ArticlePDF Available

Show me the money! An analysis of project updates during crowdfunding campaigns

Authors:
  • IBM Research, Almaden

Abstract and Figures

Hundreds of thousands of crowdfunding campaigns have been launched, but more than half of them have failed. To better understand the factors affecting campaign outcomes, this paper targets the content and usage patterns of project updates -- communications intended to keep potential funders aware of a campaign's progress. We analyzed the content and usage patterns of a large corpus of project updates on Kickstarter, one of the largest crowdfunding platforms. Using semantic analysis techniques, we derived a taxonomy of the types of project updates created during campaigns, and found discrepancies between the design intent of a project update and the various uses in practice (e.g. social promotion). The analysis also showed that specific uses of updates had stronger associations with campaign success than the project's description. Design implications were formulated from the results to help designers better support various uses of updates in crowdfunding campaigns.
Content may be subject to copyright.
Show Me the Money! An Analysis of Project Updates
during Crowdfunding Campaigns
Anbang Xu1, Xiao Yang2, Huaming Rao3, Wai-Tat Fu1, Shih-Wen Huang4, Brian P. Bailey1
1 University of Illinois, Urbana, IL, USA {xu26, wfu, bpbailey}@illinois.edu
2 Tsinghua University, Beijing, China y-x10@mails.tsinghua.edu.cn
3 Nanjing University of Science and Technology, Nanjing, China huaming.rao@gmail.com
4 University of Washington, Seattle, WA, USA wenhuang@cs.washington.edu
ABSTRACT
Hundreds of thousands of crowdfunding campaigns have
been launched, but more than half of them have failed. To
better understand the factors affecting campaign outcomes,
this paper targets the content and usage patterns of project
updates–communications intended to keep potential funders
aware of a campaign’s progress. We analyzed the content
and usage patterns of a large corpus of project updates on
Kickstarter, one of the largest crowdfunding platforms.
Using semantic analysis techniques, we derived a taxonomy
of the types of project updates created during campaigns,
and found discrepancies between the design intent of a
project update and the various uses in practice (e.g. social
promotion). The analysis also showed that specific uses of
updates had stronger associations with campaign success
than the project’s description. Design implications were
formulated from the results to help designers better support
various uses of updates in crowdfunding campaigns.
Author Keywords
Crowdfunding; updates; crowdsourcing.
ACM Classification Keywords
H.5.3 [Information Interfaces and Presentation]: Group and
Organization Interfaces - Web-based interaction.
INTRODUCTION
Crowdfunding offers a new paradigm for entrepreneurs to
initiate, expand, or advertise their business ideas [10, 25].
While the concept of crowdfunding is still nascent [4], it
has already shown immense promise. At the time of this
writing, for example, Kickstarter, the largest online
crowdfunding platform, has successfully funded 48,393
campaigns. These campaigns have generated 782 million
US dollars from more than 4.7 million people [36]. Many
campaigns have succeeded in reaching their funding goals,
however, more than half of the campaigns have failed [36].
A key challenge is therefore to understand why some
campaigns succeed while others fail. Prior work has
highlighted the relation between the project representation
and the outcome of a campaign [12, 23, 35]. The results
suggest that project creators should focus on improving the
project representation, which mainly includes the video and
textual description on the project page (see Figure 2a). For
example, the top rule for success suggested by Kickstarter
is to create a video for the project page [35].
Besides careful preparation of a project’s representation,
creating updates is also an important part of managing a
campaign [13]. The form of a project update is similar to a
blog post (see Figure 2b) and the design intent is to “keep
backers (funders) informed of a project's progress” [34].
Updates are critical to the success of a campaign. For
example, we sampled 8,529 campaigns from Kickstarter
and found that the chance of success of a project without an
update was only 32.6%. In contrast, as shown in Figure 1,
the chance of success with updates is 58.7% (χ2 = 285.18,
p< .001). This suggests that updates may be as important as
the creation of the project representation in determining the
outcome of a campaign. However, prior work has not
examined the types of updates created in a campaign, the
distribution of updates across categories or time, or how
different types of updates relate to the campaign outcomes.
In this paper, to better understand the nature of project
updates, we analyzed how creators use updates during
crowdfunding campaigns and how these updates relate to
the success of the campaigns. Our main contributions are:
Figure 1. The success rates of campaigns with updates and
campaigns without updates.
Permission to make digital or hard copies of all or part of this work for personal or
classroom use is granted without fee provided that copies are not made or distributed
for profit or commercial advantage and that copies bear this notice and the full
citation on the first page. Copyrights for components of this work owned by others
than ACM must be honored. Abstracting with credit is permitted. To copy otherwise,
or republish, to post on servers or to redistribute to lists, requires prior specific
permission and/or a fee. Request permissions from permissions@acm.org.
CHI 2014, April 26–May 1, 2014, Toronto, Ontario, Canada.
