 
Section 3, we explain our sample data, data collection process, and introduce our empirical models 
of crowdfunding succes
s. Section 4 provides results of empirical test. In Section 5, we conclude 
by providing theoretical and 
managerial
 
implication
s
 
of study, limitation
s
 
of current study, and 
future directions.
 
 
 
 
9
 
2.2 
THEORETICAL BACKGROUND AND RESEARCH HYPOTHESES
 
 
2.2.1 
Theoretical Background 
 
 
In general, there are four types of crowdfunding: donation, reward, lending, and e
quity
-
based (Burtch et al., 2013
). Academic research has begun to investigate the antecedents of funding 
success and contributor

t al. (2013

initial
 
contributions on lat
t

-
based funding. 
Ordanini et 
al.
 
(2011) investigate how participants

 
motivation
s
 
vary
 
across different platforms. 
Mollick (2012) 
ana
lyzed the impacts of project quality, social network of project creators, and project 
characteristics on funding success. Agarwal
 
et al. 
(2011
;
 
2013
) presented the positive effect
s
 
of 
offline relationships on funding investments, and the decreased effects
 
of geographical distance 
between investors and artist
-
entrepreneurs on investments. 
Kuppuswamy and Bayus (2013) 
examine
 
how
 

 
support on Kickstarter varies depending on project characteristics and 
timing.
 
Table 2
.1 summarizes 
existing
 
crowdfunding s
tudies. 
 
 
Recently, 
the 
innovation literature ha
s
 
emphasized the critical role of co
-
value creation, a 
collaborative 
process
 
between customers and companies or among actors in a various networks in 
innovations (Prahal
a
d and Ramaswamy, 2004; Nambisan and Sawhney, 2011; Lusch and 
Nambisan, 2015). Information technology is a critical enabler of a co
-
value 
creation
, which 
enables 
spontaneous
 
conversation and richness of interactions among online community 
participants and i
ncreases the 
reaches
 
of audience (Ray, Kim, and Morris, 2014). 
Researchers
 
have 
also 
investigated the role of online community with interactive web technology in innovation and 
firm performance (Pr
ahalad and Ramaswamy, 2004; Ray et al.,
 
2014; Ma and Agarwa
l, 2007). 
Online platforms facilitate opportunities for two
-
sided markets by bringing together different 
10
 
groups of users within a platform and creating economic values for both sides. Several studies 
have looked at the price structure strategies for two
-
si
ded markets, network externalities of one
-
side network, and cross
-
side network effects in online platforms (Eisenmann, Parker, and Van 
Alstyne, 2011). For a given user, the value of the platform depends largely on the number of users 

 
side; and each side typically plays a distinct ro
le in the platforms 
(Eisenmann et al.,
 
2011). 
However, transactions in two
-
sided networks often also involve sets of 
social relations. Thus, two
-
sided platforms also facilitate a network of user social inte
ractions, 

Lakhani (2013) suggest that 
there is an essential dichotomy between the roles of a platform in enabling contests vs. in enabling 
a collaborative community.
 
Prior research suggest
s that inte
rpersonal contact
 
is especially 
important in persuasion and information dissemination, and online platforms provide participants 
with initial contact points for an innovation (Phelps et al., 2004).
 
Therefore, rather than elaborating 
on the econo
mic perspective of two
-
sided networks, we focus on the social relationships of two
-
sided participants in crowdfunding platforms.
 
A considerable literature on online communities 

o and Faraj
,
 
2005), 
the role of member attachment (Ren et al.
,
 
2012) and the ability of an online community to enforce 
norms and sanctions (Chua et al.
,
 
2007).
 
Ren et al.(2012) distinguish between member attachment 
that arises from attachment to a group fr
om that of interpersonal bonds that center member 
attention on individuals. Using their typology, the member attachment in crowdfunding platforms 
such as Kickstarter exhibits both features of identity
-
based attachment, wherein there is 
considerable communi
ty engagement and communication as well as attachment to individual 

decisions.
11
 
Table 2.1 Current Crowdfunding Studies
 
Reference
 
Independent Variable
 
Dependent
 
Variable
 
Key Finding
 
Mollick (2012)
 

 
Project quality
 

 
Project characteristics 
 

 
Social network
 
Funding 
Success
 

 
Antecedents of funding success
 

 
Significant role of Project quality (video) and social network on 
funding success
 
Agrawal et al. 
(2011, 2013)
 

 
Geographical distance 
 

 
Social network (prior 
offline relation)
 

 
Project characteristics 
 
Investment 
value
 

 
Local and distant investment patterns
 

 
A reduced role for spatial proximity (distance is still significant)
 

 
Online platform does not eliminate social
-
re
lated frictions. Local 
investors invest relatively early
 

 
The geography effect is driven by investors who likely have a 
personal connection with the artist entrepreneur
 
Schwienbacher 
& Larralde 
(2010)
 

 
Compare pre
-
ordering 
mechanism vs. profit 
sharing mechanism 
 
Reward 
mechanism
 

 

increase their utility
 

 
The entrepreneur prefers pre
-
ordering if the initial capital 
requirement is relatively small, and p
rofit
-
sharing otherwise
 
Bellefalammet 
et al.(2010)
3
 

 
Modelization of 
crowdfunding: pre
-
ordering, price 
discrimination,
 
No empirical 
test
 

 
Rewards mechanism
 

 
The conditions under which crowdfunding in 
preferred to 
traditional forms 
of external funding (1) 
for
-
profit vs. non
-
profit; 
(2) choice of funding method (3) customers as investors
 
 
 
 
                                        
        
 
 
 
12
 
Table 
2.1
 

 
Gerber, Hui, 
and Kuo 
(2012)
 

 


(1) autonomy (2) 
competence(3) 
relatedness
 
Participation
 

 
Identify six creator and four
 
supporter motivations
 
Creators: raise funds, expand awareness of work, connect, gain 
approval, maintain control, to learn (unmotivated to ask for money, 
spend time, fail, gain publicity, waste money)
 

 
Supports: to collect external rewards, to help others, 
to be part of a 
community, to support a causes analogous with their personal beliefs 
(unmotivated: to wait for or not receive rewards, to be pestered for 
support
 
Ordanini et al. 
(2009)
 

 

 

 
Platform 
characteristics
 
 
Participation
 

 
Crowdfunding participants : (1) purposes: patronage, investment, 
social participation; (2) characteristics: innovative orientation, 
identification, exploitation; (3) roles and task: agent, 
shareholder(growth and development), donor(help); (4) investment 
si
ze: small, large, small
 

 
Crowdfunding firm: (1) purpose: empower artist and fans, raise 
alternative venture capital, fund social projects online; (2) service 
roles: relational mediator, engine of growth, social gatekeeper; (3) 
network effects: substitute an
 
existing intermediary, disintermediate 
from an existing intermediary, add a new intermediary
 
Lin et al. 
(2012)
 

 
Hard (credit 
information) vs. soft 
information(social 
capital)
 
Funding success
 

 
Significant impacts of social capital
 
Burtch et al. 
(2013)
 

 
Contribution 
frequency
 
Contribution 
amounts
 

 
Significant role of accumulated amounts on additional funding
 
Kuppuswamy 
and Bayus 
(2013)
 

 
Contribution pattern 
and timing
 

 
Project characteristic
 

 
Existing social 
network
 
Contribution 
amounts
 

 
How 
the 
backer

ort i
n Kickstarter varies depending on project 
and timing
 
 
13
 
2.2.2 
Information Asymmetry and Collective Evaluation in Crowdfunding 
 
 
In traditional financial markets, while start
-
ups have several avenues for external funding 
such as 
venture 
capital companies, friend
s
 
and famil
ies
 
(F&F), and angel invest
ors
, 
in practice it 
is considerably
 
difficult
 
to 
obtain
 
external funding (Gompers 
and
 
Lerner, 2001; Hsu, 2007).
 
It has 

 
her
 
human capital
 
such as 
education level, history, and prior related experience (
Stuart
, Hoang, and Hybels
, 1999
; Busenitz 
et al., 2005), 
since
 
funders may interpret such observable 
information
 
as a signal of 
the 
potential 
quality of 
the 
projects 
(Stuart et 
al., 1999).
 
Entrepreneurs

 
embeddedness in social networks
 
also 
influence the 
probability of external fundraising
 
success in traditional funding markets
 
(Jenssen
 
and 
Koening, 2002; Shane and Cable, 2002;
 
Shane and Stuart, 2002). 
 
 
Crowdfunding entails sign
ificant differences from traditional funding methods such as 
microfinancing and angel capital investment in terms of the economies of scale
 
and
 
scope of 
accessible resources (Mollick, 2012). However, in traditional funding methods, the size of project 
crea

s
 
social networks remain hidden and the social interactions among funders are not 
observable. For example, 
in a 
ventu
re capital companies syndicate
 
we may observe the formation 
but not what details 
they exchange with each other.
 
 
Rewards differ across 
crowdfunding platforms and monetary rewards are not always 
necessary. At the same time, however, since funders can contribute a small sum of money to 
projects without any profession
al
 
expertise
 
or the
 
ability to evaluate either 
the 
quality of projects 
or a

 
Online P
eer
-
2
-
P
eer 
(P2P)
 
lending is characterized by similar types of asymmetric information, in the sense that 
individuals can gather small 
a
mounts of money from crowds. In an onl
ine funding context, Lin
 
et 
al. 
(2013) found that both soft information (online friendships) and hard information (credit 
14
 
 
information) were associated with the likelihood of credit being issued.
 
However, there are 
significant differences in terms of reward
s and processes of funding
4
. 
 
 
We suggest 

The Ladder Model of Co
-
Value Creation in Crowdfunding

 
to explain the 
benefits which 
entrepreneurs
 
can obtain by using 
crowdfunding
 
platforms (Figure 
2
.
2). Like co
-
creation of value in marketing literature which 
stresses different level of involvement or impacts 
at different points in the value chain, funders

 
roles, their communities, and co
-
creation processes 
are different in the crowdfunding model. First, entrepreneurs can obtain 
financial 
resources
. 
Crowdfundi
ng allows entrepreneurs to enhance legitimacy of their projects. Entrepreneurs can 
easily access potential investors and external financial resources, and interact with them. Some 
entrepreneurs
 
raise more funding than their initial goals. Additionally, the
 
success in crowdfunding 
projects creates legitimacy for additional funding. Furthermore, entrepreneurs can obtain non
-
financial or indirect potential benefits (i.e., social capital and potential customers, delivering initial 
products and achieving innovat
ion, new opportunity and on
-
going business), which will become a 
source of potential revenue. The online community interactions and e
-
WOM not only attract 
funders, but also help entrepreneurial innovation. Funders are able to participate actively in 
entrep
reneurs

 
innovation process
. 
T
hey evaluate and share ideas about the project and 
ent
re
preneurs

 
performance. Additionally, crowdfunding communities force entrepreneurs to 
complete their initial innovation plans
 
(delivery outcomes). Lastly, based on the exi
sting online 
social capital, additional funding, and 
credibility
, entrepreneurs
 
can 
initiate 
new innovation
. Figure 
2
.
2 summaries main benefits
,
 
which entrepreneurs can obtain from using crowdfunding. 
 
 
                                        
        
 
4
 
Crowdfunding calls for entrepreneurs to provide monetary or
 
non
-
monetary rewards to 
funders, which is different from paying interest on loans, as is done in P2P marketplaces
.
 
15
 
 
 
 
Figure 
2
.
2
 
The Ladder Model of Co
-
Value Creation in Crowdfunding
 
 
Thus, Crowdfunding platforms provide with an online social space which enables 

connection

, interactions among funders and entrepreneurs or among funders, 

collaboration

, 
more than simple involvement (i.e., share experience and 
knowledge
 
about funding projects), and 

co
-
value creation

, evaluating and supporting entrepreneurs
5
. 
Figure 
2
.
3
 
shows the three major 
roles of crowdfunding platforms: (1) Project creators launch t
heir projects; (2) funders choose 
projects to which they want to contribute; and (3) platforms provide both participants with a place 
for online interactions and functions. 
 
 
                                        
        
 
5
 
W
e reframed Prahalad and Ramaswamy
 
(2004)

s argument about co
-
creation
 
in a 
crowdfunding context
.
 
16
 
 
 
 
 
Figure 
2
.
3
. Two
-
sided Interaction 
in Crowdfunding Platforms
 
 
We consider 
three distinct aspects of crowdfunding platforms that translate to success in 
fundraising. First, funders in crowdfunding platforms are engaged in terms of providing ideas and 
input, and proactive in seeking details about the project from the entrepreneur,
 
which requires 
entrepreneurs to provide business plans and details about how an idea how would executed. Thus, 

as crowdsourced input for problem
-
solving and proj
ect execution. In other words, crowdfunding 
platforms exhibit characteristics of online communities in enabling interactions between funders 
and entrepreneurs. We consider the role of network embeddedness and community engagement in 
fundraising success. Se
cond,
 
crowdfunding platforms facilitate interactions among participants 
within
 
the 
platform
 
and 
increase the 
visibility
 
of 
the 
projects and 
the 
participants, entrepreneurs 
can easily interact with funders and convey 
their
 
personal credibility, information 
about their 
17
 
 
professional organizational a
bilities
, their achievement
s
, 
track record
,
 
and
 
information about key 
funders. Such 
entrepreneurial 
human capital 
enhance
s the visibility
 
and 
legitimacy
 
of a project, 
which 
raises the
 
probability of success 
in 
fundr
aising
 
(
Zimmerman and Zeitz, 2002; 
Zott and Huy, 
2007). Finally, online social media such as blogs, wikis, crowdsourcing, and online forums have 
become important intermediaries for information transmission and for forming online 
communities (Finin et al., 
2008). Many studies have examined how electronic word
-
of
-
mouth 
influences consumer choices, content
-
generation behaviors, and product sales (Dellarocas, 2003; 
Pollock and Rindova, 2003; Grewal et al., 2003; Goes and Lin, 2014). Aggarwal et al. (2012) show 
how the valence of electronic word
-
of
-

-
making. We 
therefore examine how e
-
WOM in online social media help
s
 
crowdfunding projects to succeed. 
To gain insights into the effects of social networks in crowdfunding,
 
we analyze the connections 
between the fundraising results of projects on crowdfunding websites and the corresponding 
promotion campaigns in online social media. 
 
Figure 
2
.
4 shows the research model of our study
 
and serves as a roadmap for hypotheses 
deve
lopment
.
 
Our paper considers two
-
sided interactions in crowdfunding as occurring through 
two distinct social channels: (1) how social interactions and community engagement in 
crowdfunding influences fundraising success, 
f
inally, external impact and how e
-
W
OM on social 
media influence fundraising success. Our study is different from the existing crowdfunding studies. 
First, 
we examine digitally enabled two
-
sided network effects such as the impact of funders

 
interaction (network 
embeddedness
, engagement) and
 
entrepreneurs

 
action
s
 
on fundraising 
success
. We also examine the impact of e
-
WOM on fundraising success. In addition, we examine 
the effects of traditionally 
critical factors (entrepreneurs

 
human capital
) in fundraising success.
 
 
18
 
 
 
 
Figure 
2
. 
4
. 
Research Model
 
 
2.2.
3 
Hypotheses Development
 
 

 
 
 

characteristics of relationship influence various outcomes of interest, including behavior and 
performance (Ibarra and Andrews 1993; Lin, 1999; Wasserman and Faust, 1994). Acc
ording to the 
theory, people can access more information and resources and influence others depending on the 
strength or ties and the structural position within the networks (Wasserman and Faust, 1994; Ibarra 
and Andrews 1993). Based on Social Information 
Processing (SIP) theory, Salancik and Pfeffer 

19
 
 
attitude development. In particular, SIP plays an important role in shaping attitudes under 
conditions of uncertain
ty or ambiguity, because people are more likely to follow social 
interpretations in those conditions (Ibarra and Andrew, 1993). Studies based on Social Network 
Theory elucidate the SIP mechanisms, and provide more adequate measures for exploring the 
relati

Aydin, 1991). Network centrality is a fundamental concept that shows the extent to which a node 
is connected to other nodes in a social network (Wasserman and Faust,
 
1994; Borgatti and Everett, 

attitudes (Dean and Brass, 1985; Ibarra and Andrew, 1993). Network centrality confers access to 
resources both through better acc
ess to information as well as through the ability of central actors 
to influence others that makes them more likely to engage in similar behaviors. A central network 
position of an entrepreneur represents greater information accessibility and resource avai
lability 
(Jenssen and Koening, 2002). For example, entrepreneurs who have higher centrality within the 
entrepreneur
-
funder network have greater access to resources and their behaviors often influence 

ence of a funder amongst other funders, 

(Aral and Walker, 2012). Most prior studies on the role of social networks in funding opportunities 
have focused on
 
friends and the small offline social network of entrepreneurs. However, our study 
takes a new perspective on the role of social networks enabled by digital platforms and how online 
social interactions create business values. We study the effect of the lar
ger cohort of online 
participants and their interactions on the platforms, where the participants are mostly strangers 
rather than offline friends and acquaintances.
 
20
 
 
Several economic studies have emphasized the role of network embeddedness on 
economic outc
omes (Burt, 1997; Uzzi, 1997; Granovetter, 2005). Granovetter (2005) claimed that 

influenced by these relations. Since any social interaction transmits information, t
he embeddedness 
of social interactions or transactions in a network promotes the formation of collaboration and 
improves performance (Granovetter, 1985). Network embeddedness facilitates more efficient 
spread of information about what members are doing and
 


embedded in their social relationships and if people have higher embeddedness they will have 
more chance to get informat
ion and resources or influence others. Recently the effects of network 
embeddedness have often been reviewed in a digitally mediated community context along with 
traditional social network and social influence studies. Grewal, Lilien and Mallapragada (2006
) 
examined the effects of the network embeddedness of projects and developers on project success 
in an open source system (OSS). They found significant heterogeneity exists in the network 
embeddedness of open source projects and managers, and that network 
embeddedness influences 

OSS success. Entrepreneurship literature likewise suggests that network embeddedness helps with 
fu
ndraising (Venkataraman, 1997).
 
In a
 
crowdfunding context, network embeddedness refers to the network centrality of the 

within a whole crowdfunding network. Those funders will have more chance to i
nteract with, to 
motivate, and to attract each other (highly centralized funders) and other funders through the social 
influence mechanisms, which ultimately helps projects reach their goals. The more connected 
21
 
 
people are to each other, the more they may b
e opinion leaders (Rosen, 2009). However, funders 
who have low centrality have less social interactions with others and have limited means to 
persuade others within a network, while they individually fund projects. Many studies on the 
process of social inf
luence argue that individuals tend to ignore their preferences and private 
information, and follow the crowd, who are presumed be better informed when making investment 
decisions under conditions of uncertainty (Pollock et al. 2003). Therefore, projects wh
ich have 

propose the following hypothesis.
 
 
H
ypothesis 
1
: The network embeddedness
 

 
funders will influence its
 
probability 
of fundraising success.
 
 
A 
digital platform enables people to transmit information, share knowledge and engage in 
discussions (Finin et al., 2008). Online feedback mechanisms foster cooperation among 
community members and the publicly shared knowledge influence the entire community 
(Dellarocas, 2003). People may participate in online communities because of informational and 

other p

Venkataraman (1997) argues t
hat investors can overcome the 
problem of information asymmetry using social relationships. In the crowdfunding platform, 
funders are able to freely share comments and knowledge about projects, and to ask and to answer 

post comments in the platforms, often share in
-
depth insights 
on projects, and the reasons behind the choice they made as funders. Funders can share more 
knowledge about the projects and entrepreneurs through interactions within the digital platforms. 
By p
osting those comments, they can engage in the crowdfunding community. The volume of 
22
 
 
comments on a project is a reflection of social engagement and enthusiasm of the community 
members. More comments and discussion are proxies of how much funders are engaged
 
in the 

fundraising success.
 
 
Hypothesis
 
2
: 
The 
community engagement 

probability 
of fundraising success.
 
 
In the marketing l
iteratures, there has been the rich discussion on the power of electronic 
word
-
of
-
mouth, in which many prior studies has shown the positive e
-
WOM effects on product 
sales and advertisement (Reichheld, 2003). Aggarwal et al. (2012) find the significant impa
ct of 
e
-
WOM from blogs on the probability to get venture capitals. However, there is no study on how 
e
-


 
context. Therefore, we would 
like to investigate the e
-
WOM effects from popular social networking sources and private blogs 

-
WOM during 
the fundraising period will increas
e the probability of fundraising success. Not only does e
-
WOM 
increase the visibility of projects and entrepreneurs, but it also substitutes for unavailable quality 

 
potential 
performance (Aldrich and Fiol
,
 
1994; Sanders and Boivie
,
 
2004). Individuals seek out e
-
WOM 
messages to get more information about products and services (Schindler and Bickart, 2005). Thus, 
the increase in volume of e
-
WOM may increase the probabi
lity of funding success, because it 
increases the awareness and familiarity about projects and entrepreneurs. We suggest following 
the hypotheses.
 
23
 
 
 
Hypothesis
 
3
:
 
The volume of overall e
-
WOM for a project will be positively related to 
its
 
probability of fundraising success.
 
 

 
 
 
Entrepreneurs

 
can purposely seek legitimacy through specific actions (Suchman 1995; 
Zimmer and Zeitz, 2002
; Zott and Huy, 2007). According to 
entrepreneurship literatures, 
legitimacy (acceptance of an entrepreneur

 
actions as proper or appropriate) helps new ventures to 
overcome the liability of newness and influences the 
acquisition
 
of necessary resources (Suchman, 
1995; Zott and Huy, 2007). Ve
ntures can increase their legitimacy through several actions such as 
sharing information about their performance (patents; prior performance), credentials, 
endorsement, and social connections (Lousbury and Glynn, 2001; Zott and Huy, 2007).
 
 
 
Crowdfunding f
acilitates entrepreneurs to increase their 
visibility
 
and the 
opportunities
 
to 
access necessary resources by enhancing legitimacy of a pr
oject. 
First, the crowdfunding platform 
allows each entrepreneur to communicate with funders 
simultaneously
 
and frequen
tly. 
Entrepreneurs can continuously provide project information, project progress, and their credibility 
information, and such actions will make funders draw interpretations about the character of the 
entrepreneur and projects. 
F
or example, sharing a photo
 
of products (or a prototype) in the platform 
can cause funders (viewers) to perceive the entrepreneur as having professional process. 
Those 
actions 
can 
make them more credible and knowledgeable 
and are
 
interpreted as a good signal about 
their performance 
(Zott and Huy, 2007
). 
A
lso, funders within the platform can respond to the 
entrepreneurs

 
updated information and entrepreneurs can also 
respond back to the comments.
 
The 
interaction can be extended to the interaction among participants. Such three
-
way interactions 
(between entrepreneurs and funders, funders and entrepreneurs, and among funders) will help to 
24
 
 
increase legitimacy of a project. Also, entrepreneurs can lev
erage their funders to diffuse 
information and project stories. Once entrepreneurs share any stories about projects within the 
platform, funders can choose to indicate their interest in that information and comment on and 
share it with others. 
Thus, 
entrep
reneurs

 
(
updating behavior
s)
 
will 
influence the 
probability of 
fundraising success. 
W
e propose the following hypothesis:
 
 
Hypothesis
 
4
: The volume of 
entrepreneur
ial engagement with the crowdfunding community 
positively 
influences 
the 
probabili
ty
 
of 

fund
rais
ing success.
 
 

 
 
 
Considerable information asymmetry exists in crowdfunding. Funders may not have 
prior
 
information
 
about entrepreneurs,
 
offline
-
relationships
,
 
or prior experience with entrepreneurs. 
F
unders can contribute small amounts of money to projects without any profession
al
 
expertise 
or 
the
 
ability to evaluate 
the 

hidden information problem, in that the quality of the inno
vation (Agarwal et al., 2013; Sanders 
and Boivie 2004), entrepreneurial vision and ability are all difficult for potential funders to infer. 
Where substantial information asymmetries exist, there is high possibility of adverse selection 
(Leland and Pyle, 1
977).
 
Therefore, in 
entrepreneurial
 
context potential funders deploy 
quality
 
signals to navigate the asymmetry of information 
between
 
what they know and they nee
d to know 
(Agarwal et al
.
, 2013;
 
Janney and Folta, 2003).
 
 
Spence (1973) defines signals as 


 
and demonstrated that education is a signal in 
the labor market
. In entrepreneur finance literature, investors assume 
that 
signals from 

ervable characteristics such as prior education level and history covariate
 
with
 
25
 
 
their actual performance (
Deeds, Decarolis and Coombs, 1997; 
Stuart et al., 1999).
 
Such 
observable characteristics are germane to the viability, competence and potential value
 
of start
-
ups 
(Busenitz et al
.
, 2005). Positive information signals will increase the probability of potential 
funders

 
contribution. Therefore, entrepreneurs should convey positive quality signals to outsiders 
to become viable (Prasad, Bruton and Vozikis,
 
2000; 
Agr
a
wal
 
et al. 2011).
 
I
n crowdfunding 
platform
s
 
the 

observable 
attributes will influence the fundraising 
success.
 
In this study, we 
consider
 
two attributes such as education 
level
 
and prior 
experience
. 
Higher education ca
n be a proxy for start
-
ups

 
innovation capability.  Also, entrepreneurs who 
have prior project experience will have more skills to manage projects. W
e propose the following 
hypotheses:
 
 
 
Hypothesis
 
5

influence 
the 
probabili
ty 
of 

fund
rais
ing success.
 
 
Hypothesis
 
6

 
influence
s
 
the 
probability 
of 

fund
rais
ing success.
 
 
26
 
 
2.3 
DATA AND METHODS
 
 
2.3.1 
Data Collection
 
 
 
The data have been collected from Kickstarter.com. To 
participate in the platform, 
individuals or entrepreneurs need to sign up with the site and submit their basic personal 

profiles and social media information. 
Any individual or entrepreneur can create
6
 
as well as fund 
projects in a variety of areas. Project creators pay fees only when their goals have been reached, 
in the same way project funders can only fund a project when the pre
-
announced goal has been 
reach

profile information, a project video, and detailed project information. Members can also observe 

ccumulated amounts 
pledged, the number of comments and updates. 
W
e collected projects, funders, and entrepreneur 
data. Our sample focuses on the projects initiated within the Technology category. There are a total 
of 7
2
2 projects (March 2012 

 
January 2013
) in the dataset. We have a
 
total of 3
03
 
successful 
projects and 4
19
 
failed projects, which is almost 
the 
same 
as
 
Kickstarter.com statistics
7
. F
igure 
2
.
5 
shows 
the
 
screenshots of a project and a funder

s profile pages.
 
 
 
                                        
        
 
6
 
Currently, Kickstarter.com only allows
 
people in the United States or United Kingdom to 
create projects.
 
7
 
30 additional projects are removed from our dataset, because of insufficient information.
 
27
 
 
 
Figure 
2
.
5
 

 
 
 
We gathered project information (goal, duration, category, description, number of funders, 

from each page. We 


detailed attributes from LinkedIn. During our observation window, we collect e
-
WOM data for 
each project from online social networking sites (Twitter, Facebook and online websites) using a 
web
-
crawler. We also relied on Google searches, archive searches and data from the Kickback 
machine
8
, a site that archives Kickstarter webpages.
 
 
 
                                        
        
 
8
 
It is closed now
 
(2014).
 
28
 
 
2.3.2 
Network Analysis
 
 
The projects on Kickstarter.com are classified into more than 1
3
 
categories. We consider 
each category to be a network boundary. Our sample focuses on the projects initiated within the 
Technology category (e.g. technology, open source, and hardware). There are a total of 
722
 
projects and more than 
177
,000 funders in th
e dataset.
 
Following Wasserman and Faust (1994), 
we created 
two
-
mode 
affiliation networks of projects and funders
 
(Figure 
2
.
6)
. 
We measure 
network 
embeddedness
 
using network centrality measures (Everett and Borgatti, 2005).  To 
measure network embeddedness
 
of funders, we follow three steps. First, we formed a two
-
mode 
network by connecting funders through projects. After that we generated the one mode network of 
funders and calculate their centrality score within the whole platform. Lastly, based on those s
core, 
we calculated aggregated network 
embeddedness
 
for each project. 
 
 
 
Figure 2
.
6
 
Transpose Two
-
Mode Network to Two One
-
Mode
 
 
 
 
Two
-
mode affiliation 

 
circle: projects 
& square: funders
 
One
-
mode project 
 
One
-
mode funders 
 
 
29
 
 
2.4 
EMPIRICAL MODEL
 
 
2.4.1 
Funding Success Model
 
 
Crowdfunding success
 
is
 

entrepreneurial actions, and project characteristics. Fundraising success is (i) a
 
binary variable, i.e., 
whether a p
roject reaches its 
fundraising 
goal or not
,
 
and (ii) percentage of
 
pledged
 
amounts.
 
 
 













































































 
 
In this model, we examine the impact of network embeddedness, the volume of social 
interaction among funders, e
-
WOM
, 
entrepreneurs

 
engagement, and 
entrepreneur
 
human capital. 
We controlled for project characteristics such as size of goal and duration
, and 
control the 
geographical impacts, seasonal impact,
 
and 
reward type
s
.
 
Since the observed outcomes of CFS is 
binary 
(1=success, 
otherwise 
0
), we assu
me 


~ N (0,1) and use the general probit model.
 
 
 
Pr (




) = 

 
(


)
 
 
W
here 


 
is 1 when 
i
 
project get 
successfully
 
funded, otherwise it is zero.The positive and 
significant 







will suggest that the projects are more 
likely
 
to 
be 
success
ful
 
if higher 
network embeddedness, or
 
community interactions exist. 
The positive and significant 


 
and 


 
will suggest that the projects are more 
likely
 
to success in 
fundraising
 
if higher entrepreneurs

 
30
 
 
characteristics (education and prior 
experience
) e
xi
st. 
The impact of entrepreneurs

 
engagement
 
on the probability of funding success 
is
 
measure by 


 
and 
E
-
WOM 
is
 
measured by 


, 
respectively.
 
In the same vein, we analyze
 
the
 
percentage amount
s
 
model
 
using an OLS model.
 
 





































































 
 
 
 
31
 
 
2.5 
RESULTS AND DISCUSSION
 
 
2.5.1 
Res
ults
 
Table 2
.2
 
shows the definition of variables and descriptive statistics
. The maximum value 
of 
p
ledged percentage is 6264, the goal is 750,000, and the maximum total pledged amounts are 
2,945,885. The standard deviation of pledged percentage is 440.
89
 
a
nd mean is 170.
77
. 
 
We use 
two success variables such as Success (fundraising success) and percentage of pledged amounts 
(= goal/total funded amounts). Each project has initial fundraising goal and duration. The average 
duration is 27 days. During fundrais
ing period, we collect comments about each project in online 
social networking sites (e.g. twitter, Facebook) and calculate the total volume of electronic word 
of mouth. The maximum volume is 984, and we used square of total volume (range between 0 and 
31.
37, mean = 2.09, and SD=3.77) in the analysis.
 
There are three education levels: BS, MS, and 
PhD.
 

information from both Kickstarter.com and LinkedIn site.
 
We looked at whethe
r each project 
creator has prior crowdfunding experience (whether they created other project) or not. However, 

calculated net positive prior experience (Total po
sitive experience 

 
negative experience), which 
ranges from 
-

crowdfunding platforms. Also, based on our network analysis, we calculate the network 
embeddedness of each project. F
or the empirical analysis, we use log (Network embeddedness). 
We also control reward type of each project if the final outcome is physical products (either 
hardware or software) or warm glow type (gift, invitation, etc.). 
 
 
 
32
 
 
Table 2
. 
2 
Variable Definitions
 
and Descriptive Statistics
 
Variables
 
Definition
 
No. 
Obs.
 
Mean
 
Std. 
Dev.
 
Min
 
Max
 
Success
 
=1 if
 
a project reach the goal, 
otherwise =0
 
722
 
0.42
 
0.50
 
0
 
1
 
%_pledged
 
Percentage of funded amounts
 
722
 
170.77
 
440.89
 
0
 
6264
 
Goal
 
Funding goal (amounts)
 
722
 
39715.
7
 
70100
 
75
 
750000
 
Amounts
 
Total funded amounts
 
722
 
41152.3
4
 
175034
 
1
 
294588
5
 
Duration
 
Initial Length of the funding 
cycle for each projects (days)
 
722
 
26.89
 
13.77
 
1
 
60.04
 
E
-
WOM
 
Total quantity of comments in 
social network sites
 
722
 
18.54
 
70.50
 
0
 
984
 
Education
 

(1=BS,2=MS, 3=PhD, 
otherwise 0)
 
722
 
0
.67
 
.84
 
0
 
3
 
Experience
 

experience (Positive 

 
Negative)
 
722
 
-
0.04
 
0.38
 
-
2
 
2
 
Entrepreneur 
Engagement
 
The volume of actions (# of 
updates for a 
project)
 
722
 
4.37
 
5.11
 
0
 
33
 
Network 
embeddedness(lo
g)
 
Embeddedness of a project 
within a community
 
721
 
5.59
 
1.81
 
0
 
9.97
 
Funders 
Engagement
 
Total quantity of discussion for 
each project
 
722
 
52.98
 
168.54
 
0
 
1660
 
Num. of FB 
Friend
 
Total number of project 

FB sites = 0
 
722
 
268.58
 
471.59
 
0
 
5088
 
Word Count
 
Total number of words in 

 
722
 
82.39
 
94.92
 
0
 
1603
 
Social Presence
 
=1 if each entrepreneur share 
her personal social network site 
links, otherwise =0
 
722
 
0.68
 
0.47
 
0
 
1
 
Reward Type
 
=1 physical outcome, otherwise 
=2
 
722
 
1.20
 
0.40
 
1
 
2
 
Image
 
=1 if each entrepreneur share 
her personal image, otherwise 
=0
 
722
 
0.66
 
0.48
 
0
 
1
 
 
Figure 2.7 shows the distribution of percentage of total pledged amounts and project goals. 
Th
e distribution figures 
show the evidence of prevailing 

long tail

 
effect in crowdfunding, which 
33
 
 
provides diverse project choices
 
to funders
 
and 
financing 
opportu
nity
 
for
 
project creators. F
or the 
empirical analysis we do a log transformation of the goal and percentage of pledged percentage 
amounts. Table 2.2 provides more detailed information.
 
 
 
Figure 
2
. 
7. Distribution :Pledged Percentage and Project Goal
 
 
 
 
34
 
 
Table 2.3
 
represents the correlations 
of 
variables.
 
Table 
2
.3
 
The 
Correlation
s
 
of Variables
 
 
SG
 
PA
 
Goal
 
RW
 
DR
 
ED
 
EP
 
MO
 
EA
 
SEG
 
NE
 
LC
 
EW
 
NF
 
SG
 
1.000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
PA
 
0.784
 
1.000
 
 
 
 
 
 
 
 
 
 
 
 
 
Goal
 
-
0.294
 
-
0.365
 
1.000
 
 
 
 
 
 
 
 
 
 
 
 
RW
 
-
0.089
 
-
0.128
 
-
0.157
 
1.000
 
 
 
 
 
 
 
 
 
 
 
DR
 
0.284
 
0.197
 
0.101
 
0.010
 
1.000
 
 
 
 
 
 
 
 
 
 
ED
 
0.001
 
-
0.0083
 
0.097
 
-
0.031
 
0.106
 
1.000
 
 
 
 
 
 
 
 
 
EP
 
0.065
 
0.020
 
0.100
 
-
0.007
 
0.106
 
-
0.001
 
1.000
 
 
 
 
 
 
 
 
MO
 
0.017
 
0.032
 
0.006
 
0.032
 
-
0.447
 
-
0.125
 
-
0.025
 
1.000
 
 
 
 
 
 
 
EA
 
0.412
 
0.426
 
0.173
 
-
0.188
 
0.242
 
0.024
 
0.033
 
-
0.030
 
1.000
 
 
 
 
 
 
SEG
 
0.338
 
0.363
 
0.216
 
-
0.136
 
0.172
 
0.043
 
0.014
 
-
0.013
 
0.385
 
1.000
 
 
 
 
 
NE
 
0.611
 
0.714
 
0.126
 
-
0.285
 
0.228
 
0.059
 
0062
 
-
0.044
 
0.526
 
0.500
 
1.000
 
 
 
 
LC
 
0.069
 
0.062
 
0.030
 
-
0.118
 
-
0.041
 
-
0.013
 
0.020
 
0.095
 
0.075
 
0.061
 
0.081
 
1.000
 
 
 
EW
 
0.271
 
0.267
 
0.188
 
-
0.044
 
0.115
 
0.129
 
0.087
 
0.044
 
0.258
 
0.337
 
0.365
 
0.031
 
1.000
 
 
NF
 
0.013
 
0.010
 
0.004
 
0.065
 
0.020
 
0.073
 
0.025
 
-
0.061
 
-
0.029
 
-
0.015
 
-
0.038
 
0.010
 
0.052
 
1.000
 
SG: Goal Achieve; PA: log pledged amounts; Goal: log goal; RW: reward type; DR: duration;
 
ED: education; EP: experience; MO: Month; EA: entrepreneurial action;
 
SEG: social engagement; NE: network embeddedness; LC: location; EW: EWOM; NF: 
number of Facebook friend
 
 
35
 
 
We 
examine
 
two models
: 
(1) Fundraising success model and (2) Percentage Amount of 
Funding. We 
examine
 
the first model using probit model and OLS for the second model. To deal 
with 
het
erogeneity
 
of projects, we use cluster robustness errors. Table 
2.4
 
shows the empirical 
results
 
of our 
analysis
. 
The R
2
 
is 0.68 in fundraising success model (pseudo) and 0.760 in log 
percentage of pledged amounts model. 
Network embeddedness positively infl
uences 
both 
fundraising success (0.804***) and log percentage of pledged amounts (0.851***)
, which support
s
 
Hypothesis 
1.
 
Projects which have highly embedded funders will have more chance to get funding, 
because f
unders who have higher embeddedness influen
ce others

 
participation and increase the 
probability of 
fundraising
 
success.
 
Funders

 
engagement
 
within the platform
 
also positively 
influences crowdfunding success
 
(0.014***, 0.001*** respectively; support Hypothesis 2)
. When 
people share more 
information
 
about the project and actively interact with others within the 
platform
, the project can attract more 
funders
.
 
We also observe the significant impact of e
-
WOM 
on fundraising success (0.079***, 0.33** resp
ectively; support Hypothesis 3)
. 
We obse
rve the 
significant
 
impacts of entrepreneurs

 
engagement 
on 
fundraising
 
success
 
(0.056***, 0.051*** 
respectively; support Hypothesis 4)
. We operationalize entrepreneurs

 
engagement
 
as the 
volume
 
of updates on crowdfunding platforms. The entrepreneurs can c
onvey more knowledge than a 
simple project description through updating actions. These actions will signal the quality of 
projects and potential innovation capability. Also the effort, time, and dedication make 
entrepreneurs appear to be trustworthy.
 
Howev
er, we cannot observe the significant influence of 
entrepreneurs

 
human capital
 
such as 
prior experience and 
education on 
fundraising
 
success
 
in our 
initial test
. 
We then measure the volume of net positive prior experience (Successful prior 
experience minu
s failed experience)
 
rather than using total prior experience. We find that projects 
are more likely to get funding if project creators have prior positive success experience (only this 
36
 
 
model supports Hypothesis 6), but we cannot observe the significant im
pact of positive experience 
on the percentage of pledged amounts. 
Traditional funding literature emphasized the role of trust 
and reputation in fundraising success. Thus, start
-
ups who do have prior 
positive 
history, 
relationship, or reputation 
will be mor
e likely to get funding from funders, but the prior positive 
experience does not significantly influence total funding amounts. Additionally, none of education 
level has significant influence on either fundraising success or percentage of pledged amounts (
do 
not support Hypothesis 5). Overall, compared
 
to traditional funding methods, 
funders
 
in 
crowdfunding
 
may 
value
 
social information more than the 

 
quality.
 
 
 
 
37
 
 
Table 
2
.4
 
Results
 
Independent Variable
 
Fundraising Success
 
OLS on LogPert_Amounts
 
Coefficients
 
Std. Err
 
Coefficien
ts
 
Std. Err
 
LogGoal
 
-
 
0.95***
 
0.082
 
-
0.77***
 
0.030
 
D
uration
 
0.046***
 
0.007
 
0.018***
 
0.004
 
P
rior Experience
 
0.476*
 
0.243
 
0.066
 
0.118
 
Education1
 
-
0.196
 
0.191
 
-
0.069
 
0.101
 
Education2
 
0.011
 
0.260
 
0.109
 
0.143
 
Education3
 
-
0.323
 
0.510
 
-
0.316
 
0.226
 
Entrepreneur

 
Engagement
 
0.056***
 
0.021
 
0.051***
 
0.010
 

 
0.014***
 
0.004
 
0.001***
 
0.0003
 
Network Embeddedness
 
0.804***
 
0.094
 
0.851***
 
0.033
 
E
lectron
ic
 
Word of Mouth
 
0.079***
 
0.023
 
0.033**
 
0.013
 
Num. of FB 
Friends
 
0.0002*
 
0.0001
 
0.0002**
 
0.0001
 
M
onth
 
0.126***
 
0.030
 
0.087***
 
0.033
 
Location
 
-
0.000
 
0.0004
 
0.0000
 
0.0002
 
R
eward Type
 
0.096
 
0.211
 
0.071
 
0.117
 
Log
-
likelihood
 
-
154.80
 
-
 
R
 
0.684 (Pseudo)
 
0.760 (Adjust R
2
: 0.755)
 
Number of obs
ervation
 
721
 
721
 
two
-
tailed significance levels p<0.1, *; p<0.05, ** ; p<0.01, ***
 
 
 
We also looked that distribution of technology projects. Traditionally, technology projects 
have been often initiated in the west
-
coast (Agarwal et al.
,
 
2011), but in crowdfunding context we 
can see that projects are dispersed across the U.S. (Figure 
2
.
8). 
T
he empirical results also show the 
38
 
 
non
-
significant effect of location on 
fundraising
 
success.
 
Thus, crowdfunding platforms bring 
together large number
s of geographically dispersed funders and allow them to participate in 
interactions in online communities, contribute to start
-
ups, and create economic and social values. 
 
 
 
Figure
 
2
.
8
 
Geographical Distribution (Technology Projects)
 
 
 
 
 
39
 
 
2.5.2 
Robustness 
Checks
 
 
Endogeneity
:
 
I
n
s
trument 
V
ariable 
 
 
T
o
 
address the robustness of our estimation model
, 
results, and model identification, we 
did implement a series of robustness checks. First, we check
ed
 
potential endogeneity issues. A 
critical issue in the estimation process is reverse causality between entrepreneurial actions (i.e., 
updating information about projects) and fundraising success. Not only the increase of updating 
actions influence fundrais
ing, but also entrepreneurs who get higher funding amounts may engage 
in updating more often. To deal with this issue, we used three sets of instruments, which are 
correlated with entrepreneurial actions but are not correlated with fundraising performance.
 
 
We collect
ed
 

entrepreneurs share their personal image and personal social network sites. Additionally, using 
profile description we calculate
d
 

asic text
-
mining 
techniques (1.sentence extraction, 2. tokenization, 3. stemming, and 4. removing stop words). 
P
rior 
studies show the relationship between entrepreneurial action and their narcissistic characteristics. 
E
ntrepreneurs are more likely to be 
optimistic and have narcissistic characteristics (Busenitz
 
and 
Barney,
 
1997)
. 
Narcissistic entrepreneurs 
tend to be overconfident on their decision
s
 
and have the 
tendency to 
overestimate their 
ideas, future 
performance
, and their own abilities 
(
Benabou and
 
Tirole, 200
2
; De Meza and 
Southey
, 1996
; 
Lee, Hwang, and Chen, 2014).
 
This tendency 
contributes to entrepreneurial activities (Hirshleifer, Low, and Teoh, 2010).
 
 
 
40
 
 
Table 
2
.5
 
Instrumental Variable Models
 
Variables
 
Probit Model: 
Fundraising 
Success
 
OLS on 
LogPert_Amounts
 
Coefficients
 
Std. Err
 
Coefficients
 
Std. Err
 
Fundraising 
Success
 
LogGoal
 
-
0.94***
 
(0.082)
 
-
0.787***
 
(0.047)
 
Duration
 
0.046***
 
(0.007
 
0.014
 
(0.009)
 
 
Prior Experience
 
0.473*
 
(0.243)
 
0.083
 
(0.126)
 
 
Education
1
 
-
0.193
 
(0.189)
 
-
0.039
 
(0.124)
 
 
Education
2
 
0.010
 
(0.255)
 
0.107
 
(0.146)
 
 
Education
3
 
-
0.297
 
(0.508)
 
-
0.265
 
(0.258)
 
 
Entrepreneur 
Action
 
0.043**
 
(0.022)
 
0.119
 
(0.154)
 
 

Engagement
 
0.014***
 
(0.004)
 
0.001
 
(0.001)
 
 
Network 
Embeddedness
 
0.770***
 
(0.095)
 
0.776***
 
(0.174)
 
 
Electronic W
O
M
 
0.077***
 
(0.023)
 
0.030*
 
(0.016)
 
 
Num.of FB Friends
 
0.0002
 
(0.0002)
 
0.0002**
 
(0.0001)
 
 
Month
 
0.130***
 
(0.030)
 
0.082***
 
(0.019)
 
 
Location
 
-
0.000
 
(0.0004)
 
-
 
0.000
 
(0.0002)
 
 
Reward Type
 
0.153
 
(0.210)
 
0.107
 
(0.144)
 
Entrepreneur 
Action
 
 
Word 
count(profile)
 
-
 
0.001*
 
(0.0005)
 
-
 
Personal Image
 
-
 
0.310***
 
(0.115)
 
-
 
 
Social Presence
 
0.
412*
 
(0.234)
 
-
 
rho
 
0.265
 
-
 
Likelihood Ration test rho=0:
 
Chi
2
=3.522
 
Prob> Chi
2
 
=0.061
 
-
 
Log
-
likelihood
 
-
527.76
 
 
R
2
 
 
0.746
 
Wald Chi
2
 
(Prob >chi
2
)
 
180.33 
(0.000)
 
2132.43 (0.000)
 
Number of obs
ervation
 
721
 
721
 
two
-
tailed significance levels p<0.1, *; p<0.05, ** ; p<0.01, ***
 
41
 
 
 
Entrepreneurs who have narcissistic characteristics enjoy explaining their vision and ideas 
because of own 
overconfidence
 
and the desire to get others

 
agreements
 
about their ideas, which 
can be represented in the entrepreneurs

 
profile in crowdfunding project pages.
 