Copyright ©ACM 978-1-4503-2473-1/14/04...$15.00.
http://dx.doi.org/10.1145/2556288.2557045
Session: Crowdfunding and Crowd Storage
CHI 2014, One of a CHInd, Toronto, ON, Canada
591
An empirically-based taxonomy of the types of project
updates created during a campaign. We identified seven
types of themes in the updates and found differences
between the stated design intent of updates and how
project creators actually use project updates in practice.
We report how the different types of updates relate to
the outcomes of the campaigns. Results revealed the
relative importance of the update representation (i.e. its
content) compared to the project representation (e.g., the
presence of videos and images and description length).
Design implications for crowdfunding tools to better
support the practice of creating and using updates during
a campaign, thereby improving their effectiveness.
RELATED WORK
The factors that lead to successful fundraising have been of
great interest to researchers [5, 19]. Recent studies have
highlighted the importance of the project representation and
have found that many attributes related to the project page
(see Figure 2a) can influence the success of a campaign.
For instance, Kickstarter suggests that the number one rule
of success of a campaign is to have a video on the project
page to communicate the overall project ideas [35]. Prior
work confirms that the outcome of a campaign is related to
the presence of a video [12] and the quality of the video
[23]. Prior work has also revealed that the success of a
campaign is related to the textual content of the project
page such as the length and readability of the content [12],
and certain phrases used in the content [22].
In addition to the project representation, Kickstarter also
provides the ability to create project updates. The intent of
an update is to “keep backers (funders) informed of a
project's progress” [34]. However, creators are free to
provide any information about the project through an
update. Researchers have included project updates in their
regression analysis of campaign outcomes [18, 23]. The
number of updates was found to be positively related to the
success of a campaign and support the recommendation that
creators provide frequent updates [35]. However, because
prior work has treated updates as a single variable, it is not
known what types of updates are created, which ones occur
most often or when they occur during a campaign, or how
the different types of updates relate to campaign success.
The initial settings of a campaign such as the amount of the
funding goal and the duration of the campaign can also
predict its success. For example, Muller et al. [25] found
that the quantitative differentiator between successful and
unsuccessful campaigns is that successful campaigns have a
smaller amount of funding goals than unsuccessful
campaigns. It has also been found that the duration of a
campaign is negatively related to its success [23]. The
social network of project creators is often the initial funding
source of many campaigns and plays an important role in
determining success [1, 23]. For example, Mollick found
that the number of Facebook friends of creators is
positively correlated with the success of campaigns.
Qualitative studies have been conducted to understand the
success of crowdfunding campaigns and the motivation and
barriers to participation [11, 13, 14]. Hui et al. [13]
conducted interviews with project creators and participant
observation to understand the work needed for the creators.
They found that successful creators made large efforts in
reaching out to personal on- and off-line networks for funds
during the campaign. Yet, it is not clear how the degree of
this effort influences the outcome of the campaign.
Our research contributes to this corpus of prior work by
studying a crowdfunding site from a unique perspective –
the use of project updates. To the best of our knowledge,
we are the first to report the types of updates created during
crowdfunding campaigns and to report the relationship
between the different update types and campaign success.
(a) Project page of a campaign (b) One of the project updates of the camapign
Figure 2. (a) The project page is the main page of a campaign on Kickstarter.com. The campaign shown is raising funds for an
iPhone application that can help people find free beaches in Malibu. This campaign has four updates so far. If a user (e.g. funder)
clicks on the “update” tab, s/he will see the updates displayed in reverse chronological order. (b) An update of the campaign posted
in the middle of the campaign. Project creators in this update introduced two new rewards (poster and a beach bag) to attract
additional funders. In this update, project creators also encourage people to promote the campaign in a social network.
Session: Crowdfunding and Crowd Storage
CHI 2014, One of a CHInd, Toronto, ON, Canada
592
RESEARCH QUESTIONS AND DATA SET
Our study was designed to understand the practice of using
updates in crowdfunding campaigns and centered on two
key research questions:
RQ1: What are the usage patterns of updates? What types
of updates occur during campaigns, how often do they
occur, and when does each type of update occur?
RQ2: How are the types of updates associated with the
success of a campaign? How important is the representation
of updates compared with the representation of the main
project in explaining the success of a campaign?
A quantitative analysis was conducted for our study. For
data collection, we used the site thekickbackmachine.com
which lists the project IDs of Kickstarter projects in reverse
chronological order. Using a custom extractor, we first
collected the listed project IDs from thekickbackmachine
and then collected the corresponding content from
Kickstarter. In total, we collected publicly available data
from 8,529 campaigns on Kickstarter. The campaigns
started between March 19, 2013 and May 17, 2013. The
mean duration of a campaign in the data set was 32.1 (days)
with a standard deviation of 10.2. For each campaign, we
collected the updates that were posted before the outcome
of the campaign was determined (successful or
unsuccessful). A majority of the campaigns (58.6%) had at
least one update. Project creators can choose to provide a
private project update that can only be viewed by its
funders. We found 3,098 private updates that we could not
collect, and we only collected the updates that can be
accessed by the public (the potential funders). This
collectively provided us with a corpus of 21,234 updates.
The data set includes the content of each campaign and the
content and timestamp of each update.