Those 
entrepreneurs
 
tend to provide longer profile and to share own personal images. Additionally, 
entreprene
urs who share their personal social networking website may want share more 
information about projects information in order to show how much she
 
(
or 
he) is trustworthy. 
 
According to
 
accounting and finance literature, executives are becoming active on socia
l 
media channels (Blankespoor, Miller and White, 2013). Their personal social networking sites 
(e.g., Facebook, Twitter, or LinkedIn) can provide both indirect and direct information about a 
project, specific news about a project, funding status, 
and entre

 
day
-
to
-
day activities. 
Thus, p

, conversation and status
 
and get 
better informed about project creators and projects.
 
Prior literature suggests that executives

 
activation of personal social
 
network sites are related to the increasing the levels of 
disclosure
 
and 
alleviating information 
asymmetry (Saxton and Anker, 2013; Blankespoor et al., 2013). Therefore, 
entrepreneurs who activate their online social networking sites and share them in the
 
profile page 
are more willing to provide information about projects and interact with funders.  
 
We performed 
the seemingly unrelated 
b
ivariate probit regression for the goal achievement model and used the 
two
-
stage least square method for the percentage 
of pledged amounts model. Table 
2
.5
 
shows the 
significant influence of all social interactions variables in probit model with instruments variable 
of goal achievement. However, in the percentage of pledged amounts model, entrepreneurial 
actions do not sign
ificantly influence the percentage of pledged amounts. This information can 

mechanism, project creators try to convey more information about their projects and i
nteract with 
42
 
 
funders. However, once a project reaches the initial goal, entrepreneurs may not pay too much 
attention or effort to raise additional funding. Thus, entrepreneurial action may not be a critical 
factor for additional fundraising. At the same ti
me, funders who fund projects after projects have 
less risk (already reached the goal), may care about social information (e.g., who fund the projects, 
how others think of, and popularity of projects) more than information from entrepreneurs. Thus, 
in the 
percentage of pledged amounts model social information plays an important role.
 
This also 
explain
s
 

2
.
9
 
depicts an 
apparent power
-
law distribution, which the 20% of projects accounts for nearly 80% of cumulative 

current popularity. This is 
a 
commonly observed ph
enomenon, rich
-
get
-
richer or preferential 
attachment in networks (Barabasi and Albert, 1999; Easley and Kleinberg, 2010). Not every 
project get
s
 
the same amount of attention in crowdfunding platforms, and a few projects might 
achieve higher visibility. Fun

high popularity among members in online communities, and show some positive outcomes. 
 
 
 
Figure 
2
.
9
 
% of Cumulative Amounts and Number of Projects
 
 
0.00%
20.00%
40.00%
60.00%
80.00%
100.00%
120.00%
0
500000
1000000
1500000
2000000
2500000
3000000
3500000
1
12
23
34
45
56
67
78
89
100
111
122
133
144
155
166
177
188
199
210
221
232
243
254
265
276
287
298
309
pledged_amount
cumulative percentage
80%marker
43
 
 
Selection Bias 
-
 
Heckman Selection 
Model
 
 
We are suspicious of whether selection bias exists in 
the OLS model, 
because
 
the 
unobserved characteristics of successfully funded projects may create estimation bias in the 
model 
of pledged percentage
 
amounts. In order to deal with this selection 
biases
 
problem, we
 
perform 

 
selection model (Heckman, 1979)
. First, we estimate 
a
 
p
robit model
 
(dependent 
variable
: fundraising success =1, otherwise 0)
 
to obtain estimates of 

 
and compute 

 

If projects rea
ch the original funding goals, 




 
otherwise 




.  
 
 
First Stage Model: Pr (




) = 

 
(




), 










 
; 
 
 
Second Stage Model: 










 
 
 
Next, we estimate the expected value of 
Y
, conditional on 
S
=1, and 


: 
 
 
E(










) =




 
+ E(










E(












, 














)
 
 
 
If the residuals are correlated 
with one another and
 
we estimate OLS Model without considering 
the first model, we will get the 
biased
 
estimation. To test for bias, we examine the relationship 
between the residuals for the two stages (stage1 and stage2). 
 
 
E(












) =




 
+




, 
 
E(









  




 
, 

 
: 
the inverse Mills ratio
 

 
Second Stage Model: 















, E(






)  =




  
(

 
E(








 
 
 
If the 
unobservable
 
in the selection model are correlated with 
the
 
unobservable in the stage 
2 model, in 
other
 
words t

o
 
in OLS (outcome equation) 
model is significant
, we have biased estimates without correction. This is basically saying that the 
44
 
 
unobservable
 
in the selection of the probit model (Fundraising success) are also affecting the stage 
2 model. If the 
unobservabl
e
 
in stage 1 are unrelated to the 
unobservable
 
in stage 2, then we can 
say that the stage 1 does not affect stage 2 results. This is another way to say that selection into the 
sample of stage 2 is a random process, unaffected by different 
unobservable
. Tab
le 
2
.6
 
shows the 
results of Heckman

s Two Stage Selection Estimation. There are 721 projects are used for the first 
stage probit model and 303 projects are used for the second stage OLS Model (Log_Pert). The 
model estimation is 
statistically
 
significant
 
(
W
ald 


 
= 235.50***). We add the highest education 
level (PhD) to 
the 
probit model as an 
exogenous
 
covariate, but do not add to the second stage OLS 
Model. Our results show that the 
significant
 
coefficient of 

 
and a positive rho value in the OLS 
model, whi
ch indicates that the second stage model has selection bias. Thus, we want to explain 
the results based on Heckman
 
Selection model.
 
The results show the significant impact of Funders

 
engagement and network embeddedness along with LogGoal
 
(negative), but w
e cannot observe 
effects of other independent 
variables
. Compared to the goal achievement model, pledged 
percentage amounts mainly depend on the social interaction within the platform. Both positive
 
prior
 
experience
 
and entrepreneurial actions do not signi
ficantly influence percentage of pledged 
amounts.
 
 
 
 
45
 
 
Table 2
.6
 

 
St
age
 
Selection Model
 
Indep
en
dent
 
Variables
 
Probit Success
 
Model
 
OLS LogPert 
(Success=1)
 
Heckman Selection 
Model
 
 
Coefficients
 
Std. err
 
C
oefficients
 
Std. err
 
C
oefficients
 
Std. 
err
 
LogGoal
 
-
 
0.95***
 
0.082
 
-
0.
3223
***
 
0.030
 
-
0.430
***
 
0.050
 
D
uration
 
0.046***
 
0.007
 
-
0.005
 
0.004
 
0.001
 
0.004
 
P
rior Experience
 
0.476*
 
0.243
 
-
0.061
 
0.079
 
-
0.016
 
0.080
 
Education
1
 
-
0.196
 
0.191
 
-
0.003
 
0.081
 
 
-
0.049
 
0.082
 
Education
2
 
0.011
 
0.260
 
0.111
 
0.106
 
 
0.104
 
0.106
 
Education
3
 
-
0.323
 
0.510
 
-
 
-
 
-
 
-
 
Entrepreneur 
Engagement
 
0.056***
 
0.021
 
0.001
 
0.007
 
0.001
 
0.007
 

 
0.014***
 
0.004
 
0.
001
***
 
0.0002
 
0.001***
 
0.0002
 
Network 
Embeddedness
 
0.804***
 
0.094
 
0.395***
 
0.035
 
0.531***
 
0.050
 
E
-
Word of
 
Mouth
 
0.079***
 
0.023
 
0.00
3
 
0.003
 
0.008
 
0.009
 
Num.of FB Friends
 
0.0002*
 
0.0001
 
0.000
 
0.000
 
0.000
 
0.000
 
M
onth
 
0.126***
 
0.030
 
0.00
6
 
0.012
 
0.159
 
0.013
 
Location
 
-
0.000
 
0.0004
 
-
0.000
 
0.000
 
-
0.000
 
0.0001
 
Reward Type
 
0.096
 
0.211
 
-
0.101
 
0.103
 
-
0.053
 
0.102
 

 
-
 
-
 
0.491***
 
0.124
 
rho
 
-
 
-
 
0.79
 
Log
-
likelihood
 
-
154.80
 
-
 
-
 
R
2
 
0.55 (Pseudo)
 
0.529(Adjust 
R
2
: 
0.507)
 
0.55
 
Number of obs
 
72
1
 
303
 
303
 
two
-
tailed significance levels p<0.1, *; p<0.05, ** ; p<0.01, ***
 
 
 
46
 
 
Sentiment Analysis of Electronic Word of Mouth 
 
 
To 
test the effect
s
 
of e
-
WOM, we used total volume of online comments on an each project.
 
Prior studies on e
-
WOM show
 
contradict
ory
 
results (Aggarwal et al., 2012). Thus, we expect that 
the sentiment of e
-
WOM may differently influence fundraising success. To analyze e
-
WOM and 
sentiment of comments
,
 
we used 
web crawling
 
tools we collect comments 
that
 
mention each 
crowdfunding project. 
Considering relevance, we used data only talked about each project during 
fundraising period. 
To identify sentiment of each comment, we followed Lexicon based sentiment 
analysis approach. The sentiment (
negative, neutral, positive), 
were extracted from tra
nslated 
messages through the sentiment analysis based on Lexicon (Taboada et al., 2011; Pang and Lee, 
2008; Melville et al
.
, 2009). 
To calculate the valence of e
-
WOM, we use Janis
-
Fadner coefficient 
of imbalance, which is the most popular valence measureme
nt in communication research (Pollock, 
Shier & Slattery, 1995; Pollock & Whitney, 1997). 
The Janis
-
Fadner Coefficient of Imbalance
 
(single
-
score content analysis)
 
is calculated as follows: 
 
If f>
 
u 
(or the sum of the 

favorable

 
attention scores is greater
 
than the sum of the 

unfavorable

 
attention scores), 
Coefficient of Imbalance (answers lie between 0 and +1): 

















 


 
use the following formula: Coefficient of Unfavorable 
Imbalance (answers lie between 0 and 

1): 


















 
 
where
,  
f
 
= 
sum of attention scores coded positiv
e or favorable; 
u
=
 
sum of attention scored 
coded unfavorable
 
or negative; 
n=sum of 
attention scores coded balanced
 
or 
neutral
; 
r
 
=
 
f+
 
u+
 
n
.
 
The statistics yielded from these formulas can vary from +1.00 to 

1.00 and permit 
quantitative comparison of each e
-
WOM coverage of crowdfunding projects. Scores ranging 
47
 
 
between zero and +1 indicate
 
favorability while scores between zero and 

1 show unfavorability 
toward crowdfunding projects. However, our results do not show the significant impacts of the 
valence of e
-
WOM and significantly different sentiment effect on fundraising success. There are
 
two possible interpretations. First, most of e
-
WOM
s
 
(during fundraising period
s
) have positive 
sentiment
s
, because funders cannot experience and see the actual outcomes of projects and 
evaluate their experience. Funders often write comments to share proje
ct information and attract 
additional funders in order to successfully fund the projects to which they want to contribute. 
Because of that, the total volume of e
-
W
OM does significantly influence, but the valence does not 
significantly influence fundraising
 
success. 
 
 
Heterogeneity of Projects
 
 
 
Figure 
2
.
10
 
shows the percentage
 
of projects and backers
 
by reward types in our dataset.
 
Funders
 
can receive a tangible outcome, 
such as 
the final outcome of a project, or simple gifts 
(e.g., thank
-
you message, small gifts). 48.0% of crowdfunding projects in Technology category 
need to deliver hardware and 20% of projects provide funders with simple gifts or 
invitation
. 
Different types of r

ome 
funders may contribute to a project with non
-
economic motivation but others contribute to a 
project to pre
-
order the product. To explain people

s charity giving behaviors
,
 
Andreoni et 
al. 
(1990) introduce the warm glow theory, an internal satisfaction that comes from 
the
 
act of 
giving
. 
Unlike the perfect 
altruism
, under this theory, funders view others

 
contribution as imperfect 
substitutes of own contribution.
 
To deal with the unobserv
ed heterogeneity of projects result from 
expected rewards, we employ 

-
test for equality of 
means. 
To test this, we first aggregate current categories into two groups based on if outcomes are 
48
 
 
physical products (h
ardware, software, both) or warm glow type (gift, invitation, other product, 
nothing).
 

actions in two groups. We observed the significant mean difference in the size of
 
project (goal, 
percentage pledged amounts, network embeddedness, and entrepreneurial actions). However, we 
cannot find difference in other variable such as e
-

deliver physical products (hardware, software) usuall
y have larger goal than warm
-
glow type of 
project
s
. Therefore, those projects should attract more funders and money (regardless the size of 
project goal), and entrepreneurs also need to actively engage in projects to obtain sufficient 
funding to manufactur

activity (engagement, eWOM) does not vary across projects and reward types. 
 
 
 
Figure 
2
.
10
 
% of 
Project
s
 
and Backers by Reward Types
 
 
 
 
12.0%
48.0%
4.8%
0.4%
3.0%
10.6%
21.1%
% Projects by Reward Type
Gift
Hardware
Invitation
Nothing
4.15%
29.16%
39.49%
3.00%
3.92%
3.28%
2.35%
14.11%
0.54%
# Backers by Reward type
All
Gift
Hardware
Invitation
Nothing
49
 
 
2.6 
CONCLUSION AND IMPLICATIONS
 
 
 
In digitally mediated networks, most funders are 
anonymous;
 
often
 
they
 
are not 
part of
 
the same social network as the entrepreneurs. Rather than 
being influenced by 
traditional social 
norm
s
, funders 
are likel
y 
to 
be influenced by social interaction
s 
and 
peer influence within the 
community
 
that forms around the project
. Despite the important role of network characteristics in 
social influence of 
an online 
funding community on funding success, 
these topics have not received 
much
 
prior attention. However, th
ere 
is a great deal of literature that
 
emphasize
s
 
the role of social 
influence in different areas
,
 
such as online reviews (Moe and Schweidel, 2012) 
and 
user
-
generated 
contents (Susarla et al., 2012). 
I
n this study, we focus on the factors 
enabled
 
by featur
es of online 
communities such as funders

 
network 
embeddedness
, 

engagement with the online crowdfunding community, electronic word of mouth effects and 
entrepreneurs

 
human capital
. Our empirical results show significa
nt impacts of features of online 
communities on fundraising success in a crowdfunding platform.
 
 
T
his 
study
 
make
s
 
several 
contributions to 
the 
literature. First, we demonstrate the 
significant
 
impacts of two
-
sided networks on crowdfunding success. We exami
ne the impact of 

social learning

 
through social influence and interaction
s
 
on the probability of fund
rais
ing 
success. A few prior 
crowd
funding researchers 
have 
investigate
d
 
the impact of social influence 
and 
the 
social network
s
, but they 
have 
not actuall
y measure
d
 
social influence from the network 
characteristics
.
 
Additionally, while an each individual can decide to contribute 
to 
a certain project, 
the funding decision (i.e., whether each project creator will get funding or not) only depends upon 
the coll
ective intelligence. Therefore, online collaboration and interactions are more important in 
crowdfunding context. 
T
he interactions in online communities and social networking sites will 
50
 
 
easily provide funders with project information which alleviates the i
nformation asymmetry 
problem. 
 
 
Our empirical results support the significant impact of network embeddedness, 
both 
community and entrepreneurs

-
WOM
 
effects
 
on probability of fundraising 
success
. However, in the model examining the percent

 
engagement
 
do not significantly influence the percentage of pledged amounts. The results explain 

chanism, project 
creators need to
 
achieve their initial goals to get funding. Therefore, until a project reaches the 
initial goal a project creator attempt
s
 
to contact more funders and provide
s
 
them with project 
information. 
Once a 
project
 
reaches the original goal, the project creator
 
ma
y pay less attention or 
effort to raise additional funding. Therefore, entrepreneurial 
engagement
 
may not be a critical 
factor for additional fundraising. Our results in the selection bias model also support thi
s 
mechanism. 
 
 
A Rich
-
get
-
rich (or preferenti
al attachment) model can provide an alternative explanation 

may care about social information (e.g., who fund
ed
 
the projects, how others think of
 
them
, and 

network embeddedness within the crowdfunding platform significantly influence the percentage 

he probability of 
fundraising success.
 
However, we do not observe the impact of entrepreneurs

 
characteristics such 
as education 
level 
on fundraising success. In crowdfunding platform, funders

 
decisions depend 
more on social relationship, social 
interacti
ons
,
 
e
-
WOM,
 
and relationship with entrepreneurs than 
entrepreneurs

characteristics.
 
51
 
 
Entrepreneurs can obtain many benefits by using crowdfunding projects. First, they can 
easily access and obtain 
financial 
resources
. Crowdfunding allows entrepreneurs to e
nhance 
legitimacy of their projects. Compared to the traditional financing context, entrepreneurs can easily 
access potential investors and interact with them. Additionally, the success in crowdfunding 
project creates legitimacy for additional funding. Acc
ording to Mollick and Kuppuswamy (2014)

s 
survey, entrepreneurs who 
successfully
 
get funding from crowdfunding platforms are more likely 
to get the high percentage of ongoing venture investment and the high level of revenue (over 
$100,000). Furthermore, en
trepreneurs can obtain non
-
financial or indirect potential benefits (i.e., 
social capital and potential customers, delivering initial products and achieving innovation, new 
opportunity and on
-
going business), which will become a source of potential revenue
. 
Crowdfunding projects can be featured within the platform and outside of the platform (social 
media), and 
entrepreneurs
 
can obtain 
social capital 
(funders and friends in online social media) 
from their own projects. Entrepreneurs easily access potential 
customers who are either ardent 
funders or the 
anonymous
 
public. The online community interactions and e
-
WOM not only attract 
funders, but also help entrepreneurial innovation. Funders are able to participate actively in 
entrepreneurs

 
innovation process
. 
T
hey evaluate and share ideas about the project outcomes. 
Additionally, crowdfunding communities force entrepreneurs to 
complete their initial innovation 
plans
 
(manufacture & delivery outcomes). Lastly, based on the existing online social capital, 
addition
al funding, and 
credibility
, entrepreneurs
 
can 
initiate 
new innovation
.
 
However, t
his study has a few limitations. First, we measure social interactions and 

engagement 
using 
the 
volume of transactions. Secondly, we only focus one platform 
an
d a specific category (Technology). In future research, we intend to investigate the nature of 

52
 
 
can be categorized into a few criteria based on the characteristic
s of contents and information. 
Future work can apply our empirical models in different category and crowdfunding platforms. 
 
 
Despite a few limitations, o
ur study 
help
s
 
researcher
s
, potential funders
,
 
and entrepreneurs 
to 
understand 
the role of
 
online 
communities (i.e., funders

 
network 
embeddedness
, 


 
electronic word 
of mouth effects
) in crowdfunding success, and 
to
 
explain why some projects are more likely to 
succee
d in attracting funding.
 
Start
-
ups have suffered because of the lack of legitimacy (
Aldrich
 
and 
Fiol
, 1994)
. Crowdfunding 
offers the promise of
 
allow
ing
 
start
-
ups to enhance 
their 
legitimacy 
and provide a new venue for external resources.
 
At the same time,
 
however,
 
lack of success in 
fundraising c
ould hurt 

 
reputations. 
Indeed, more than
 
50% of Kickstarter projects 
fail to reach the
ir
 
goal. 
We also provide insights for potential funders and entrepreneurs as well as 
provide managerial implicati
ons to entrepreneurs who participate in crowdfunding platforms about 
what kind of entrepreneurial strategies should be pursued. The results also provide important tips 
for crowdfunding platform designers on how to design their platform. 
 
 
 
53
 
 
CHAPTER 3.
 
ESSAY
 
2: DYNAMIC FUNDRAISING PATTERNS AND ENTREPRENEURIAL 
PERFORMANCE IN CROWDFUNDING PLATFORMS
 
 


,
 
Tim Berry
 
 
3.1 
INTRODUCTION
 
 
 
The emergence of crowdfunding has demonstrated strong potential to unleash 
considerable changes in the business environment. 
This online fundraising platform provides 
entrepreneurs with new opportunities for funding, and ultimately fosters entrepreneurship
 
and new 
firm creation. The question of how entrepreneurs successfully raise money using such a platform 

certain issues have also come to light. According to CNN
 
Money (December 18, 2012), 84% of 

delivering the promised products. On May 1st, 2014, the first lawsuit was brought against 
crowdfunding project creators who faile
d to
 
deliver the promised outcome
9
. 
Mollick (2013) 
demonstrated that only 24% of successfully funded 
projects in 
Kickstarter delivered their promised 
outcomes

ranging from simple thank
-
you cards to physical products (e.g., hardware, software, 
games, and so
 
on)

to funders on time. The predicted success of project creators lacks external 
measures of validation that could help 
in
 
screening the quality of creators. Project quality also 
lacks quantifiab
le information. Hence, 
the 
crowdfunding platforms may 
have t
he ones that have 
ability to raise financial resources, but they are less likely to be successful at
 
generat
ing
 
successful 
                                        
        
 
9
 
http://www.washingtonpost.com/blogs/govbeat/wp/2014/05/02/the
-
first
-
state
-
lawsuit
-
over
-
a
-
crowdfunding
-
project
-
is
-
about
-
a
-
deck
-
of
-
cards/
 
54
 
 
innovations. Additionally, many crowdfunding platforms do not have strict enforcement 
mechanisms in place to eliminate hoax projects 
or to prevent moral hazard p
roblems among project 
creators
10
.
 
Kickstarter, one of the most popular crowdfunding platforms, does state that 

quality of the produ
cts (Kickstarter blog, 2
012). Rather, the platform does
 
guarantee the quality of 


ickstarter allows funders 
who have neither the professional expertise nor the ability to evaluate the quality of a project or 

funders may be drawn to a particula
r project based on its external visibility and bandwagon effects, 

crowdfunding is beneficial to entrepreneurs for accessing financial resources, they have a 
fundamen
tal responsibility to initiate and execute innovation, and to deliver the promised outcome 
by the predefined delivery deadline. Yet, more than 50% of technology projects funded through 
Kickstarter.com failed to deliver their outcomes on time. A negative en
trepreneurial performance 

negative experiences influence their funding behaviors. This will ultimately influence the overall 
ecology of digital funding and entrepre
neurship. Therefore, it is necessary to investigate 

that performance. There is a dearth of research, however, on entrepreneurial performance after 
                                        
        
 
10
 
In 2014
,
 
Kickstarter
 
changed their policy 
about
 
refunds and project creation to improve the rate of 
execution success
 
and to eliminate moral hazard and agent problems, but it still does not have 
a 
strict 
policy on project execution and delays. 5
% of projects are fraudulent projects, as determined by 
entrepreneurs returning funded amounts or failing to update their pages 
through 
the promised delivery 
date in Kickstarter
 
(Mol
l
ick, 2013)
 
55
 
 
fundraising 
success or on the mechanisms that help to explain why some successfully funded 
projects could not deliver their promised outcomes on time. To date, crowdfunding research has 
mainly focused on the antecedents of fundraising success. A few exploratory studie
s have 
examined how project delays are influenced by wording errors in project pages, by the number of 
backers, and by the percentage of funded amounts (Mollick, 2013). Nonetheless, the insights 
gained from such research are insufficient to explain the cri
tical variables that lead to successful 
(or failed) project performance. Moreover, they ignore why some projects with similar pledged 
amounts fail to deliver outcomes; why some over
-
funded projects cause negative project 
performance; and what the underlyin
g mechanisms are for over
-
funded projects. 
 
Success in technology project execution and innovation are driven by multiple factors 

capacity for innovation, pr
oject management skills, available resources, and environmental factors 
(Tatikonda and Rosenthal, 2000; Shane and Stuart, 2002; Lee, Lee, and Pennings, 2001). Success 
in project execution is related to project characteristics such as the novelty, complexit
y, and 
outcome types of the project; the level of product newness and project complexity often lead to 
undesirable project performance (Tatikonda and Rosenthal, 2000). Entrepreneurs spend more time, 
money, and effort on producing physical products, with th
e risk and uncertainty of innovation 
increasing as the level of newness rises. Project development appears to be a struggle when 
entrepreneurs have limited experience and execution capability (Gupta and Wilemon, 1990; 
Wheelwright and Clark, 1992). Prior en
trepreneurship and innovation studies have emphasized the 
critical role of entrepreneurial characteristics on entrepreneurial innovation performance (Shane 
and Stuart, 2002). Entrepreneurs with a higher education level and more relevant experience are 
more
 
likely to perform innovation successfully than others, because such resources are highly 
56
 
 

-
Barton, 1992; 
Stuart and Abetti, 1990). However, entrepreneurial performance in crowdf
unding projects may not 
depend only on their characteristics or innovation capability because of the unique attributes of 
fundraising in crowdfunding platforms. Entrepreneurs who have similar levels of experience and 
education will perform differently depe
nding on the fundraising environment, which includes the 
type of funders, the project characteristics, funding amounts and trends, and interactions between 
funders and entrepreneurs. Entrepreneurs may fail to manage a project because of a lack of early 
mom
entum (resources) or because of sudden hype
-
demands, which disperse their resources to 
other activities (i.e., attracting more funding, revising original plans, or finding necessary 
resources). 
 
Therefore, understanding the dynamics of fundraising could ex
plain more about why some 
projects with similar funding percentages differ in their performance or have different fundraising 
patterns. The few existing studies that link the total percentage of accumulated amount of funding 
with delays in delivery have no
t taken the history of fundraising dynamics over the fundraising 
period into account. The heterogeneity of the dynamics of crowdfunding projects cannot be 
explained sufficiently by simply looking at the total pledged amounts, overfunding (above the 
initial
 
fundraising goal amounts), and their impact on crowdfunding project performance, 
especially regarding on
-
time delivery. This paper investigates dynamic fundraising patterns in 
crowdfunding projects, how/why those patterns exist, and how different dynamic 
patterns are 
associated with project performance. The main research questions we address are: 
 
(1) Are there different dynamic fundraising patterns among crowdfunding projects? 
 
(2) Do the identified fundraising patterns help predict entrepreneurial perfor
mance? If so, how do 
the fundraising patterns influence entrepreneurial performance? 
 
57
 
 
 
We explain the mechanisms using such theoretical lenses as entrepreneurship, innovation, 
and bandwagon effects. Fundraising trends in crowdfunding combine three differen
t mechanisms: 


heterogeneous combinations that can be either a positive or 

entrepreneurial performance. By employing the functional data analysis (FDA) method, this study 

enables us to explain the differe
nt trajectories. We treat each funding trajectory as a unit of analysis, 
and identify four of the most critical fundraising patterns by performing functional principal 
component analysis (FPCA). Data from technology crowdfunding projects were collected fro
m a 
reward
-
based crowdfunding platform in the U.S. We scrutinized the performance of projects 
initiated and successfully funded from March 2012 to January 2013. There were a total of 303 
successfully funded projects that provided delivery information and d
aily funding data. After 
eliminating projects with goals of less than $200 and 
duration
 
of fewer than 7 days, the study had 
a sample of 285 successfully funded projects. Project performance was measured in three ways. 
First, did the project creators delive
r the promised outcome on time? In project management, 
budget, time, and cost are key performance measures. In crowdfunding platforms, project creators 
set a specific delivery date when they initiate a project. In this context, on
-
time delivery is an 
impor
tant and observable performance measurement. Prior innovation studies have suggested 
different innovation performance measures. Commercialization

bringing new products 
successfully to market

is one important measure because introducing a new product to mar
ket is 
a signal of the first stage of successful innovation (Balachandra and Friar, 1997). Thus, we also 
investigated factors that influence the commercialization of a final project outcome. Crowdfunding 
58
 
 
provides benefits to entrepreneurs by creating legit
imacy for additional funding opportunities. We 
therefore investigated which project creators received additional funding (e.g., venture capital, 
private equity) after successful fundraising in crowdfunding platforms. We measured project 
performance in thre
e ways: (1) delivery, (2) commercialization, or (3) venture capital funding. A 
functional regression model (FReg) shows that late infusion patterns (higher than average funding 
near the ending date) predicts negative performance in on
-
time delivery. The pa
ttern also shows 
that under
-
funding in the middle of fundraising and a sudden hyper
-
infusion in the last few days 
have negative impacts on on
-
time delivery and commercialization. Our main model does not 
observe the significant impact of early funding on pe
rformance. In our sample, only 3.7 % of 
projects had early funding patterns. This can explain a limitation of crowdfunding, in which it is 
difficult to raise early funding in crowdfunding platforms due to the inherent nature of projects 
and entrepreneurs. 
The study also finds 
a positive, critical role of 

in on
-
time project delivery and obtaining venture capital funding
. However, e
ntrepreneurial 
characteristics do not significantly infl
uence commercialization
. The FDA allows u
s to identify 
shapes in the trajectory curves that are associated with superior crowdfunding performance. We 

enhancing model fit and forecasting accuracy, whe
n compared with accumulated funding measures. 
This indicates the important role of 
the history of funding 
dynamics on performance. 
 
We adopted a latent instrumental
 
variable (LIV) approach (Ebbes
 
et al. 2005) to control 
for potential endogeneity concerns w
ith respect to the effect of fundraising patterns on project 
performance. Unobserved entrepreneurial quality factors may influence both fundraising type and 
project performance, which could lead to a spurious effect of funding patterns on project 
performan
ce. In particular, such unobserved quality factors could influence both early momentum 
59
 
 
and project performance. In this study, we used a Bayesian statistical estimation method and a 
Markov Chain Monte Carlo Simulation (MCMC) for LIV to increase the accurac
y of the 
estimation. We observed that significant coefficients of latent classes on early funding (the 
posterior means do not include zero in 97.5% of credible intervals). Additionally, to address the 
heterogeneity of projects, we clas
sified projects into 
two groups
 
such as 
crowd
-
based projects and 
exper
t
s

based projects
 

ANOVA (FANOVA) results show that different mean values between the two groups, and the 
permutation F
-
Test represent sign
ificant F
-

Crowd
-
based projects have a significant increasing pattern at that point in time. This also provides 
evidence of how dynamic fundraising patterns can address the heterogeneity of crowdfunding 
proje
cts and the underlying funding mechanisms. 
 
 
This study contributes to the entrepreneurship and crowdfunding literature by providing a 
comprehensive model to explain both project performance in crowdfunding markets and how 
fundraising patterns themselves i
nfluence project performance in crowdfunding markets. This 
study identifies the dynamic fundraising pattern using fine
-
grain data and advanced statistical 
models; it explains the underlying mechanisms of those dynamics and examines the impact of 
those patt
erns on crowdfunding project performance
.
 
Finally, it explains how fundraising patterns 
influence project performance in crowdfunding markets. The empirical evidence provides 
meaningful managerial and theoretical implications.
 
 
In the following section, we
 
review the extant theories and the relevant entrepreneurship 
and crowdfunding literature, proposing research hypotheses that link fundraising patterns with 
project performance. Next, we explain the data and introduce our empirical model. We then present 
t
he empirical results of our analysis. Finally, we discuss the theoretical and managerial 
60
 
 
implications of this study and suggest future research directions
.
 
 
3.2 
THEORETICAL BACKGROUND AND RESEARCH HYPOTHESES
 
 
3.2.1 
L
iterature
 
Review
 
 

 
There are different types of investo
r
s
,
 
based on 
their tendency to select projects and 
make 
investment 
decisions
.
 
Prior literature on finance, angel investments, and venture capital 
investments have demonstrated different types of investor strategies (tend
encies), how those are 
formed, and the relationship between investor strategies and financial performance. Grinblatt et al. 
(1995) discussed herding behaviors in combination with the alleged tendency of investors to follow 
momentum
-
based fads in mutual fun
ds purchases. Tyszka et al. (2008) argued that investment 


inves

-
taking attitudes (Siegel and Hoban, 1982; 
Filbeck, 
Hatfield, and Horvath
, 2005
). 
 

characteristics (e.g., passion, overco
nfidence, perception of entrepreneurial potential and attributes, 
and subjective judgement), and examined how those characteristics lead the investment decision
-
making in angel and venture capital investments (Mitteness, Sudek, and Cardon, 2012; Maxwell, 
J
effrey, and Levesque, 2011). Gompers and Lerner (2000) propose that the start
-
up evaluations 
depend on start
-
up characteristics as well as those of the investors. Others proposed that the impact 
61
 
 
of the available information and knowledge about a project an

investment decision (Campbell and Kirmani, 2000; Edmiston and Fisher, 2006; Wang, 2009). 
 
Unlike traditional efficient market hypothesis (EMH), lots of private or internal 
information which could not timely reflect the value
 
of start
-
up projects exists in start
-
up 
investment. Since not all inform
ation is publicly available, 
there are information asymmetry issues. 

-
up investment cannot always be explained by 
popular economic th
eories such as the Expected Utility Theory. A few prior studies of start
-
up 


ctive judgements 
(heuristics), and cognitive biases (i.e., overconfidence, overreaction, representative bias, and 
human errors) in information processing and decision
-
making (Koellinger et al., 2007; Maxwell 
et al., 2011). Additionally, in contrast to econ
omic theory, which assumes individuals consider only 
their own wealth and will not sacrifice to help others, it is not difficult to observe investors who 

-
stage start
-
u
p 
funding (Angel investments and crowdfunding). Along with the behavioral economics perspective, 

early funding comes from people who know the entrepreneurs pers
onally (i.e., friends and family), 
and are often for relatively large amounts of money (Ag
ra
wal et al., 2011). Table 
3.
1 provides a 
summary of the extant literature on investors, investment types, and start
-
up investments.
 
Although prior studies have addre
ssed different aspects of investor characteristics, 
investment tendencies, and related financial performance, none of them provide a comprehensive 
frame for investor types using multiple criteria. Since investment behaviors often occur for 
multiple reasons
, it is valuable to integrate multiple perspectives and different defined investor 
62
 
 
types. Furthermore, there is no prior study on crowdfunding that addresses either the types of 
investors or defines them with multiple criteria. Compared to other funding me
thods (such as 
Angel investments and VC investments), crowdfunding is an online investing process that has a 
specific goal and duration. Additionally, this platform provides high visibility of both the projects 
and the fundraising process. Thus, investors 
can easily observe the probability of fundraising 
success for the projects, and update their own belief in them. Offline investment (e.g. angel and 
venture capital investments) could not provide a real
-
time fundraising process, nor is it easy for 
funders t

 
Participants in crowdfunding platforms have very diverse background knowledge, 
experiences, motivations and social relationships. According to prior crowdfunding studies, the 
majority of funders in crowdfunding ar
e laymen who do not have professional knowledge of 
projects and entrepreneurs. Therefore, in a crowdfunding platform, funders experience substantial 

funding dec
isions. Most crowdfunding platforms do not scrutinize either the projects or the project 
creators. Instead, they shift this responsibility to the funders (Kickstarter blogs, 2012). As a result 

s shaped by: available information 
and knowledge about a project and an entrepreneur (Campbell and Kirmani, 2000; Edmiston and 
Fisher, 2006; Wang, 2009); his/her level of enthusiasm for the project (Barnewall, 1987; Filbeck
 
et al.
, 2005); risk tolerance (S
iegel and Hoban, 1982; Filbeck et al., 2005; Wang, 2009); resource 

, 
Hirshleifer, and Welch
,1992).
 
By combining prior studies, we classify investors with multip
le criteria. 
Table 3.2 provides 
the summary of the characteristics of investor types.
 
We classify funders into three categories: (1) 
63
 
 
Enthusiastic funders
, (2) 
Momentum funders
, and (3) 
Conservative funders
. 
First, 
e
nthusiastic 
funders
 
are
 
people who are excited about projects ideas, have 
ample 
knowledge and experience, 
are willing to take a risk and contribute larger
 
amounts
 
of 
money to projects in the early funding 
stage. The enthusiastic funders enjoy finance and analysis, and want to be
 
fully engaged in 
decision
-
making for their investments (Barnewall, 1987). The financial literature indicates 
that 

s
 
and characteristics 
are an important 
influence 
on their 
financial risk 
behaviors and risk tolerance (Siegel and 
Hoban, 1982; Schooley and Worden, 1999; Filbeck et al., 
2005). Active investors are more likely to take risk
s
 
(Barnewall, 1987) and
 
their
 
higher knowledge 
levels 
lead 
to 
different information search and investment behaviors (Campbell and Kirmani, 
2000). 
U
s
ing the Myers
-
Briggs Type Indicator (MBTI)
, 
Filbeck
 
et al.
(2005) f
ou
nd 
that 
the higher 

thinking, and perce
ption
. Thus, enthusiastic funders will make fundin
g decision in the early stage
s,
 
with larger amounts of funding. They tend to 
a 
have higher risk tolerance level than other types of 
investors. 
 
Second, m
omentum funders believe large increases in the funding amounts will be 
followed by good performance. Th
e basic idea of their funding strategy is that once a trend is 
established, it is more likely to continue in that direction
, rather
 
than move against the trend, thus 

during 

ly 


their own
 
funding 
decision
s
 
because 
they
 

 
about the probability of fundraising. 
Bandwagon effects refer to a tendency for people to follow the previous behaviors of critical 
mass

which corresponds to the crowd under conditions of uncertainty or ambiguity

and adopt 
64
 
 
or defer to the same behaviors t
hat others have already accepted
 
(Abrahamson and Rosenkopf, 

; Rosenkopf and Abrahamson, 1999). 
E
arlier studies emphasized 
the idea that individuals 
would 
make rational decisions to follow others based on their utility 
(Leibenst
ein, 1950). The more recent approach o
n
 
bandwagon effects does not always involve 
rational decisions
 

). Bandwagon 
effects 
may 

t the benefits 
that will accrue from adopting 
a
 
behavior (rational efficiency theories), or because of social, 
environmental or external pressures force 
those involved 
to adopt 
an 
idea, technology, policy, and 
product that 
a
 
number of organizations or individuals have already adopted (fad theories)
 
(Abrahamson and Rosenkopf, 1993

effects are motivated by several sources, including legitimacy, social relations, peer 
pressure, 

bandwagon effects have documented the momentum effect



s
 
or behaviors
 
(Xiong and Bharadwaj, 2014). 
Momentum funders will wait until projects achieve a certain threshold 
level 
and 
then contribute to 
them.
 
Final
ly, conservative funders seek to preserve their money by investing in lower risk 
projects (Barnewall, 1987; Filbeck 
et al., 2005) 
that have 
already reached the
ir
 
funding goals or 
were 
initiated by project creators with high visibility and prior relevant experience. The superstar 

-
takes
-

-
la
w 
distribution or Pareto distribution (Brynjolfsson, Hu, and Smith, 2003; Easley and Kleinberg, 
2010). Conservative investors have risk tolerances rangi
ng from low to moderate.
 
 
65
 
 
Table 
3
.1 
Literature R
eview 
 
Section
 
Topic areas
 
Authors
 
Key findings
 

Enthusiasm
 
 
Investment 
strategy (Investor 
types)
 
Grinblatt
 
et al.
 
(1995)
 
-
a
nalyzed the extent to which mutual funds purchase based on their past returns 
as well as their tendency to exhibit herding 
behaviors (the tendency to invest with 
the herd, in combination with the alleged tendency of investors to follow 
momentum based fads by buying past winners)
 
-
m
omentum investing affects the performance of the funds: purchase above 
intrinsic values cause low
er future performance
 
Investment 
strategy (Investor 
types)
 
Tyszka
 
et al.
 
(2008)
 
-
i
dentified four types of investment strategies (long
-
run momentum strategy, 
long
-
run contrarian strategy, short
-
run momentum strategy, and short
-
run 
contrarian strategy) for
 
predicting uncertain next events; 
-
 
argued that investment 

 
-
show the tend to predict that the next event will be a continuation of the recent 
rend (momentum strategy); found
 
that under the uncertainty about the next 
events, short
-
run momentum strategy is prevail.
 
Information 
sharing among 
Investors
 
Shiller and 
Pound (1989)
 
-
h
erding behaviors
 
-
found that the impact of direct interpersonal communications and word of mouth 
communications on individual investors investment
 
Investment 
strategy (Investor 
types)
 
Hirshleifer, 
Subrahmanyam, 
and Titman
 
(1994)
 
-
h
erding Behaviors
 
-
analyzed trading behaviors and equilibrium information acquisition when some 
investors receive common 
private information before others
 
-
found that herding patterns (mimicking earlier trades)
 
Investment 
strategy (Investor 
types)
 
Morrin
 
et al.
 
(2002)
 
-
explored patterns of decision making among professional security analysists and 
observed different 
investment patterns (momentum and contrarian patterns)
 
Investment 
strategy (Investor 
type and 
characteristics)
 
Kubinska
, 
Markiewicz, and 
Tyszka
 
(2012)
 
-
examined disposition effect for stock trading participants
 
-
found that contrarian investors are more 
prone to the disposition effect thant are 
momentum traders
 
 
66
 
 
Table 
3

)
 

Enthusiasm
 
Investment 
strategy 
 
De Bondt (1993)
 
-
f
ound that non
-

(momentum strategy: financial laymen 
consistently bet on trend continuation)
 

Characteristics
 
Mitteness, 
Sudek, and 
Cardon, (2012)
 
-
a
ngel investor characteristics
 
-
p
erceived passion is likely to play a significant role in the funding decision 
process.
 
-

characteristics influence how perceived passion translates into 
evaluations of funding potential.
 

Characteristics
 
Maxwell, 
Jeffrey, and 
Levesque (2011)
 
-
studied early stage business angel decision making
 
-
found that angel investors do not use 
a fully compensatory decision model 
wherein they weight and score a large number of attributes. Rather, they use a 
shortcut decision making heuristic known as elimination
-
by
-
aspects to reduce the 
available investment opportunities to a more manageable size
.
 
-
Investment decisions are made according to two stages
 

Characteristics
 
Madill, Haines, 
and Rlding
 
(
2005
)
 
-
identified aspects that differentiate business angels as early stage investors fro
m 
other investors.
 

Characteristics
 
Gompers and 
Lerner (2000)
 
-
h
ow start
-
up evaluations might not depend on start
-
up characteristics only, but 
also those of the investors
 
-
show that market conditions impact VC valuations 
 
Social 
relationships
 
Hochberg
, 
Ljungqvist, and 
Lu
 
(2010)
 
-
 
f
ound net
working affect valuations of newly founded companies.
 
Information 
 
Cumming and 
Dai (2011)
 
-
showed u shape relationship between fund size and firm valuations
 
-
 

bargaining power and 
valuations in addition to venture quality and market 
conditions
 
 
 
 
67
 
 
Table 
3
.1 

)
 

Enthusiasm
 
Information 
 
Sanders and 
Boivie (2004)
 
-
h
ow information asymmetry can lead to opportunistic behaviors in form of 
adverse selection (i.e., hidden information) and moral hazard (i.e., hidden 
actions). For this reason investors struggle to get valuable and reliable 
information. 
 
Information
 
Zheng,
 
Liu, and 
George
 
(2010)
 
-
two kinds of information that influence investor evaluation: internally generated 
information on the s
t
art
-

information on the start
-

-
company network attributes 
 
Information
 
Binks, Ennew, 
and Reed
 
(1992)
 
-
i
nformation asymmetry (investors try to evaluate companies based on the 
information they are provided by the founders or able to collect
 
Risk and 
Investment
 
Siegel and 
Hoban, 1982
 
-

and characteristics are an important influence on 
their financial risk behaviors and risk tolerance
 

characteristics
 
Mitteness et 
al.
(2012)
 
-

entrepreneurial pot
ential and attribute, and subjective judgement) and examined 
how those characteristics lead investment decision making in angel investment 
and venture capital investment
 
Entrepreneurial 
characteristics
 
Hsu (2007)
 
-
c
haracteristics of founders are important determinants in VC evaluations. (prior 
experience in founding, both human capital (e.g., training and prior professional 
experience), social capital (e.g. social skills and charisma) of the start
-
up 
founders are all
 
positively correlated with higher evaluations. 
 

characteristics 
 
Gompers and 
Lerner (200
1
)
 
-
start
-
up evaluations depend on start
-
up characteristics as well as those of the 
investors.
 

characteristics
 
Koellinger et 
al.(
2007
)
 
-
own 
subjective judgement (heuristics), cognitive biases (i.e., overconfidence, 
overreaction, representative bias, and human errors) in information processing 
and decision making based on the behaviors economics
 

enthusiasm
 
Barnewall 
(
1987
)
 
-
enthusi
astic funders enjoy finance and analysis, and want to be fully engaged in 
decision
-
making for their investments
 
Information and 
Professionalism
 
Campbell and 
Kirmani (
2000
)
 
-
higher knowledge levels lead to different information search and investment 
behav
iors
 
 
 
 
68
 
 
Table 
3
.1
 

 

Enthusiasm
 
Risk and 
enthusiasm
 
Filbeck et al. 
(2005)
 
-

extraversion, intuition, thinking, and perception.
 
Bandwagon 
Effects
 
Abrahamson 
and 
Rosenkopf
, 
1990; Fiol and 

 
 
-
b
andwagon effects refer to a tendency for people to follow the previous 
behaviors of critical mass

which corresponds to the crowd under conditions of 
uncertainty or ambiguity

and adopt or defer to the same beh
aviors that others 
have already accepted
 
Entrepreneurial 
improvisation
 
Resource 
dependency
 
Zimmerman and 
Zeitz, 2002
 
-
r
esources are crucial to new venture growth
 
Resource 
dependency 
 
Christensen and 
Bower, 1996
 
-

innovation can often be ascribed to insufficient 
resources or expertise
 
Resource 
dependency
 
Salancik and 
Pfeffer, 1978;
 
-
argued that entrepreneurs have strong external resource dependence; and should 
attract external actors to provide those resources by 
motivating their beliefs and 
feelings that the entrepreneurs are worthy, appropriate, and competent.
 
Resource 
dependency
 
Baker
, Miner, 
and Eesley 
(2003)
 
-

constrained by exte
rnal resources, and their actions are contingent upon the 
availability of resources
 
Resource 
dependency & 
acquisition of 
resources
 
Lounsbury and 
Glynn, (2001)
 
-
 
define cultural entrepreneurship as the process of storytelling that mediates 
between extant 
stocks of entrepreneurial resources and subsequent capital 
acquisition and wealth creation
 
Heterogeneous 
resource and 
performance
 
Teece, Pisano, 
and Shuen 
(1997)
 
-
 
c
ompany capabilities are heterogeneous among the market participants which 
lead to differe
nt company performance
 
Project 
characteristics
 
Project newness 
 
Wallace
, Keil, 
and Rai
 
(2004)
 
-
 
technological newness and application size influence project failure 
 
Project newness
 
Cooper
 
(1980)
 
-
 
project newness negatively influence new product 
success
 
Novelty, 
complexity
 
Tatikonda and 
Rosenthal (2000)
 
-
technology novelty, project complexity, and product development project 
execution success: a deeper look at task uncertainty in product innovation
 
 
 
 
69
 
 
Table 
3
.2
 
Investors Types 
 
 
Criteria
 
Investor Type
 
Enthusiastic
 
Investors
 
Momentum
 
Investors
 
Conservative 
Investors
 
Funding Timing
 
Early
 
Relative 
 
Late
 
Risk Tolerance
 
High
 
Low to medium
 
Low to medium
 
Level of Enthusiasm
 
High
 
Medium
 
Low
 
Resource Allocation
 
Large amounts
 
Small to medium
 
Small
 

Behaviors
 
Low
 
High
 
High
 
 

 
 
E
ntrepreneurs ultimately seek to generate new business value, 
so 
they engage in efforts to 
identify resources and opportunities (Shane &
 
Venkataraman, 2000). Resources are crucial 
to
 
new 
venture growth (Zimmerman and Zeitz, 2002), and the failure of 
a 
firm

s innovation can 
often 
be 
ascribed to insufficient resources or expertise (Christensen and Bower, 1996). The degree of 
newness
 
and
 
diff
iculty, relative to the skills and experience of the firm influences their survival 
(Christensen and Bower, 1996). However, most start
-
ups lack the financial, human, and 
technological resources because of 
their 
limited or nonexistent record
s
 
of performance
 
(Zimmerman and Zeitz, 2002). As a result, entrepreneurs have strong external resource 
dependence (Salancik and Pfeffer, 1978; Christensen and Bower, 1996; Zimmerman and Zitz, 
2002), and should attract external actors to 
provide those 
resources by motivati
ng their belief
s
 
and 
feeling
s
 
that the entrepreneurs are worthy
,
 
appropriate
,
 

strategic orientation and innovation success are 
often 
constrained by external resources, and their 
actions are contingent upon the avail
ability of resources (Baker et al., 2003; Christensen and Bower, 
7
0
 
 
1996; Lounsbury
 
and Glynn, 2001; 


execution of novel ac

,
 
and whether it 
meets existing goals and potential possible outcomes (Baker et al., 2003; Moorman and Miner, 
1998; Hmieleski, Corbett, and Baron, 2013). Accordingly, entrepreneurs must be capabl
e of 
framing executable strategic decisions to achieve the
ir
 
original goals and to move their firms in a 
more promising direction using the resources available to them in the moment (Hmieleski et al. 
2013). 
 


the 
opportunities and risks of their projects, according their own innovation capabilit
ies
 
(e.g., prior 
experience, knowledge) and funding trends, which will appear in different strategic actions (i.e., 
resource allocation and
 
social engagement). Entrepreneurs experience conflicts and challenges 
when facing the two main dilemmas 
of
 
crowdfunding platforms: (1) failure to attract adequate 
initial funding, causing a distraction from their focus
,
 
and (2) the risks associated with u
nexpected 
hype
-
funding. Because of limited resources, entrepreneurs 
need to 
effectively allocate resources 
such as time and effort. Entrepreneurs who achieve early 
funding 
momentum 
in
 
the overall 
fundraising process can put more effort into actual production than those who need to focus on 
raising money. To attract more funding, the latter type of entrepreneurs will be more likely to 
engage 
in
 
the
 
social community to explain the plausi
bility and credibility of their business. These 

 
differently. The former 
scenario 
may 
positively influence performance, 
whereas 
the latter one may not have a positive impact on 
performance. 
 
 
E
ntr

s
 
and knowledge from higher education are 
71
 
 
associated with 
increased 
innovation performance (Shane and Stuart, 2002). Entrepreneurs who 
have a higher education level and greater relevant experience are more likely to achie
ve 

innovation capabilit
ies
, including project management skills (Lee et al., 2001; Leonard
-
Barton, 
1992; Stuart and Abetti 1990). 
However, t
heir innovation capability
 
is often contingent on the 
availability of resources. According to 
the 
previous literature, improvisation behaviors can 
generate 
different 
effects on the performance of firms under specific circumstance
s
 
(Miner, Bassoff, 
and Moorman, 2001). Thus, dependin
g on the availability of resources and their own ability to 
cope with risks,
 
entrepreneurs 
with 
prior experience and a deep knowledge of their projects 
may 
or 
may 
not do justice to their 

execution. For example, over
-
funded projects will 
burden
 
entrepreneurs with the obligation to fulfill their commitments to funders by the deadline 
because of unexpected high demands with constrained resources and time
.
 
 
Project Characteristics
 
 
Project execution success is related to project characteristics suc
h as novelty, complexity, 
and outcome types. Tatikonda and Rosenthal (2000) point
ed out 
how project technology novelty 
(i.e., the newness of products) and project complexity (i.e., the number of product functions 
embodied in the product) influence performa
nce in product development. High
ly
 
novel and 
complex projects increase the level of task (
e.g.,
 
new project development) uncertainty. 
Table 3.3
 
In crowdfunding, each funder will receive rewards, such as the final outcome of a project, 
or simple gifts (e.g.
, thank
-
you message
s
, small gifts). In 
the 
Technology 
group 
(from March 2012 
to January 2013) in Kickstarter, 48.0% of crowdfunding projects
 
needed to deliver hardware,
 
and 
21.2% of them need
ed
 
to deliver software, respectively. 20% of 
the 
projects provide
d
 
funders with 
72
 
 
simple gifts or invitation
s
 
(Figure 3.1)
. Entrepreneurs need to spend more time, money, and effort 
on producing physical products than intangible ones
,
 
and the risk and uncertainty of innovation 
increase
s
 
as the level of newness rises. Project and innovation performance is rooted in multiple 

skills, 
and 
available resources (Lee et al., 2001; Tatikonda and R
osenthal, 2000; Shane and Stuart, 
2002).
 
 
 
Figure 
3
. 1. % of Projects by Reward Types
 
 
Moving Toward Conceptualization of 
Fundraising Patterns
 
in Crowdfunding
 
 
Landstrom (1998) argue that investment as a process in which decision making criteria 
may 
vary in the course of time
,
 
and the complexity of the investment decision process can be better 
understood if that process is broken down into several stages. 
The 
purpose of 
theoretical 
development is to highlight the importance of considering the shape of
 
the fundraising patterns, as 
researchers can miss important information regarding entrepreneurial performance in 
crowdfunding projects. Most crowdfunding platforms
11
 
use all
-
or
-
nothing mechanisms in which 
                                        
        
 
11
 
W
e collected data from a platform 
using the
 
all
-
or
-
nothing mechanism.
 
12.0%
48.0%
4.8%
0.4%
3.0%
10.6%
21.1%
% Projects by Reward Type
Gift
Hardware
Invitation
Nothing
Other product
73
 
 
each project has a predefined duration and funding 
goal. In order to obtain funding from the 
platform, project creators 
must 
raise funds by the predefined deadline
,
 
while 
funders are 
only 
able 
to contribute to a certain project 
within 
that time period, otherwise 
the 
entrepreneurs 
do not 
get 
funding nor do 
funders have a chance to contribute to the projects. Before the predefined funding 
deadline, 
there is 
little information available about the final outcome, and funders cannot judge its 
value
,
 
entrepreneurial potential 
or 
innovation performance. Additionall
y, entrepreneurs who 
launch crowdfunding projects can neither assure fundraising success upfront nor predict 
the 
exact 
demands for the outcomes. These uncertainties lead 
to 
different perception
s
, attitude
s
, and 
behaviors of funders and entrepreneurs toward
 
crowdfunding projects. Those differences are 
represented by the heterogeneity 
found 
in fundraising patterns. In this paper, we argue that 
fundraising trends in crowdfunding combine three different mechanisms
:
 

(orientation toward cr
owdfunding projects)
;
 

;
 
and (3) Project 
characteristics. The different speeds and directionality of 
the 
three mechanisms will engender 
various combinations, representing diverse fundraising trajectories and patterns. Entrepr

performance in crowdfunding projects tends to be contextually dependent on such heterogeneous 
combinations 
that
 

P

ial 
performance. 
 
In crowdfunding platforms, project crea
tors raise
 

 
through a process 


 
Funders can contribute small amounts of money to projects 
without any professional expertise or the ability to evaluate the quality of projects as well as 

als. Accordi
ngly, some funders 
will be more 
likely
 
to be 
drawn to a particular 

nature or entrepreneurial performance. Additionally, entrepreneurs who have si
milar levels of 
74
 
 
experience and education wil
l achieve different performance
 
levels 
depending on the fundraising 
environment. Such environments include project characteristics, funding amounts and trends, 
funders, and interaction
s
 
between funders and entrep
reneurs. In turn, the fundraising process 
results 
are a 
combination of dynamics among funders, entrepreneurs, and inherent project 
characteristics, which may vary across the projects. 
Previous research in marketing has shown that 

tentions and growth patterns can have a significant impact on future sales 
and prerelease buzz dynamics can continuously influence the dynamics of purchase intentions 
(Chintagunta and Lee, 2012). 
Xiong and Bharadwaj (2014) found 
ways in which 
the trajectory of 
buzz influences future marketing performance; they demonstrated the critical role that trajectory 
plays in increasing the predictive power of their research 
model. 
As we discussed earlier, there are 
heterogeneities in funder type
s
, entrep

By taking time and space dimensions into account in explaining 
the 
dynamics among 
the 
three 
forces (funders, entrepreneurs, and projects), we are able to identify previously unobserved 
dynamic 
patterns 
that 
influence project performance. The fundraising trajectory can explain the 
heterogeneity of 
the 
dynamics of crowdfunding projects. Fundraising dynamics can continuously 

t
,
 
and influence the 
dynamics of entrepreneurial performance. Understanding the dynamics of fundraising could 
explain more about why some projects with similar funding percentages differ in their performance 
or 
why 
they have different fundraising patterns.
 
In order to study the relationship between dynamic 
fundraising patterns in crowdfunding projects and the subsequent performance of entrepreneurs, 
we 
have 
conceptualize
d
 
four 
possible 
fundraising pattern
s that
 
consider
 
the
 
characteristics of 
the 
three forc
es (funders, entrepreneurs, and projects): (1) Early momentum pattern, (2) Bandwagon 
and late infusion funding pattern, (
3
) 
U
nderfunding and late burst pattern, and (4) 
R
ational funding 
75
 
 
pattern (average growth pattern). The four fundraising patterns are co
nceptually related to the 
phenomenon of resource scarcity, bandwagon effects and 
the 
characteristics of 
the 
three forces. 
Early momentum funding patterns impl
y
 
a low level of financial resource scarcity for the 
en
trepreneur in the early stage. 
Bandwagon an
d late infusion funding pattern represents projects 
which attract funders continuously 
throughout
 
the funding period, and 
then 
suddenly attract 
a 
huge 
number of funders
. U
nderfunding and late burst patterns represent projects which do not have high 
visibil
ity (either projects or entrepreneurs) and fail to attract funder
s
 
during 
the 
majority 
of 
the 
time period, but reach the funding goal right before the ending date. 
In 
rational funding patterns
,
 
f
unders may contribute to the projects not because of 
the 
band
wagon or momentum effect, but 
because of project and entrepreneurial characteristics. Furthermore, entrepreneurs can predict the 
future demand and will be able to make executable product development plans. 
Table 
3
.
3
 
explain
s
 
four fundraising patterns and their characteristics. 
 
 
Table 
3
.
3
 
Fundraising Patterns and Characteristics
 
 
Factors
 
Fundraising Patterns 
 
Early 
momentum
 
Late infusion
 
Underfunding
 
late burst
 
Rational 
funding pattern
 
Funding Variation 
Time
 
Early
 
Late
 
Late
 
Steady
 
Reaching Thresholds
 
Early
 
Early to mid
 
Late
 
Mid
 
Slack Resource
 
High
 
Mid
 
Low
 
Mid
 

Attention
 
Execution
 
Execution
 
Fundraising
 
Execution
 
Risks from 
D
emands 
P
rediction
 
Low
 
High
 
High
 
Low
 
Entrepreneurs

 
Visibility
 
High
 
High or Mid
 
Low
 
Mid or Low
 
 
 
76
 
 
3.2.2 
Research Model
 
 
Different dynamics in fundraising patterns exist in crowdfunding projects, and these 
dynamics may influence entrepreneurial performance because 
they
 
can explain more about the 
impact of unobserved entrepreneurial contingent behaviors or the nature of funding
. 
We measure 
project performance in three ways: (1) delivery, (2) commercialization, 
and 
(3) funding from 
alternative sources such as venture capi
tal and private equity after projects are initiated in 
crowdfunding platforms. 
Figure 3
.2
 
shows 
the research model of our study and serves as a roadmap 
for 
hypotheses development
.
 
 
 
 
Figure 3
.
2.
 
Research Model
 
 
 
 
77
 
 
3.2.3 
Hypothesi
s Development
 
 
F
undraising trajector
ies
 
can explain the heterogeneity of 
the 
dynamics of crowdfunding 
projects. Understanding the dynamics of fundraising could explain more about why some projects 
with similar funding percentages differ in their performance
,
 
or 
why 
they h
ave different 
fundraising patterns.
 
The 
fundraising patterns will differently influence entrepreneurial 
performance. Early momentum funding patterns impl
y
 
a low level of financial resource scarcity 
for the entrepreneur in the early stage. Prior studies hav
e shown that early funding in a campaign 
is often driven by friends and families
,
 
or 
by 
experts who have more knowledge about the projects 
(Agrawal et al.
, 2013). Thus, earlier funding momentum may represent a high probability of 
success in project executi
on. Early stage funding can create slack resources, which are defined as 
resources in excess of what is required, allowing an entrepreneur
 
time 
to adapt successfully to 
internal and external pressure and changes (Bourgeois, 1981;
 
Sharfman
, Wolf, Chase, and
 
Tansik
, 
1988). 
S
lack resources can positively influence entrepreneurial performance (Sharfman et al., 
1988). Project execution and innovation performance can be negatively influenced by hype
r
-
funding (late infusion) in two ways. 
First,
 
hype
r
-
funding creat
es risk regarding the execution of 
more outcomes, and resource allocation decisions could be suboptimal, with resources being 
constrained by the 
entrepreneur 
fulfilling commitments to his or her backers. Entrepreneurs are 
less likely to manage increased ex
pectations, 
thus 
undermining their performance. Also, 
entrepreneurs may fail to manage a project because of 
a 
lack of early momentum (resources) and 
sudden hype
r
-
demands 
that 
disperse their resources to other activities (i.e., attracting more funders, 
seek
ing necessary resources, and cha
nging original project plans). 
Projects might experience 

effects, as noted above, are not always based on rational decision
-
making.
 
On occasion, over
-
78
 
 
conformity can have 
a 
negative impact on society (Sunstein, 2005). The effects can often cause 
problems of over
-
legitimacy, raising money that exceeds initial goals and that overstate an 

-
legitimacy enforced b
y social communities in crowdfunding platforms 
may, in some cases, negatively influence project performance. To have stunning success
,
 

fundraising 
scope
,
 
can be either a plight or an 
opportunity. It all depend
s on how the entrepreneur copes with it. Entrepreneurs will face the 
challenge of dealing with 
a 
large and complex order fulfillment process. It is difficult for 
entrepreneurs to fathom, in advance, the exact scale of a project. In crowdfunding projects, w
hen 
a late hype
r
-
growth pattern is observed, the bandwagon effects may negatively i
nfluence potential 
performance.
 
After all, the projects may be of poor quality or the entrepreneurs 
may be 
ill
-
prepared 
to execute
 
them
. 
 
Additionally, some projects are underfunded for most of the fundraising period and then 
have a late infusion; such a pattern is likely to negatively influence project performance. These 
projects do not 
have high visibility of projects or entrepreneurs and 
do not 
actually 

attention
 
during the fundraising period.
 
The late burst is mainly driven by 
the 

fundraising efforts. The combination of 
a 
lack of crowd attention and 
the 

scattering reflects the quali
ty of 
the 
projects and entrepreneurial potential. Thus, this type of 
trajectory negatively influences project performance. 
 
Lastly, p
roject
s
 
with average growth patterns
 
are likely to be positively associated with 
project performance, because entrepreneurs
 
do not experience sudden demands and crowds have 
consistently paid attention to the projects.
 
Thus, entrepreneurs are more 
likely
 
to make executable 
product 
development
 
plans and focus on project execution. In addition, funders may contribute to 
the proje
cts not because of bandwagon or momentum effect, but because of project and 
79
 
 
entrepreneurial quality. 
Therefore, the 
fundraising patterns will differently influence 
entrepreneurial performance.
 
W
e propose the following hypothesis.
 
 
 
Hypothesis
: Entrepreneur
ial performance is 
differently 
influenced by 
fundraising 
patterns.
 
 
 
80
 
 
3.3 
DATA
 
 
 
3.3.1 
Data Collection
 
 
We collected funding projects data from Kickstarter.com, one of the most popular reward
-
based crowdfunding platforms in the U.S. Our data focuses on projects in the Technology category, 
which 
were 
initiated and successfully funded between March 2012 and Ja
nuary 2013. A prior study 
has shown that, within a crowdfunding platform, the project performance patterns (delivery rate) 
do not differ across categories (Mollick, 2013). 
There were
 
303 successfully funded projects
.
 
A
fter 
eliminating projects that d
id
 
not
 
provide any delivery information 
or 
longitudinal data, 
had 
small
-
size funding (less than $200), 
were 
short duration projects (less than 10 days), and 
those that 
canceled projects after fundraising success, there 
were
 
a total of 285 projects that provide
d
 
delivery 
information. The observation window was closed on December 2014.  
 
 
 
 
81
 
 
3.4 
METHODOLOGY AND EMPIRICAL MODEL
 
 
3.4.1 
Functional Data Analysis
 
We employed 
the Functional Data Analysis (FDA) method 
t
o identify fundraising patterns
,
 
model the longitudina
l fundraising process
,
 
and test our hypotheses. FDA has become important 
in statistics in other disciplines (Sood, James, and Tellis, 2009)
 
though
 
it is not common in 
Information Systems. The central paradigm of FDA is to treat each function or curve as 
a
 
unit of 
observation. Each function was examined as 
a
 
unit of observation
,
 
and this approach assumed 
smoothness and permitted as much flex
ibility as required by the data
 
(Ramsay and Silverman, 
2005; Sood et al., 2009). We appl
ied
 
the FDA approach by treatin
g the daily cumulative funding 
data of each project as 285 curves or functions. By taking this approach, we 
could
 
extend several 
standard statistical methods for use on the curves themselves (Sood et al., 2009). FDA can 
effectively incorporate entire fundr
aising histories (Ramsay and Dalzell, 1991). To explain the 
relationship between fundraising patterns (observed daily funding dynamics for each project until 

a single point of time), we face
d
 
the challenge of dimensionality and regressing multi
-
dimensional 
vectors on a scalar variable. 
 
The FDA approach helps overcome this challenge because it allows most of the variability 
in fundraising across 
the 
fundraising days 
to be captured 
with a few functional principal 
components, significantly reducing the dimensionality (Ramsay and Silverman, 2005). Functional 
principal component analysis (FPCA) help
ed
 
us to identify the 
shape 
patterns in the fundraising. 
FPCA provides a parsimonious, finite
-
dimensional representation for each curve
,
 
which helps us 
to 
understand variations among the curves. Additionally, this allow
ed
 
us to perform functional 
82
 
 
regression by treating the functional principal component sc
ores a
s independent variables. 
Prior 
stud
ies
 
show that 
a 
functional data analysis approach provides more accurate prediction
s
 
than other 
traditional approach
es
 
that
 
us
e
 
information from only one curve (Sood et al., 2009).
 
 
To identify fundraising patterns, we cr
eated functional data and assessed the trends
 
of the 
data

,
 
comprised of 
daily observations. 
We used penalized smoothing splines
 
(PENSS)
 
to recover the underlying continuous smooth 
fundraising dynamic
 
curve 





 
for each new product j ( j= 285) by removing random influences 
(Ramsay and Silverman 2005, Sood et al., 2009). The spline basis is defined by the sequence of 
kn
ots on the daily interval. 





 

 
T=60). Then the 
smoothing spline estimate is defined as the function, h(t)
,
 
which minimizes for a given value of 

>0 (
Sood et al., 2009).
 
 




































  
(1)
 
Roughness 
penalty
 
= 













 
The first squared error term in Equation (1) forces h(t) to provide a close fit to the observed 
data. The smoothing parameter 

 
controls the trade
-
off between 
the 
smoothness of function h and 
data fit (Foutz and Jank, 2010). We follow
ed
 
the standard practi
ce of choosing 

 
as the value that 
provides the smallest cross
-
validated residual sum of squared errors (
Sood et al., 2009). Figure 3
.
3
 
shows the mean of 
general cross
-
validation (
GCV
)
 
criterion 



-
2
, the mean of 

10
-
2
 
in our analysis.  
 
83
 
 
 
 
Fi
gure 3
.
3
 
Smoothing Parameter 
Lambda(

) and Mean of GCV
 
 
3.4.2 
Functional 
Principal Component Analysis
 
 
Next, we perform
ed a
 
functional principal component analysis (FPCA) to identify the 
most critical features of fundraising patterns by displaying the modes of functional variation. 
FPCA is used to extract the common temporal ch
aracteristics of a set of curves (Ramsay
, Hooker, 
and Graves
, 2009). 
Smoothing t
echniques 
we
re incorporated into the functional principal 
components analysis. We decompose
d
 
curves into functional principal scores. We denote
d
 
by 

























 
the 
i
 
smooth curves that are our approximations to the fundraising 
curves for each project. 
 



































 
(2)
 
Subject to the following orthogonality constraints: 
 















, 













 
1




. (3)
 
Where, 




 
is 
the 
eigenfunction
,
 
which represents the principal component func
tions and
 




 
the average c
urve over the entire population
.
 
T
he principal component scores
,
 


,
 
correspond to 
the i
th
 
curve representing the amounts that 





,
 
which varies in t
he direction defined by 





. 
 
84
 
 
3.4.3 
Empirical Model
 
(
Func
tional Regression Model)
 
 
 
We perform
ed
 


principal component scores as the independent variables. We examine
d
 
the impact of the ide
ntified 
fundraising patterns on en

 
project performance is influenced by 
multiple factors
, our study focuses on the
 
impact of fundraising patterns
 
on entrepreneurial 
performance. P
roject performance refers to whether entrepren
eurs delivered 
promised 
outcomes 
on time, commercialized projects, and obtained venture capital funding
. 
Project performance was 
measured with a binary variable
 
(deliver
:
 
1, otherwise: 0; commercialization: 1, venture funding:
 
1, otherwise: 0).
 
To test the
 
causal relationship between functional observations and 
the 
scalar 
response variable, we used 
a 
functional regression model, 
in particular, 
we used 
a 
functiona
l 
logistic regression model to 
forecast a binary response variable from a functional predictor whose 
observations are functions (Aguilera, 
Escabias, 
and Valderrama, 2006). 
The 
following is our 
performance model. The impact of fundraising patterns (

p
c
jp
) on project performance was 
exami
ned along with controlling other variables (

). 
 
 
Performance Model:  
y
j

0
 

P
p
=1 

p
c
jp
)
 

J
j
=1 

 
j
 

j
 
 
(4)
 
 
Where 

p
 
is the coefficient of the 
p
th FPCA score c
jp
, and 

 
is the vector of the coefficients of other 
variables. 
 
 
 
 
 
85
 
 
Control Variables
 
 
We also controlled several covariates to eliminate alternative explanation
s
. We control
led
 
entrepreneurial characteristics (i.e., prior experience, education), social capital (the number of 
funders, the volume of online community interactions), project cha
racteristics such as size of goal, 
average funding amounts, outcome type, seasonal effect
s
, entrepreneurial actions during 
the 
fundraising period, and geographical location. According to 
the 

performance is signaled by their
 
observable characteristics such as education and prior experience 
(Shane and Stuart, 2002; Stuart, Hoang and Hybels, 1999). 
Entrepreneurs who have a higher 
education level and more relevant experi
ence, knowledge, and skills 
will be more likely to have 
suc
cessful innovation
s
 
than others, because those resources are highly associated with 

-
Barton, 1992; Stuart and Abetti 
1990
)
. Using 
the 
resource
-
based view theory, prior research in strategic 
management has 

resources) play in their performance. Start
-

acquiring of intangible resources for survival or
 
growth, and
 
on
 
internally accumulated resources 
or capabilities (Barney, 1991; Lee et al.
,
 
2001). Lee et al. (2001) examined the 
influence of internal 
capabilit
ies
 
(i.e., technological knowledge, production skills) on technological start
-

Add
itionally, in IT project management research, the success of a project often depends on 

and Gauvreau, 2004). In venture capital literature, the most importan
t criteria for venture 
capitalists

 
funding decisions are the quality of the entrepreneurs and their team, 
as well as 
the 
entreprene


ty and 
characteristics are 
important
 
factors 
in crowdfunding project performance. Prior entrepreneurship 
86
 
 
and strategy literature has emphasized the role of social capital in entrepreneurial performance 

o
ften conditional on its social network (Granovetter, 1985). There is a correspondence between 

The crowdfunding community networks can provide both availabl
e resources and community 
knowledge and ideas about projects
,
 

entrepreneurs who have a strong network within a crowdfunding community are more likely to 
achieve a higher performance. Additionally, in cro
wdfunding platforms, online crowds can work 
as information regulators, because they can easily post their opinions and questions about project 
performance and shar
e them with other participants.
 
Entrepreneurs in crowdfunding platforms 
cannot ignore these a
ctions because it will negatively impact their legitimacy. Furthermore, the 
current community participants are more likely to be repeat funders or future customers. Thus, 
entrepreneurs might try to complete the projects due to their own will and the pressu
re of online 
communities. We measure
d
 

participation in online community as well as the total number of funders.
 
 
Measurement
 
 
 
Adams et al.,
(2006) reviewed prior innovation performance literatu
re and summarized 
innovation management measurement areas as being (1) inputs, (2) knowledge management, (3) 
innovation strategy, (4) organization culture, (5) portfolio management, (6) project management, 
and (7) commercialization. To measure crowdfunding
 
project performance, this study 
distinguished three outcomes, depending on the in
novation process
: (1) complete initial innovation 
plans (project performance), (2) introducing new products into the market (innovation 
87
 
 
performance), and (3) future growth an
d financial opportunities. The dependent variables are on
-
time delivery, commercialization, and obtaining additional funding (venture capital funding).
 
 
(1) 
Complet
ing
 
their initial innovation plans (delivery outcomes) 
-
Input Process
 
 
We observed
 
the actual performance of
 
projects and gathered project delivery, and 

blogs and websites. We defined project performance as whether entrepreneurs delivered their 
p
romised outcomes to funders. Project performance was coded using a binary variable (1: on time 
delivery, otherwise 0).
 
 
(2) 
Innovation performance: Introduce New Product into
 
the
 
Market
 
 
In addition to examining project performance in crowdfunding platform
s, we also 
examined whether each entrepreneur commercialized his or her product in the market. 
Commercialization can be a good measure of innovation. Commercial success in the market should 
be explained with more sophisticated methods. This study, however,
 
did not measure the actual 
sales or the commercial success. Funders hope t
o see positive outcomes from
 

 

market. Commercialization data was 

If the final outcomes are sold in the market, we coded 1(otherwise 0).
 
 
(3)
Growth/Financial Opportunity
 
 
The fundraising success in a crowdfunding project creates legitimacy for additional 
funding. Entrepreneurs who successfully obtain funding from crowdfunding platforms are more 
88
 
 
likely to get a high percentage of ongoing v
enture investment. Thus, we als
o
 
wan
ted 
to investigate 
which types of projects 
received
 
external funding a
fter initial 
fundraising success in a 
crowdfunding platform. Venture capital investment data was collected from 
PrivCo 
and 
Cr
u
nchbase. 
If project creators 
received
 
venture capital funding, it was coded 1, otherwise 0. 
 
 
Table 
3
. 
4
 
Frequency Table for Dependent Variables
 
DV
 
Delivery
 
Commercialization
 
Venture Capital Funding
 
Codes
 
Frequency
 
Percent
 
Frequency
 
Percent
 
Frequency
 
Percent
 
0
 
190
 
66.67
 
131
 
49.80
 
228
 
86.69
 
1
 
95
 
33.33
 
132
 
50.20
 
35
 
13.31
 
Total
 
285
 
100
 
263
 
100
 
263
 
100
 
 
Table 
3
.
4
 
shows
 
the
 
frequency of 
the 
dependent variables. In our sample, 33% of projects 
delivered 
their 
final outcomes on time. Over 6
7
% of 
the 
project
s
 
failed to deliver outcomes to 
funders by the estimated delivery date. With a two
-
year win
dow, 50.
2
% of projects 
were
 
ready to 
sell in 
the 
market
. The study included projects sold in public market places such as Amazon and 
app
-
stores (Android, apple store)
 
and private webs

31
% of 
the 
projects were able to get additional funding from venture capital and private investors. 
 
Table 
3
.
5
 
shows the definition of 
the 
variables and descriptive statistics. The maximum 
value of 
the 
pledged
 
percentage is 6264 (the maximum total pledged amounts are 2,945,885) and 
the log (goal) is 13.53($750,000). The standard deviation of 
the 
pledged percent
age is 440.89 and 
the 
mean is 170.
8
.
 
F

w
as
 
collect
ed
 
d
uring 
the fundraising period
. The maximum volume of comments is 1660. There are three education 

input 0. We looked at whether each project creator had prio
r crowdfunding experience or not 
89
 
 
(whether they created other projects or not). We also controlled for the reward type of each project 
if the final outcome was physical products (either hardware or software) or 
a 
warm glow type (gift, 
invitation, etc.). For
 
the second and third model (commercialization and venture capital funding), 
we ruled out projects whose outcome was not a physical produc
t (e.g., event, conference, one
-
time 
lecture). We used 263 projects for the second and third analyses. 
 
 
Table 
3
.5
 
Variable Definitions and Descriptive Statistics
 
Variables
 
Definition
 
# of 
Obs.
 
Mean
 
Std. 
Dev.
 
Min
 
Max
 
Delivery
 
=1 if a project deliver outcomes on 
time, otherwise =0
 
285
 
0.332
 
0.470
 
0
 
1
 
Commercializat
ion
 
=1 if a project commercialize the 
product, 
otherwise =0
 
263
 
0.504
 
0.500
 
0
 
1
 
VC Funding
 
=1 if a project get additional 
external funding, otherwise =0
 
263
 
0.133
 
0.340
 
0
 
1
 
%_pledged
 
Percentage of funded amounts
 
285
 
170.
8
 
440.89
 
0
 
6264
 
logGoal
 
Funding goal (amounts)
 
285
 
9.19
 
1.529
 
5.
30
 
13.5
3
 
Av
g
_amount
s
 
Average of funded amounts (total 
amounts/number of funders)
 
285
 
147.0
 
251.94
 
5.05
 
2469.
8
 
Education
 

(1=BS,2=MS,3=PhD, otherwise 0)
 
285
 
0.69
 
.831
 
0
 
3
 
Experience
 

 
285
 
0.154
 
0.36
 
0
 
1
 
Entrepreneur 
Engagement
 
The volume of actions (# of 
updates for a project)
 
285
 
3.59
 
4.42
 
0
 
29
 
Number of 
Funders
 
Total number of funders for each 
project
 
285
 
849.44
 
1715.7
 
4
 
11281
 
Funders 
Engagement
 
Total quantity of discussion for 
each project
 
285
 
134.89
 
258.48
 
0
 
1660
 
Num. of FB 
Friend
 

friend, if there is no FB sites = 0
 
285
 
274.73
 
398.11
 
0
 
3109
 
Location
 
Geographical location
 
285
 
345.50
 
218.38
 
110
 
999
 
Reward Type
 
=1 software, 2 hardware
 
263
 
1.693
 
0.462
 
1
 
2
 
 
 
90
 
 
3.5 
EMPIRICAL ANALYSIS AND RESULTS
 
 
3.
5
.
1
 
Explor
ing the Fundraising Patterns
 
 
We explore
d our data and found
 
that 
different fundraising patterns exist. Figure 
3.
4
 
shows
 
the
 
overall fundraising trajectories of crowdfunding projects. The 
left panel shows 
the 
cumulated 
funding trajectory curves for crowdfunding projects and the right panel represents the 
corresponding velocity curves (the first order differentiation) which represent the rate of change 
in funding amounts. 
These figures show the different history
 
and evolutionary path of fundraising 
across projects in our data.
 
 
 
[The Cu
mulated Funding Trajectory Curves]
 
 
 
[The 
Corresponding Velocity Curves]
 
Figure 
3
. 
4
 
Smoothing Sp
l
ines of Crowdfunding Projects 
 
* (Y axis: Cumulative Funding amounts/goal, X 
axis: Days)
 
 
 
 
 
 
91
 
 
 
Figure 
3.
5
 
shows the exam
ples of funding trends for 
different 
crowdfunding pr
oject
s
 
in 
our data. Each project
 
has different funding amounts and evolutionary paths.
 
 
 
 
Figure
 
3
.
5
 
Examples of Funding Trajectory for Each Project
 
 
Figure 
3.
6
 
shows two projects that have similar total funding 
but
 
have different fundraising
 
trajectories
.
 
The momentum effect also indicates that an upward trend in early funding is not the 
same as a high volume of early funding. Project A and B have the 
same pledged percentage (193%, 
pledged amounts/goals) amounts,
12
 
however, we observed different fundraising trajectories. 
Project A got more funding at an early stage while Project B got funding at a late stage. The entire 
history of fundraising dynamics ca
n contain additional information to forecast entrepreneurial 
outcomes above and beyond the accumulated funding amounts.
 
                                        
        
 
12
 
In crowdfunding platform
s
, each project has 
a different
 
goal (funding amounts)
;
 
to standardize
,
 
we 
calculate
d
 
the percentage
s
 
of funding amounts.
 
92
 
 
 
 
Figure
 
3
.
6
 
Examples of Different Fundraising Patterns
 
 
3.5.
2
 
Results
 
 
To reduce the dimensionality constraints, we conducted 
a 
functional principal component 
analysis (FPCA). The results of 
the 
FPCA showed the dominant modes of variation in the data. 
Because principal components are often easier to interpret 
if they are rotated, 
a 
VARIMAX rotation 
was performed. Figures 
3
.
7
 
and 
3
.
8
 
present the FPCA results (Figure 
3
.
8
: harmonics of FPCA) 
using 
a 
cumulative percentage of 
the 
daily funding amounts (= daily funding/project goal) with 
a 
VARIMAX rotation. We observ
e
d
 
the four most popular fundraising patterns (see 
the 
eigen value 
numbers in Figure 
3
.
9
)
,
 
w
hich explain 99% of variation 
in
 
funding
.
 
The eigenvalue graph shows 
that 
the optimal factors number 4. 
 
 
The first fundraising pattern
 
explains variations in the early stage of funding. This 
represents projects that obtain early infusions (above the average) 
in the first 
30% of 
the 
funding 
duration
,
 
and then exhibit average growth after achieving a certain momentum.
 
3.7$ of projects 
follo
ws this patterns.
 
The second
 
pattern 
explain
s
 
the large variation
 
of funding amounts
 
in the late 
Project A
:
 
Xprotolab 
Portable Oscilloscope
 
 
Project B
:
 
 
Lake 
Surfboard
 
93
 
 
stage of fundraising, which represent a project that has average growth in most of fundraising 
period and then, hype
-
funding (
a 
late funding infusion)
 
in the l
ate 
funding stage.
 
We found that 
21.6% of technology projects have late infusion patterns.
 
T
he third model represents under
-
funding 
through 
most of the fundraising period
, followed by
 
late hype
-
funding. Projects in this group do 
attract some funders in the
 
beginning of fundr
aising, before losing attention, 
and then get extra 
funding right before the ending date. 
This pattern is the most popular in crowdfunding context. 
More than 50% of projects have type 3 pattern. 
The fourth model shows av
erage growth fund
ing 
(19.1% of projects) patterns. 
 
Type 1. 
Early Momentum
 
(3.7% of projects
) 
 
Type 2. 
Late Infusion
 
(21.6% of 
projects
)
 
 
 
 
 
 
 
Type 3 Underfunding
 
(55.
6% of 
projects
)
 
Type 4 Rational Funding
 
(19.1% of 
projects
)
 
 
 
 
 
 
 
Figure 
3
.
7
 
Funding Principal Component Analysis
 
* 
A principal component represents 
a 
variation around the mean. The solid line is the mean curve 
of funding (t), the +++++
 
(
-----
) line represents adding (subtracting) some amount of the FP
CA
.
 
94
 
 
 
 
Figure 
3
.
8
 
Harmonics of FPCA Patterns
 
 
 
 
Figure 
3
.
9
 
Eigenvalue Number
 
 
We examine
d
 
the relationship between the shape of fundraising patterns and crowdfunding 
project performance. 
Table 
3
.
6
 
presents the functional regression analysis results. To elim
inate 
possible alternative explanations, we controlled for location effects, outcome types, and 
entrepreneurial 
experience
. In the initial project model (delivery), t
he results showed that the 
second
 
and third fundraising patterns have negati
ve impacts on 
project delivery.
 
95
 
 
Table 
3.
6
 
Results of Initial Project Performance 
 
Independent 
Variable
 
Initial Project Performance  
 
On Time Delivery
 
Initial Project Performance
 
On Time Delivery
 
Coefficients
 
Std. Err
 
Coefficients
 
Std. Err
 
Type 1
 
Early momentum
 
-
 
-
 
0.070
 
0.107
 
Type 
2
 
Late infusion
 
-
 
-
 
-
 
0.186*
 
0.108
 
Type 3 
underfunding
 
-
 
-
 
-
 
0.210**
 
0.106
 
Type 4
 
Rational funding
 
-
 
-
 
0.206*
 
0.115
 
Pledged percent
 
-
0.00
0
 
0.0000
 
-
 
-
 
Education1
 
0.206
 
0.2795
 
-
 
0.006
 
0.174
 
Prior Experience
 
0.736**
 
0.356
 
0.721**
 
0.363
 
LogGoal
 
-
 
0.268**
 
0.127
 
-
 
0.350***
 
0.127
 
Num. of FB 
Friends
 
0.0003
 
0.0003
 
0.0004
 
0.0003
 

Engagement(t
-
1)
 
-
 
0.003**
 
0.0014
 
-
0.003**
 
0.046
 
Number of 
funders
 
0.0002*
 
0.0001
 
0.0004*
 
0.0001
 
Average Funding
 
0.001*
 
0.0007
 
0.001*
 
0.0007
 
Month
 
0.107**
 
0.048
 
0.107**
 
0.048
 
Location
 
-
0.005
 
0.0006
 
-
0.005
 
0.0006
 
Reward Type
 
0.096
 
0.211
 
0.096
 
0.211
 
Log
-
likelihood
 
-
162.203
 
-
162.203
 


 
0.10 (Pseudo)
 
0.110 (Pseudo)
 
Number of obs
.
 
285
 
285
 
T
wo
-
tailed significance levels p<0.1, *; p<0.05, ** ; p<0.01, ***
 
 
Projects that suddenly obtain
ed
 
hype
-
funding limit entr

ies,
 
and make it difficult for entrepreneurs to perform their initial production plans. Additionally, 
underfunded projects with high variation
s
 
in the late stage of fun
ding
,
 
also negatively influence 
96
 
 

 
Entrepreneurs may focus more on luring funders
,
 
to achieve 
fundraising success
,
 
and the inherent quality of the project and the entrepreneurs may not be good 
enough to d
eliver outcomes on time. 
Fi
nally,
 
rational funding (average growth trajectory) has a 
positive impact on performance. In this case, entrepreneurs can predict potential demands and 
prepare
 
for them. In addition,
 
longer horizontal fundraising to reach a funding goal may inc
rease 
scrutiny from the crowd.
 
However, the signi
ficant impact of 
early momentum model was
 
not 
obser
ved
. This shows the limitation of crowdfunding projects. Only 3.7% of projects are of this 
type, which means projects in crowdfunding have difficulty raising earl
y fund
ing momentum. 
Although we 
did not
 
find the significant impact of the ear
ly momentum model, we still f
ound
 
the 
positive directionality of th
is pattern on project delivery.
 
We also observe
d
 
the positive impact of 
entrepreneurial prior experience on pro
ject performance.
 
The second model is co
mmercialization success
 
(see Table 3.7)
. We f
ound
 
a significant 
negative impact of underfunding on commercialization. This result reinforces our argument about 
fundraising patterns and the quality of the project and entrepreneurs. The inherent quality of the 
project and of entrepreneurs may not be good e
nough to introduce a product to market
. For 
commercialization, we f
ound
 
that entrepreneurs

 
experience
 
had 
no significant influence
 
on 
c
ommercialization
. This result 
indicates
 
that 
commercialization requires other capabilities such as 
marketing and technol
ogy development. Thus the crowdfunding experience may not be sufficient 
to explain prior 
commercialization
 
experience.
 
 
97
 
 
 
In the third model, we investigate
d
 
factors 
that
 
influence the opportunity to get extra 
funding from venture capitalist
s
 
or private equity
 
(see Table 3.8)
. We
 
did not
 
find
 
a 
significant 
impact 
of fundraising patterns 
on venture capital funding.
 
However, we f
ound
 
that experience is a 
critical factor in v
enture capital funding. 
Additionally, we observed the significant impact of size 
Table 
3
.7
 
Results of New Product into Markets
 
Independent 
Variable
 
New Product into Markets
 
Commercialization
 
New Product into Markets
 
Commercialization
 
Coefficients (p
-
value)
 
Std. Err
 
Coefficients (p
-
value)
 
Std. Err
 
Type 
1
 
Early momentum
 
-
 
-
 
0.136
 
0.135
 
Type 
2
 
Late infusion
 
-
 
-
 
-
0.170
 
0.113
 
Type 3 
underfunding
 
-
 
-
 
-
0.190*
 
0.112
 
Type 4
 
Rational funding
 
-
 
-
 
0.119
 
0.119
 
Pledged 
percent
 
0.001***
 
0.0004
 
-
 
-
 
Education1
 
-
0.250
 
0.280
 
-
0.092
 
0.168
 
Prior Experience
 
-
0.075
 
0.386
 
-
0.076
 
0.399
 
LogGoal
 
-
 
0.562***
 
0.142
 
-
0.591***
 
0.146
 
Num. of FB Friends
 
0.0004
 
0.0004
 
0.0005
 
0.0004
 
Entrepreneur 
Engagement(t
-
1)
 
0.007
 
0.032
 
0.021
 
0.034
 

Engagement(t
-
1)
 
0.0009
 
0.001
 
0.001
 
0.0009
 
Number of funders
 
-
 
0.0003**
 
0.0001
 
-
 
0.0004***
 
0.0001
 
Average Funding
 
0.00005
 
0.0007
 
-
0.00002
 
0.0007
 
Month
 
0.094**
 
0.049
 
0.087*
 
0.049
 
Location
 
-
0.002
 
0.0006
 
-
0.0003
 
0.0006
 
Reward Type
 
-
0.145
 
0.309
 
-
0.181
 
0.311
 
Log
-
likelihood
 
-
159.075
 
-
155.44
 


 
0.127
 
0.15
 
Number of obs
.
 
263
 
263
 
two
-
tailed significance levels p<0.1, *; p<0.05, ** ; p<0.01, ***
 
98
 
 
of projects on venture capital funding. Therefore, in venture capital funding, observable 
entrepreneurial experience and popularity of projects are more important than other f
actors
 
 
 
Table 
3
.8
 
Results of Growth/Financial Resource
 
Independent 
Variable
 
Growth/Financial Resource
 
Venture Capital Funding
 
Growth/Financial Resource
 
Venture Capital Funding
 
Coefficients
 
Std. Err
 
Coefficients
 
Std. Err
 
Type 
1
 
Early momentum
 
-
 
-
 
0.204
 
0.159
 
Type 
2
 
Late infusion
 
-
 
-
 
-
0.247
 
0.227
 
Type 3 
underfunding
 
-
 
-
 
-
0.224
 
0.192
 
Type 4
 
Rational funding
 
-
 
-
 
0.143
 
0.174
 
Pledged Percent 
 
0.001
*
 
0.0004
 
-
 
-
 
Education1
 
0.120 
 
0.264
 
0.153
 
0.485
 
Prior Experience
 
1.052**
 
0.647
 
1.075**
 
0.629
 
LogGoal
 
1.283***
 
0.595
 
1.283***
 
0.289
 
Num. of FB 
Friends
 
0.001
 
0.001
 
0.0006
 
0.0006
 
Entrepreneur 
Engagement(t
-
1)
 
-
0.001
 
0.001
 
-
0.002
 
0.044
 

Engagement(t
-
1)
 
0.0007
 
0.001
 
0.0005
 
0.001
 
Number of funders
 
-
0.0001
 
0.0002
 
-
0.0004*
 
0.0002
 
Average Funding
 
0.0002
 
0.0007
 
0.0001
 
0.0007
 
Month
 
0.150*
 
0.080
 
0.151*
 
0.083
 
Location
 
-
0.0001
 
0.001
 
-
0.0007
 
0.001
 
Reward Type
 
-
0.812*
 
0.498
 
-
0.880*
 
0.524
 
Log
-
likelihood
 
-
72.29
 
-
68.82
 


 
0.29
 
0.33(Pseudo)
 
Number of obs
.
 
263
 
263
 
T
wo
-
tailed significance levels p<0.1,
 
*; p<0.05, ** ; p<0.01, ***
 
99
 
 
3.5.
3
 
Robustness Checks 
 
 
Potential Endogeneity: Latent Instrumental Variable Approach
 
 
Endogeneity can arise from
 
multiple sources
: (1) relevant omitted variables, (2) 
measurement error
s
 
in the 
repressors
, (3
) the problem of 
self
-
selection, (4) simultaneity, and (5) 
serially 
correlated errors in the presen
ce
 
of a lagged dependent variable
 
(Zhang et al., 2009
)
. One 
way to 
overcome 
problems of endogeneity is to find instruments, based on economic theory or 
intuition (Greene, 200
0). 
Instruments are variables that 
correlated with
 
independent variables
, 
but 
are uncorrelated with the error term. Hence, instrumental variables cannot have a direct effect on 
the dependent variable. However, 
there a
re 
c
hallenges in using instrument varia
bles. In many cases, 
it is not easy to find available in
strument variables, 
and 
some
 
of those
 
available 
might not be a 
good quality instrument. 
Using bad quality instruments may result in estimates that are even more 
biased than OLS estimates (Bound et 
al., 1995; Hahn and Hausman, 2003
; Zhang et al., 2009
). 
 
We observed different fundraising dynamics and 
the 
influence of such patterns on 
entrepreneurial performance. 
To eliminate potential endogeneity issues from omitted variables, we 
controlled the impac

community, and entrepreneurial characteristics) on entrepreneurial project performance. 
Additionally, 
reverse causality may not be problematic in our context, because 
there is 
a 
time 
gap 
between 
fundraising and entrepreneurial project performance. The only concern we have is,
 
regarding
 
potential endogeneity issues from u
nobservable variables 
that
 
could influence both 
fundraising patterns and entrepreneurial performance.
 
Especially, 
an 
unobservable 
entrepreneurial 
quality
 
could influence both early momentum funding pattern
s
 
and entrepreneurial performance.
 
However, i
t is not easy to observe entrepreneurial quality information in crowdfunding projects
. 
100
 
 
Additionally, there is difficulty in
 
identifying 
an 
observable instrument variable for dynamic 
fundraising patterns. To resolve this potential endogeneity problem, we adopt
ed
 
a 
latent 
instrumental variable (LIV) approach.
 
This is a modeling approach to account for reg
ressor
-
error 
dependencie
s and
 
is used when observable instrumental variab
les are difficult to identif
y
 
(Ebbes 
et al., 2005)
. 
Th
e
 
Latent Instrumental Variables (LIV) approach estimates regression parameters 
regardless of the presence of regressor
-
error correlations
13
 
(Zhang et al, 
2009). 
This method solves 
the endogeneity problem without observable instrument variables (Ebbes et al., 2005; Zhang et al., 
2009). 
As this method does not rely on observable instruments, 
we can also avoid 
issues 
of 
availability, validity, and weakness of the instruments 
(Zhang et al
.
, 2009; Ebbes
 
et al.
, 2005). 
The 
main 
ide
a
 
of this approach is to introduce a binary unobserved 
instrumental variable
 
that partitions 
endogenous predictors into two parts, one uncorrelate
d and the other correlated with the error term 
in the main equation
 
(Ebbe
s 
et al., 2005; Zhang et al., 2009
). The inclusion of a latent instrument 
in a system of regression equations is similar to a linear structural relations model, in which a latent 
fact
or that is uncorrelated with t
he error term is specified
 
(Zhang et al., 2009)
. Ebbes et al. (2005) 
provide 
a way 
to estimate models with an endogenous explanatory variable. Zhang et al. (2009) 
extend the LIV approach by developing a Bayesian formulation th
at has the advantage of providing 
valid inferences and conducting sig
nificant tests.
 
To deal with 
a 
potential endogeneity problem (
the 
early momentum pattern), we 
introduce
d
 
an early momentum equation
 
(



14
. We specif
ied
 
the model as follows:
 
                                        
        
 
13
 
Traditional instrumental variable (IV) is typically used to correct for regress
-
error 
correlations.
 
 
14
 
We calculated and used the percentage of early funding amounts 
for 



 
(total funding 
amounts for early 30% of duration to measure (i.e. if the duration is 60days, it will be the total 
percentage of pledged amounts of first 20days). 
 
101
 
 




















 
(a)
 
























 
(b)
 
Where, 


: 
project performa
nce
; 

: coefficients of regressors;





 
instrumented 
variables; 


: other covariates; 


: manifest variable; 


: coefficient of manifest variable; 


:unobserved latent instrument variables; 

: coefficients of latent instrument variables
 
 
In the
 
early funding model (Equation 
b), the instrumented
 
average early funding (



) is a 
function of entrepreneur
ial characteristics
 


 
and an 
un
observed LIV, 


. We assume
d
 
that 


 
follows a Bernoulli distribution, 


 
~ B (


), where 


= P (


=1) is the instrument probability. 
The latent instrument 


 
partitions the variation in 


 
in such a way that a part of it (



) is 
uncorrelated with entrepreneurial performance error 



 
and another part of it 



 
is correlated 
with 



. The parameter 

 
represents the effect of the latent instrument 


 
on the average early 
funding. The instrumented average early funding
,
 



, instead of observed average early funding 
(



, appears in the entrepreneur
ial performance model
 
(Equation 
a
)
. 
The LIV effectively remove
s
 
early funding endogeneity from the model and enables consistent estimates of the effect of average 
early fu
nding on entrepreneurial performance (



.
 
W
e estimate
d
 
the model with 
a 
Markov Chain 
Monte Carlo (MCMC) using R and WinBugs. 
We check
ed
 
a few convergence diagnostics.
 
A trace 
plo
t is a plot for the iteration number against the value of the parameter at e
ach iteration. 
Figure 
3
.10 
shows
 
a 
well
-
mixed 
trace plot for each parameter
 
and density plots of the parameters.
 
We also 
d
o
 
not find significant autocorrelation. To determine
 
the
 
optimal number of burn
-
in, we conducted 
Gelman
-
Rubin diagnostics with three chains. 
W
ith 10,000 burn
-
in, we cannot observe over
-
dispersed starting points and three chains are well mixed. Next, we determine
d
 
the optimal number 
of 
iteration (length of chain

 
The MCMC 
102
 
 
estimation results for the entrepreneurial performance and
 
early funding equations appear
 
in Table 
3
.9
.
 
We find significant 
p
arameters 

 
on averag
e early funding
. 
The Posterior 
coefficient
s
 
of 


represent
 
significant
 
e
ffects of unobserved two groups
 
on
 


. Thi
s shows that early momentum 
could
 
be endogenous and unobserved project or entrepreneurial quality influence
s
 
early funding 
amounts. 
Howeve
r, we d
id
 
not observe the significant impact of 



 
(early funding momentum) 
on entrepren
eurial performance. This result
 
shows
 
that early momentum can be driven by prior 
visibility or unobserved 
entrepreneurial 
quality factors, but
 
the early momentum 
alone is not 
sufficient to
 
influence entrepreneurial performance. 
 
Table 
3
.9
 
MCMC Bayesian 
Results of 
Latent Instrument Model (Posterior)
 
(Burn
-
in 10,000; Iteration 20,
000)
 
*** significant
 
Independent 
Variable
 
Posterior
 
Quantiles for each variable
 
Mean
 
SD
 
2.5%
 
97.5%
 


 
-
1.780
 
0.059
 
-
1.157
 
1.187
 


 
-
2.347
 
0.077
 
-
1.780
 
1.277
 


 
-
1.309***
 
0.059
 
-
2.476
 
-
1.552
 


 
-
1.702
 
0.039
 
-
9.521
 
6.210
 


 
-
1.152
***
 
0.039
 
-
1.928
 
2.495
 


 
5.185
***
 
0.386
 
4.097
 
-
4.039
 
103
 
 
 
 
 
Fi
gure 3
.
10
 
Trace Plots and Density Plots
 
 
104
 
 
Heterogeneity of Projects: Crowd
-
 
vs. Expert
-
 
Based Funding
 
 
As we discussed earlier in this paper, funders

g
ies
-
how much and wh
en 
they contribute to projects
-
 
var
y
 
in accordance with their ability to evaluate project quality, 
their 
enthusiasm, and the tolerance of risk level.
 
Thus, some funders 
contribute to projects with
 
a great 
deal of
 
knowledge
 
and higher amounts of money in order to access 
an 
early version of in
novation 
outcomes, while 
others are drawn to a particular project 
base
d on 
simple hedonic interest, 
philanthropy
, 
fads
, or bandwagon effects rather than the intrinsic quality of 
the 
projects. 
Such 
a 
motivation makes funders avoid high
-
risk investment
s
,
 
and
 
lead
s
 
to
 
late
 
funding participation with 
small amounts of money
.
 
Therefore, w
e expect that projects which gathered money from those 
investors wil
l have more chances to represent
 
the late hype
-
funding patterns than others.
 
To 
address 
the 
heterogeneity of projects and the related fundraising patterns, we classified projects 
into two groups: 
(1) Crowd
-
based projects
 
and (2) Expert
-
based projects based on the data we 
collected from Kickstarter.com.
 
However, 
since 
Kickstarter.com does not prov
ide each
 

 
classif
ied
 
these groups using reward data. First, 
we 
investigated every level
 
of rewards, corresponding contribution amounts and the number of 
contributors for each project
. With those data, we calculated 
the 
contribution percentages of each 
reward level. Next, we 
investigated whether
 
a project 
received
 
more than 50% of 
its 
funding 
amounts from the 50% of lower reward levels
.
15
 
If projects belon
g to this group, we coded the
m
 
as crowd
-
based projects, otherwise we
 
coded 
them as 
expert
-
based projects. 
Figure 
3
.
11
 
shows the 
evolutionary fundraising path
s (left figure: funding trends; right figure: velocity of funding trends)
 
for 
the 
two grou
ps. Blue lines represent expert
-
based funding projects and 
red lines 
represen
t 
crowd
-
based projects
. 
In general, r
ed lines more represent hype
-
funding projects than blue ones in 
                                        
        
 
15
 
A l
ower rewards level is equivalent to lower contribution amounts. 
 
105
 
 
the figures.
 
To analyze the effect of 
a 
project

s
 
type on the shape
 
of crowdfunding curves (i.e., the 
fitted curves as responses), we adopted 
a 
functional analysis of variance (FANOVA) method. 
 
 
 
Figure 
3
.
11
 
Plots of Crowd Vs. Expert
-
based Model
 
(Red line: Crowds; Blue: Experts
)
 
 
To perform a FANOVA, we need to define a linear model. 
W
e
 
can then
 
find the best 
parameters (using regression).
 
Y(t) is 
a 
functional response, 


is the weights, 


 
is either 0 or 1. 
 
 
























 
 
There are 69 projects in
 
Group 1 (crowd
-
based, funding from relatively small a
mounts) 
and 216 projects in 
Group 2 (expert
-
based, funding from relat
iv
ely higher amounts). Figure 3
.12 
shows the means of the fitted funding curves for the crowd
-
based projects (n=69, black curve) and 
the expert
-
based projects (n=216, grey curve). As we see in this figure, crowd
-
based projects have 
a 
higher mean value of f
unding in the late stage than the expert
-
based funding.
 
106
 
 
 
 
 
Figure 3
.1
2
 
Mean Curves for Crowd
 
and
 
Expert
 
Types
 
 
Figure 3
.1
3
 
represent
s
 
the 
estimated 
functional regression coefficients
 
(right figure:



, left 
figure:



)
 
with corresponding confidence Interval
s (shaded). With the crowd
-
based projects as 
the reference group, the estimate discrepancy to reference group (



 
is d
ifferent from
 
the 
reference group. 
 
 
 
Type1: Crowd
-
based project
 
Type2: Expert
-
based project
 
107
 
 
 
   
 
Figure 3
.13
 
Plots of estimated beta curves with functional confidence intervals
 
 
 
Next, we perform
ed
 
pairwise comparisons of two groups using 
a 
functional permutation 
F
-
Test (1,000 replications) to check whether 
the 
two groups statistically differ
ed
 
from each other. 
Figure 3
.14 show 
show
s
 
the permutation F
-
Test results and p
-
value curve. We found statistically 
a 
significant p
-
value at the time of 60% of entire duration (40days/60days). Crowd
-
based projects 
have a significant increasing pattern at tha
t time point (pointwise 0.05 critical value). Therefore, 
crowd
-
based projects are more likely to have the late
 
hype
r
-
funding pattern
,
 
where the majority of 
funding comes from funders who have less expertise and are less sensitive to entrepreneurial 
innovat
ion performance. This also provides evidence of how dynamic fundraising patterns can 
address 
the 
heterogeneity of crowdfunding projects and underlying funding mechanisms. 
 
 
 
 
 
108
 
 
 
 
 
  
 
Figure 3
.
14
 
Plo
ts of Permutation F
-
Test and P
-
V
alue
 
 
 
 
109
 
 
3.6 CONCLUSION 
AND IMPLICATIONS
 
 
 
This study investigate
d
 
the dynamics of fundraising patterns and their impact on 
crowdfunding project performance. Using 
the 
entrepreneurship theory and bandwagon effects as 
a theoretical lens, we examined which fundraising dynamics (inc
orporating projects, investors, 

d
 
or hinder
ed
 

n IT
-
enabled funding platforms. 
 
Employing functional data analysis (FDA) methods, we identif
ied different fundraising 
dynamics across crowdfunding projects. FDA provides a set of techniques that can improve the 
prediction of empirical models (Sood et al., 2009). 
A f
unctional principal component analysis 
(FPCA) help
ed
 
us to identify the patterns o
f shapes in the fundraising. FPCA provides a 
parsimonious, finite
-
dimensional representation for each curve
,
 
which helps us understand the 
variations among the curves. Additionally, this allows us to perform functional regression by 
treating the functional
 
principal component scores as independent variables. Prior stud
ies
 
show 
that 
a 
functional data analysis approach provides 
a 
more accurate prediction than 
a
 
traditional 
approach of using information from only one curve (Sood et al.
, 2009).
 
 
Our empirical analysis results show different fundraising patterns that had yet to be 
explained. 
T
he results 
also 
show that the performance of projects 
with
 
late
 
funding
 
infusion was 

 
lack of preparation 
for unexpected demand
s
 
for outcomes resulting from slack resources. Additionally, projects that failed to create early 
momentum and were underfunded for most of the fundraising period, but had a late infusion
,
 
showed negative performances. This sheds light
 
on the process by which entrepreneurs generate 
heterogeneous value
,
 
and their ability to execute projects given ostensibly similar circumstances 
110
 
 
(in the sense of fulfilling the original funding goals). To improve performance, entrepreneurs need 
to prepare
 
for project execution early in their innovation process
,
 
and be equipped with the 
capability to cope with extreme hype
r
-
attention
,
 
so 
they do 


 
Online communities can provide available resource
s
,
 
as well as 
community knowledge and 

projects 
through
 
their own will
,
 
and from the pressure of online communities. In crowdfunding 
platforms, online crowds c
an work as information regulators because they can easily post their 
opinions and questions about project performance
,
 
and share them with other participants. 
Entrepreneurs in crowdfunding platforms cannot ignore these actions because doing so would hurt 
t
heir legitimacy. In addition, 
their
 
ab
ility
 
to obtain funds in crowdfunding platforms can signal 
their legitimacy to financial institutions and future investors. Failing to deliver their outcomes to 
online funders on time is a negative signal to future inv
estors. Thus
,
 
for the survival of their 
business and future funding, it is critical for sound entrepreneurs to deliver promised outcomes on 
time. 
 
Our study also supports the long
-

Entrepreneurs who
 
have pri
or visibility
 
have 
a better 
chance 
of gaining 
early momentum than 
others. We have emphasized that in order for entrepreneurs to achieve better performance they 
need to gain legitimacy and access to resources early in the fundraising stage. By doin
g this, they 
can reduce the burden of resource dependency and focus on 
the 
execution of projects. Projects in 
crowdfunding platforms need to overcome the inherent difficulty in obtaining early funding in 
crowdfunding platforms by using other intermediaries
 
such as online social media and establishing 
good track records. Along with this, crowdfunding platform providers need to implement design 
111
 
 
tools that enhance the visibility of new projects. Otherwise, crowdfunding becomes a place where 
project creators ar
e able to raise funds without ever delivering outcomes to their funders. 
In 
a
ddition, funders need to consider fundraising trajectories to avoid the risk of not having outcomes 
on time
,
 
and 
of 
investing in projects of low quality 
or 
entrepreneurs with 
low 
potential. We also 
f
ou

project experience positively influences project performance. 
 
T
his study has some limitations that could be addressed in future research. Fir
st, in many 
cases, delays are caused by unexpected external elements such as shipping problems, legal 
screenings, and so on. In a future study, it would be valuable to consider how entrepreneurs cope 
with these issues. While our findings are limited to one
 
technology category and platform, it would 
be interesting to investigate whether a different category has different dynamic patterns
,
 
and how 
they are associated with entrepreneurial performance. 
 
Despite the limitations, t
his study
 
provides managerial im
plications for both start
-
ups who 
want to raise funds and investors who contribute to projects and 
are 
seek
ing
 
sound projects in 
crowdfunding platforms. Entrepreneurs should emphasize their prior performance. If they have 
none, 
they should increase their
 
v
isib
ility to obtain early 
resources
.
 
Investors should look at the 

Our finding
s
 
offer novel and important implications for the theory and practice of project
 
performance in crowdfunding platforms. We expect that our work will contribute to and promote 
future research that enhances our understanding of crowdfunding participation and performance 
regarding resource access and new firm creation.
 
 
 
112
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
REFERENCES 
 
 
 
113
 
 
REFERENCES
 
 
Abrahamson, E., & Rosenkopf, L. (1993). Institutional and competitive bandwagons: Using 
mathematical modeling as a tool to explore innovation diffusion.
 
Academy of management 
review
,
 
18
(3), 487
-
517.
 
Adams, R., Bessant, J., & Phe
lps, R. (2006). Innovation management measurement: A 
review.
 
International Journal of Management Reviews
,
 
8
(1), 21
-
47.
 
Agarwal, R., Gupta, A. K., & Kraut, R. (2008). Editorial overview

 
The interplay between digital 
and social networks. Information Systems
 
Research, 19(3), 243
-
252.
 
Aggarwal, R., Gopal, R., Gupta, A., & Singh, H. (2012). Putting money where the mouths are: 
The relation between venture financing and electronic word
-
of
-
mouth.
 
Information 
Systems Research
,
 
23
(3
-
part
-
2), 976
-
992.
 
Agrawal, A. K.,
 
Catalini, C., & Goldfarb, A. (2011).
 
The geography of 
crowdfunding
 
(No.w16820). National Bureau of Economic Research.
 
Agrawal, A. K., Catalini, C., & Goldfarb, A. (2013).
 
Some Simple Economies of Crowdfunding. 
NBER Working Paper No. w19133.SSRN:http://ssr
n.com/abstract=2281044
 
Aguilera, A. M., Escabias, M., & Valderrama, M. J. (2006). Using principal components for 
estimating logistic regression with high
-
dimensional multicollinear data.
 
Computational 
Statistics & Data Analysis
,
 
50
(8), 1905
-
1924.
 
Aldrich, 
H. E., & Fiol, C. M. (1994). Fools rush in? The institutional context of industry 
creation.
 
Academy of management review
,
 
19
(4), 645
-
670.
 
Andreoni, J. (1990). Impure altruism and donations to public goods: a theory of warm
-
glow 
giving.
 
The economic journal
, 464
-
477.
 
Aral, S., Dellarocas, C., & Godes, D. (2013). Introduction to the Special Issue: Social Media and 
Business Transformation: A Framework for Research. 
Information Systems 
Research
,
 
24
(1), 3
-
13.
 
Aral, S., & Walker, D. (2012). Identifying influentia
l and susceptible members of social 
networks.
 
Science
,
 
337
(6092), 337
-
341.
 
Baker, T., Miner, A. S., & Eesley, D. T. (2003). Improvising firms: bricolage, account giving and 
improvisational competencies in the founding process.
 
Research policy
,
 
32
(2), 255
-
2
76.
 
Balachandra, R., & Friar, J. H. (1997). Factors for success in R&D projects and new product 
innovation: a contextual framework.
 
Engineering Management, IEEE Transactions 
on
,
 
44
(3), 276
-
287.
 
114
 
 
Barabási, A. L., &
 
Albert, R. (1999). Emergence of scaling in random networks. 
Science
,
 
286
(5439), 509
-
512.
 
Barnewall, M. M. (1987). Psychological characteristics of the individual investor.
 
ICFA 
Continuing Education Series
,
 
1987
(2), 62
-
71.
 
Barney, J. (1991). Firm resources
 
and sustained competitive advantage.
 
Journal of 
management
,
 
17
(1), 99
-
120.
 
Belleflamme, P., & Peitz, M. (2010). Platform competition and seller investment 
incentives.
 
European Economic Review
,
 
54
(8), 1059
-
1076.
 
Belout, A., & Gauvreau, C. (2004). Factors i
nfluencing project success: the impact of human 
resource management.
 
International journal of project management
,
 
22
(1), 1
-
11.
 
Bénabou, R., & Tirole, J. (2002). Self
-
confidence and personal motivation. 
Quarterly journal of 
economics
, 871
-
915.
 
Bikhchandani,
 
S., Hirshleifer, D., & Welch, I. (1992). A theory of fads, fashion, custom, and 
cultural change as informational cascades.
 
Journal of political Economy
, 992
-
1026.
 
Binks, M. R., Ennew, C. T., & Reed, G. V. (1992). Information asymmetries and the provision 
of 
finance to small firms.
 
International Small Business Journal
,
 
11
(1), 35
-
46.
 
Blankespoor, E., Miller, G. S., & White, H. D. (2013). The role of dissemination in market liquidity: 

 
The Accounting Review
,
 
89
(1), 79
-
112.
 
Borgatti, S. P., & Everett, M. G. (2006). A graph
-
theoretic perspective on centrality.
 
Social 
networks
,
 
28
(4), 466
-
484.
 
Boudreau, K. J., & Lakhani, K. R. (2013). Using the crowd as an innovation partner.
 
Harvard 
business review
,
 
91
(4), 60
-
69.
 
Brynjolfsson
, E., Hu, Y., & Smith, M. D. (2003). Consumer surplus in the digital economy: 
Estimating the value of increased product variety at online booksellers.
 
Management 
Science
,
 
49
(11), 1580
-
1596.
 
Bound, J., Jaeger, D. A., & Baker, R. M. (1995). Problems with ins
trumental variables estimation 
when the correlation between the instruments and the endogenous explanatory variable is 
weak. Journal of the American statistical association, 90(430), 443
-
450.
 
Bourgeois, L. J. (1981). On the measurement of organizational sl
ack.
 
Academy of Management 
review
,
 
6
(1), 29
-
39.
 
Burt, R. S. (1997). The contingent value of social capital.
 
Administrative science quarterly
, 339
-
365.
 
Burtch, G., Ghose, A., & Wattal, S. (2013). An empirical examination of the antecedents and 
115
 
 
consequences 
of contribution patterns in crowd
-
funded markets.
 
Information Systems 
Research
,
 
24
(3), 499
-
519
 
Busenitz, L. W., & Barney, J. B. (1997). Differences between entrepreneurs and managers in large 
organizations: Biases and heuristics in strategic decision
-
makin
g.
 
Journal of business 
venturing
,
 
12
(1), 9
-
30.
 
Busenitz, L. W., Fiet, J. O., & Moesel, D. D. (2005). Signaling in Venture Capitalist

New Venture 
Team Funding Decisions: Does It Indicate Long

Term Venture 
Outcomes?.
 
Entrepreneurship Theory and Practice
,
 
29
(
1), 1
-
12.
 
Campbell, M. C., & Kirmani, A. (2000). Consumers' use of persuasion knowledge: The effects of 
accessibility and cognitive capacity on perceptions of an influence agent.
 
Journal of 
Consumer Research
,
 
27
(1), 69
-
83.
 
Christensen, C. M., & Bower, J. 
L. (1996). Customer power, strategic investment, and the failure 
of leading firms.
 
Strategic management journal
,
17
(3), 197
-
218.
 
Chua, C. E. H., Wareham, J., & Robey, D. (2007). The role of online trading communities in 
managing internet auction fraud.
 
MIS 
Quarterly
, 759
-
781.
 
Chintagunta, P. K., & Lee, J. (2012). A pre
-
diffusion growth model of intentions and 
purchase.
 
Journal of the Academy of Marketing Science
,
 
40
(1), 137
-
154.
 
Cooper, R. G. (1980). Project NewProd: Factors in new product success.
European J
ournal of 
Marketing
,
 
14
(5/6), 277
-
292.
 
Cumming, D., & Dai, N. (2011). Fund size, limited attention and valuation of venture capital 
backed firms.
 
Journal of Empirical Finance
,
 
18
(1), 2
-
15.
 
De Meza, D., & Southey, C. (1996). The borrower's curse: optimism, 
finance and 
entrepreneurship.
 
The Economic Journal
, 375
-
386.
 
De Bondt, W. P. (1993). Betting on trends: Intuitive forecasts of financial risk and 
return.
 
International Journal of forecasting
,
 
9
(3), 355
-
371.
 
Dean, J. W., & Brass, D. J. (1985). Social intera
ction and the perception of job characteristics in 
an organization.
 
Human Relations
,
 
38
(6), 571
-
582.
 
Deeds, D. L., Decarolis, D., & Coombs, J. E. (1997). The impact of firm specific capabilities on 
the amount of capital raised in an initial public 
offering: Evidence from the biotechnology 
industry.
 
Journal of Business Venturing
,
 
12
(1), 31
-
46.
 
Dellarocas, C. (2003). The digitization of word of mouth: Promise and challenges of online 
feedback mechanisms.
 
Management science
,
 
49
(10), 1407
-
1424.
 
Dholakia
, U. M., Bagozzi, R. P., & Pearo, L. K. (2004). A social influence model of consumer 
participation in network
-
and small
-
group
-
based virtual communities.
 
International journal 
116
 
 
of research in marketing
,
 
21
(3), 241
-
263.
 
Easley, D., & Kleinberg, J. (2010).
 
Net
works, crowds, and markets: Reasoning about a highly 
connected world
. Cambridge University Press.
 
Ebbes, P., Wedel, M., Böckenholt, U., & Steerneman, T. (2005). Solving and testing for regressor
-
error (in) dependence when no instrumental variables are avai
lable: With new evidence for 
the effect of education on income.
Quantitative Marketing and Economics
,
 
3
(4), 365
-
392.
 
Eisenmann, T., Parker, G., & Van Alstyne, M. (2011). Platform envelopment. 
Strategic 
Management Journal
, 32(12), 1270
-
1285.
 
Everett, M. G., 
& Borgatti, S. P. (2005). Extending centrality.
 
Models and methods in social 
network analysis
,
 
35
(1), 57
-
76.
 
Fiol, C. M., & O'Connor, E. J. (2003). Waking up! Mindfulness in the face of 
bandwagons.
 
Academy of management review
,
 
28
(1), 54
-
70.
 
Finin, T., Jos
hi, A., Kolari, P., Java, A., Kale, A., & Karandikar, A. (2008). The information 
ecology of social media and online communities.
 
AI Magazine
, 
29
(3), 77.
 

-
Generated Content: 
Evidence
 
from Online Product Reviews. 
Information Systems Research
, 222 

 
238
 
Granovetter, M. (1985). Economic action and social structure: the problem of 
embeddedness.
 
American journal of sociology
, 481
-
510.
 
Granovetter, M. (2005). The impact of social structure 
on economic outcomes.
 
The Journal of 
Economic Perspectives
,
 
19
(1), 33
-
50.
 
Grewal, R., Cline, T. W., & Davies, A. (2003). Early
-
entrant advantage, word
-
of
-
mouth 
communication, brand similarity, and the consumer decision
-
making process.
 
Journal of 
Consumer P
sychology
,
 
13
(3), 187
-
197.
 
Grewal, R., Lilien, G. L., & Mallapragada, G. (2006). Location, location, location: How network 
embeddedness affects project success in open source systems.
 
Management Science
,
 
52
(7), 
1043
-
1056.
 
Greene, W. H. (2008). The economet
ric approach to efficiency analysis. The measurement of 
productive efficiency and productivity growth, 92
-
250.
 
Grinblatt, M., Titman, S., & Wermers, R. (1995). Momentum investment strategies, portfolio 
performance, and herding: A study of mutual fund behav
ior. 
The American economic 
review
, 1088
-
1105.
 
Gompers,
 
P., and Lerner, J. (2001).
 
The venture capital revolution.
 
The Journal of Economic 
Perspectives
 
15.2: 145
-
168.
 
117
 
 
Gupta, A. K., & Wilemon, D. L. (1990). Accelerating the development of technology
-
based ne
w 
products.
 
California management review
,
 
32
(2), 24
-
44.
 
Ebbes, P., Wedel, M., Böckenholt, U., & Steerneman, T. (2005). Solving and testing for regressor
-
error (in) dependence when no instrumental variables are available: With new evidence for 
the effect of
 
education on income. Quantitative Marketing and Economics, 3(4), 365
-
392.
 
Edmiston, K., & Fisher, M. G. (2006).
 
Financial education at the workplace: Evidence from a 
survey of Federal Reserve Bank employees
 
(No. 2006
-
02). Federal Reserve Bank of 
Kansas Ci
ty.
 
Filbeck, G., Hatfield, P., & Horvath, P. (2005). Risk aversion and personality type.
 
The Journal of 
Behavioral Finance
,
 
6
(4), 170
-
180.
 
Foutz, N. Z., & Jank, W. (2010). Research note
-
prerelease demand forecasting for motion pictures 
using functional sha
pe analysis of virtual stock markets. 
Marketing Science
,
 
29
(3), 568
-
579.
 
Hahn, J., & Hausman, J. (2003). Weak instruments: Diagnosis and cures in empirical econometrics. 
American Economic Review, 118
-
125.
 
Heckman, J. J. (1979). Sample selection bias as a s
pecification error.
 
Econometrica: Journal of 
the econometric society
, 153
-
161.
 
Hirshleifer, D., Subrahmanyam, A., & Titman, S. (1994). Security analysis and trading patterns 
when some investors receive information before others.
 
Journal of Finance
, 1665
-
16
98.
 
Hirshleifer, D., Low, A., & Teoh, S. H. (2012). Are overconfident CEOs better innovators?.
 
The 
Journal of Finance
,
 
67
(4), 1457
-
1498.
 

and firm performance: 
A study of dispositional and environmental moderators.
 
Strategic 
Entrepreneurship Journal
,
 
7
(2), 138
-
150.
 
Hochberg, Y. V., Ljungqvist, A., & Lu, Y. (2010). Networking as a barrier to entry and the 
competitive supply of venture capital.
 
The Journal of Finan
ce
,
65
(3), 829
-
859.
 
Hsu, D. H. (2007). Experienced entrepreneurial founders, organizational capital, and venture 
capital funding.
 
Research Policy
,
 
36
(5), 722
-
741.
 
Ibarra, H., &
 
Andrews, S. B. (1993). Power, social influence, and sense making: Effects of network 
centrality and proximity on employee perceptions. 
Administrative science quarterly
, 277
-
303.
 
Janney, J., & Folta, B. (2003). Signaling through private equity placements a
nd its impact on the 
valuation of biotechnology firms.
 
Journal of Business Venturing
,
 
18
(3), 361
-
380.
 
Jenssen, J. I., & Koenig, H. F. (2002). The effect of social networks on resource access and business 
118
 
 
start
-
ups.
 
European Planning Studies
,
 
10
(8), 1039
-
10
46.
 
Kirsch, D., Goldfarb, B., & Gera, A. (2009). Form or substance: the role of business plans in 
venture capital decision making.
 
Strategic Management Journal
,
 
30
(5), 487
-
515.
 


entrepreneurial behavior.
 
Journal of economic psychology
,
28
(4), 502
-
527.
 

momentum investors.
 
Journal of Behavioral Finance
,
 
13
(3), 214
-
225.
 
Kup
puswamy, V., & Bayus, B. L. (2013).
 
Crowdfunding Creative Ideas: the Dynamics of Projects 
Backers in Kickstarter
. SSRN Working Paper, http://papers. ssrn. com/sol3/papers.cfm.
 
Landström, H. (1998). Informal investors as entrepreneurs.
 
Technovation
,
18
(5), 3
21
-
333.
 
Lee, J. M.
, 
Hwang, B. H., 
&
 
Chen, H.(2015). Behavioral Differences between Founder CEOs and 
Professional CEOs in Large Public Firms: The Role of Overconfidence. 
Available at 
SSRN:
 
http://ssrn.com/abstract=2510549
 
 
Lee, C., Lee, K., & Pennings, J. M
. (2001). Internal capabilities, external networks, and 
performance: a study on technology

based ventures.
 
Strategic management journal
,
 
22
(6

7), 615
-
640.
 
Leenders, R. T. A., & Gabbay, S. M. (Eds.). (1999).
 
Corporate social capital and liability
. Springer 
Science & Business Media.
 
Leibenstein, H. (1950). Bandwagon, snob, and Veblen effects in the theory of consumers' 
demand.
 
The Quarterly Journal of Economics
, 183
-
207.
 
Leland. E., and Pyle, D. (1977), information asymmetries, fin
ancial structure and financi
al 
intermediation
. Journal of Finance
 
32, 371
-
387.
 
Leonard

Barton, D. (1992). Core capabilities and core rigidities: a paradox in managing new 
product development.
 
Strategic management journal
,
 
13
(S1), 111
-
125.
 
Lin, N. (1999). Building a network theory of 
social capital.
 
Connections
,
 
22
(1), 28
-
51.
 
Lin, M., Prabhala, N. R., & Viswanathan, S. (2013). Judging borrowers by the company they keep: 
friendship networks and information asymmetry in online peer
-
to
-
peer 
lending.
 
Management Science
,
 
59
(1), 17
-
35.
 
Lusch
, R. F., & Nambisan, S. (2015). Service innovation: A service
-
dominant (SD) logic 
perspective.
 
MIS Quarterly
, 39(1), 155
-
175.
 
Lounsbury, M., & Glynn, M. A. (2001). Cultural entrepreneurship: Stories, legitimacy, and the 
acquisition of resources.
 
Strategic 
management journal
,
22
(6

7), 545
-
564.
 
119
 
 
Ma, M., & Agarwal, R. (2007). Through a glass darkly: Information technology design, identity 
verification, and knowledge contribution in online communities. 
Information systems 
research
,
 
18
(1), 42
-
67.
 
Madill, J. J., Ha
ines, Jr, G. H., & RIding, A. L. (2005). The role of angels in technology SMEs: A 
link to venture capital.
 
Venture Capital
,
 
7
(2), 107
-
129.
 
Maxwell, A. L., Jeffrey, S. A., & Lévesque, M. (2011). Business angel early stage decision 
making.
 
Journal of Busines
s Venturing
,
 
26
(2), 212
-
225.
 
Melville, P., Gryc, W., & Lawrence, R. D. (2009, June). Sentiment analysis of blogs by combining 
lexical knowledge with text classification. In
 
Proceedings of the 15th ACM SIGKDD 
international conference on Knowledge discovery 
and data mining
 
(pp. 1275
-
1284). ACM.
 
Miner, A. S., Bassof, P., & Moorman, C. (2001). Organizational improvisation and learning: A 
field study.
 
Administrative science quarterly
,
 
46
(2), 304
-
337.
 
Mitteness, C., Sudek, R., & Cardon, M. S. (2012). Angel invest
or characteristics that determine 
whether perceived passion leads to higher evaluations of funding potential.
 
Journal of 
Business Venturing
,
 
27
(5), 592
-
606.
 
Moe, W.W., D.A. Schweidel. (2012). Online Product Opinions: Incidence, Evaluation, and 
Evolution.  
Marketing Science. 31(3), 372
-
386.
 
Mollick, E. (2012). The Dynamics of Crowdfunding: Determinants of Success and Failure. 
SSRN.doi:10.2139/ssrn.2088298.
 

Journal of 
Business Venturing, Forthcoming
.
 
Mollick, E. R., & Kuppuswamy, V. (2014). After the campaign: Outcomes of crowdfunding.
 
UNC 
Kenan
-
Flagler Research Paper
, (2376997).
 
Moorman, C., & Miner, A. S. (1998). Organizational improvisation and organizational 
memory.
 
Academy of management Review
,
 
23
(4), 698
-
723.
 
Morrin, M., Jacoby, J., Johar, G. V., He, X., Kuss, A., & Mazursky, D. (2002). Taking stock of 
stockbrokers: Exploring momentum versus contrarian investor strategies and 
profiles.
 
Journal of Consumer Research
,
 
29
(2), 188
-
198.
 
Nambisan, S., & Sawhney, M. (2011). Orchestration processes in network
-
centric innovation: 
Evidence from the field.
 
The Academy of Management Perspectives
,
 
25
(3), 40
-
57.
 
Ordanini, A., Miceli, L., Pizzetti, M., Parasuraman, A. (2011). Crowd
-
funding: transforming 
customers into investors through innovative service platforms. 
Journal of Service 
Management
,
 
22
(4), 443
-
70.
 
Pang, B., & Lee, L. (2008). Opinion mining and sentiment analysis. 
Foundations and trends in 
120
 
 
information retrieval
,
 
2
(1
-
2), 
1
-
135.
 
Phelps, J. E., Lewis, R., Mobilio, L., Perry, D., & Raman, N. (2004). Viral marketing or electronic 
word
-
of
-
mouth advertising: Examining consumer responses and motivations to pass along 
email.
 
Journal of advertising research
,
 
44
(04), 333
-
348.
 
Polloc
k, T. G., & Rindova, V. P. (2003). Media legitimation effects in the market for initial public 
offerings.
 
Academy of Management Journal
,
 
46
(5), 631
-
642.
 

Cu

Journal of 
International Communication
, 2(2), 67
-
86.
 
Pollock, J. C., & Whitney, L. M. (1997). Newspapers and racial/ethnic conflict: Comparing city 
demographics and nationwide reporting on th
e Crown Heights (Brooklyn, NY) incidents. 
Atlantic Journal of Communication, 5(2), 127
-
149.
 
Prahalad, C. K., & Ramaswamy, V. (2004). Co

creation experiences: The next practice in value 
creation.
 
Journal of interactive marketing
,
 
18
(3), 5
-
14.
 
Prasad, D., 
Bruton, G. D., & Vozikis, G. (2000). Signaling value to business angels: the proportion 
of the entrepreneur's net worth invested in a new venture as a decision signal.
 
Venture 
Capital: An international journal of entrepreneurial finance
,
 
2
(3), 167
-
182.
 
Ram
say, J. O., & Dalzell, C. J. (1991). Some tools for functional data analysis. 
Journal of the 
Royal Statistical Society. Series B (Methodological)
, 539
-
572.
 
Ramsay, J. O., Hooker, G., & Graves, S. (2009). Exploring Variation: Functional Principal and 
Canoni
cal Components Analysis. In
 
Functional Data Analysis with R and MATLAB
 
(pp. 
99
-
115). Springer New York.
 
Ramsay, J. O., & Silverman, B. W. 
(
2005
)

ed.
 
Rice, R. E., & Aydin, C. (1991). Attitudes toward n
ew organizational technology: Network 
proximity as a mechanism for social information processing.
 
Administrative science 
quarterly
, 219
-
244.
 
Ray, S., Kim, S. S., & Morris, J. G. (2014). The central role of engagement in online 
communities.
 
Information Syst
ems Research
,
 
25
(3), 528
-
546.
 
Ren, Y., Harper, F. M., Drenner, S., Terveen, L. G., Kiesler, S. B., Riedl, J., & Kraut, R. E. (2012). 
Building Member Attachment in Online Communities: Applying Theories of Group 
Identity and Interpersonal Bonds.
 
Mis Quarterl
y
,
36
(3), 841
-
864.
 
Reichheld, F. F. (2003). The one number you need to grow.
 
Harvard business review
,
 
81
(12), 46
-
55.
 
121
 
 
Rosen, E. (2009).
 
The anatomy of buzz revisited: Real
-
life lessons in word
-
of
-
mouth marketing
. 
Crown Business.
 
Rosenkopf, L., & Abrahamson, 
E. (1999). Modeling reputational and informational influences in 
threshold models of bandwagon innovation diffusion.
Computational & Mathematical 
Organization Theory
,
 
5
(4), 361
-
384.
 
Salancik, G. R., & Pfeffer, J. (1978). A social information processing appr
oach to job attitudes and 
task design.
 
Administrative science quarterly
, 224
-
253.
 
Saxton, G. D., & Anker, A. E. (2013). The aggregate effects of decentralized knowledge 
production: Financial bloggers and information asymmetries in the stock market.
 
Journal
 
of Communication
,
 
63
(6), 1054
-
1069.
 
Sanders, W. M., & Boivie, S. (2004). Sorting things out: Valuation of new firms in uncertain 
markets.
 
Strategic Management Journal
,
 
25
(2), 167
-
186.
 
Sapienza, H. J. (1992). When do venture capitalists add value?.
 
Journal
 
of Business 
Venturing
,
 
7
(1), 9
-
27.
 
Schindler, R. M., & Bickart, B. (2005). Published word of mouth: Referable, consumer
-
generated 
information on the Internet.
 
Online consumer psychology: Understanding and influencing 
consumer behavior in the virtual world
, 35
-
61.
 
Schwienbacher, A., & Larralde, B.
 
(2010). Crowdfunding of Small Entrepreneurial Ventures. 
Handbook of Entrepreneurial Finance, Oxford University Press, Forthcoming. 
 
Siegel, F. W., & Hoban, J. P. (1982). Relative risk aversion revisited.
 
The 
Review of Economics 
and Statistics
, 481
-
487.
 
Singh, P. V. (2010). The small
-
world effect: The influence of macro
-
level properties of developer 
collaboration networks on open
-
source project success.
 
ACM Transactions on Software 
Engineering and Methodology (
TOSEM)
,
 
20
(2), 6.
 
Singh, P. V., Tan, Y., & Mookerjee, V. (2011). Network effects: The influence of structural capital 
on open source project success.
 
MIS Quarterly
,
 
35
(4), 813
-
829.
 
Shane, S., & Cable, D. (2002). Network ties, reputation, and the financing 
of new 
ventures.
 
Management Science
,
 
48
(3), 364
-
381.
 
Shane, S., & Stuart, T. (2002). Organizational endowments and the performance of university start
-
ups.
 
Management science
,
 
48
(1), 154
-
170.
 
Shane, S.,
 
& 
Venkataraman, S. 
(
2000
)
. The promise of entrepreneu
rship as a field of research
.
  
Academy of management review
,
 
(
25
:1), pp.217
-
226.
 
Salancik, G. R., & Pfeffer, J. (1978). A social information processing approach to job attitudes and 
task design.
 
Administrative science quarterly
, 224
-
253.
 
122
 
 
Sharfman, M. P., 
Wolf, G., Chase, R. B., & Tansik, D. A. (1988). Antecedents of organizational 
slack.
 
Academy of Management Review
,
 
13
(4), 601
-
614.
 
Shepherd, D. A., Douglas, E. J., & Shanley, M. (2000). New venture survival: Ignorance, external 
shocks, and risk reduction s
trategies.
 
Journal of Business Venturing
,
 
15
(5), 393
-
410.
 
Spence, M. (1973). Job market signaling.
 
The quarterly journal of Economics
, 
87
(3), 355
-
374.
 
Schooley, D. K., & Worden, D. D. (1999). Investors' asset allocations versus life
-
cycle 
funds.
 
Financial 
Analysts Journal
,
 
55
(5), 37
-
43.
 
Shiller, R. J., & Pound, J. (1989). Survey evidence on diffusion of interest and information among 
investors.
 
Journal of Economic Behavior & Organization
,
12
(1), 47
-
66.
 
Sood, A., James, G. M., & Tellis, G. J. (2009). Function
al regression: A new model for predicting 
market penetration of new products.
 
Marketing Science
,
28
(1), 36
-
51.
 
Stuart, T., Hoang, H., & Hybels, R. (1999). Interorganizational endorsements and the performance 
of 
 
entrepreneurial ventures. 
Administrative scie
nce quarterly
,
 
44
(2), 315
-
 
349.
 
Stuart, R. W., & Abetti, P. A. (1990). Impact of entrepreneurial and management experience on 
early performance.
 
Journal of business venturing
,
5
(3), 151
-
162.
 
Suchman, M. (1995). Managing legitimacy: Strategic and institution
al approaches.
 
Academy of 
 
management review
,
 
20
(3), 571
-
610.
 


 
NYUL Rev.
,
 
80
,  pp. 962.
 
Susarla, A., Oh, J. H., & Tan, Y. (2012). Social networks and the dif
fusion of user
-
generated 
content: Evidence from YouTube.
 
Information Systems Research
,
 
23
(1), 23
-
41.
 
Taboada, M., Brooke, J., Tofiloski, M., Voll, K., & Stede, M. (2011). Lexicon
-
based methods for 
sentiment analysis.
 
Computational linguistics
,
 
37
(2), 267
-
3
07.
 
Tatikonda, M. V., & Rosenthal, S. R. (2000). Technology novelty, project complexity, and product 
development project execution success: a deeper look at task uncertainty in product 
innovation.
 
Engineering Management, IEEE Transactions on
,
 
47
(1), 74
-
87.
 
Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic 
management.
 
Strategic management journal
,
 
18
(7), 509
-
533.
 
Teo, H. H., Wei, K. K., & Benbasat, I. (2003). Predicting intention to adopt interorganizational 
linkages: An instit
utional perspective.
 
Mis Quarterly
, 19
-
49.
 
Tyszka, T., Zielonka, P., Dacey, R., & Sawicki, P. (2008). Perception of randomness and predicting 
uncertain events.
 
Thinking & Reasoning
,
 
14
(1), 83
-
110.
 
Uzzi, B. (1997). Social structure and competition in interf
irm networks: The paradox of 
123
 
 
embeddedness.
 
Administrative science quarterly
, 35
-
67.
 
Venkataraman, S. (1997). The distinctive domain of entrepreneurship research. 
Advances in   
 
Entrepreneurship, Firm Emergence and Growth
,
 
3 J. Katz, ed., JAI Press, 
119
-
138
.
 
 
Wallace, L., Keil, M., & Rai, A. (2004). How software project risk affects project performance: 
An investigation of the dimensions of risk and an exploratory model*.
 
Decision 
Sciences
,
 
35
(2), 289
-
321.
 
Wasko, M. M., & Faraj, S. (2005). Why should I share
? Examining social capital and knowledge 
contribution in electronic networks of practice.
 
MIS quarterly
, 35
-
57.
 
Wasserman, S., & Faust, K. (1994).
 
Social network analysis: Methods and applications
 
(Vol. 8). 
Cambridge university press.
 
Wheelwright, S. C., &
 
Clark, K. B. (1992).
 
Creating project plans to focus product development
. 
Harvard Business School Pub..
 
Xiong, G., & Bharadwaj, S. (2014). Prerelease buzz evolution patterns and new product 
performance.
 
Marketing Science
,
 
33
(3), 401
-
421.
 
Zhang, J., Wedel,
 
M., & Pieters, R. (2009). Sales effects of attention to feature advertisements: a 
Bayesian mediation analysis. Journal of Marketing Research, 46(5), 669
-
681.
 
Zheng, Y., Liu, J., & George, G. (2010). The dynamic impact of innovative capability and inter
-
fi
rm network on firm valuation: A longitudinal study of biotechnology start
-
ups.
 
Journal 
of Business Venturing
,
 
25
(6), 593
-
609.
 
Zimmerman, M. A., & Zeitz, G. J. (2002). Beyond survival: Achieving new venture growth by     
 
 
building legitimacy.
 
Academy of Ma
nagement Review
, 27(3), 414
-
431. 
 
Zott, C, and Huy, Q (2007). "How entrepreneurs use symbolic management
 
to acquire 
resources."
 
Administrative Science Quarterly
 
52.1 (2007): 70
-
105.
 
http://
money.cnn.com/2012/12/18/technology/.../kickstarter
-
ship.../index.h
tml
 
https://www.kickstarter.com/blog/kickstarter
-
stats