ANALYZING UPDATES (RQ1)
To identify the types of themes in the project updates, we
applied Latent Dirichlet Allocation (LDA), an unsupervised
generative method that is often used to discover hidden
themes in documents and the words associated with each
theme. The method enables analysis of large amounts of
unlabeled documents by clustering words that frequently
co-occur. Several steps were performed to process the data.
Step 1: Sample the data. The updates were not distributed
evenly across the projects. For example, some projects had
many updates while others had only a few. If we directly
applied topic modeling on the updates, the identified
themes and the words associated with a theme might be too
specific to the content of some of the projects. To avoid
this, we randomly selected three updates from a project or
all the updates from a project if it had fewer than three.
Step 2: Clean the data. We converted the text to lowercase
and removed the punctuation characters such as ();:‘”.
Step 3: Create “Bag of words”. We adopted a bigram “bag
of words” model, a common approach in computational
linguistics [21]. That is, we used single words (unigrams),
and two word phrases (bigrams) to represent the text.
We first attempted to perform LDA on the preprocessed
text by treating each update as a document. However, we
found that the resulting themes had too much overlap (i.e.
many themes shared the same words), making the
interpretation of the themes difficult. This is a known
limitation of standard LDA. An alternative that has been
applied in the text mining community is to decompose the
document into finer granularities such as the sentence level
in order to detect more specific topics and reduce the
overlap [15] 1. Thus, we applied LDA at the sentence-level,
meaning that each sentence was treated as a document.
Step 4: Decompose the updates into sentences. We
separated all text into sentences based on terminal
punctuations (“.”, “?”, and “!”). We excluded sentences that
had fewer than three unigrams (15% of the sentences were
removed); as these sentences were too short for successful
model training and were usually not too meaningful. By
applying LDA at the sentence-level, we found the themes
easier to interpret because there was less overlap.
Researchers typically examine the output of different
themes in order to decide the number of unique themes
[32]. Following this data-driven approach, two experts
familiar with crowdfunding reviewed the outcomes from
the LDA models. The experts started by fixing a large
number of themes (30 in this case), and reduced the number
if they could find duplicates (e.g. themes described by the
same set of words). The experts were able to finalize seven
unique themes from the LDA results and category labels
were assigned to the themes and any disagreements were
resolved by discussion.
We then created a dictionary based on the results of LDA
and used the dictionary to assign themes to the updates.
Specifically, we constructed dictionaries to represent the
identified themes by selecting the top 60 words (unigrams
and bigrams) that were most strongly associated with each
theme according to the LDA model. We also excluded a
few words that are related to project categories or locations
(e.g. game, music, and New York). This dictionary-based
approach was applied to the entire data set. We classified an
update to belong to a theme if it contained at least two
unigrams or bigrams from the corresponding dictionary. By
this definition, a majority of the updates belonged to only a
single theme.
To verify the reliability of the produced taxonomy, we
recruited two people to code a sample of the updates based
on the taxonomy. First, the coders received training in
which they were introduced to the categories, definitions,
1 Sentence-level LDA works well when most of the sentences
contain only a single theme. This was the case for our data set. <
1% of the sentences in our data set included multiple themes.
Session: Crowdfunding and Crowd Storage
CHI 2014, One of a CHInd, Toronto, ON, Canada
593
and examples in the taxonomy. They then coded updates on
a small sample of the data and resolved disagreements.
Then, they independently coded a random sample of 200
updates. There was good agreement between the coders and
the dictionary-based method derived from the taxonomy
with a Fleiss' kappa of 0.77.
Types of Updates
The themes identified in the updates demonstrated the
creative use of project updates on the crowdfunding site.
The stated design intent of project updates is to help project
creators keep their funders informed on the development of
the project [34]. This intention was confirmed by the
identified theme Progress Report (see Table 1). More
interestingly, six other unexpected uses of updates were
identified: 1) Social Promotion encourages people to
promote a project and spread the word in their social
networks. 2) New Content introduces new project ideas to
the existing content of the project. 3) New Reward provides
new reward levels for a pledge. Reward was something
creators offered people in exchange for their pledges of
funds. 4) Answer Questions from people such as funders
and potential funders. 5) Reminders encourage people to
pledge. Figure 3 shows the percentages of updates in each
of these themes. We will elaborate these themes below.
Social Promotion is the most popular theme of the updates
during the crowdfunding campaigns (23% of the updates).
Project creators frequently encouraged supporters to
promote a project in social media. This pattern revealed the
emerging needs of project creators to gain additional
exposure to their projects through the use of social media.
Category Dictionary words Examples
Social
Promotion
facebook page, twitter, help spread,
tweet, follow facebook, please share, tell
friend, share link, friends, family
“Please tweet or post the link to this update on all your social
network sites to help spread the work about…”
“One of the best ways you can stay involved is to like our Facebook
page.”
Progress
Report
progress report, new update, first update,
last update, half done, get done,
milestone reached, journey continues,
halfway, moving ahead, ahead schedule
“We are more than half done with our remodel windows, a few
paint touch ups, and bakery area left.”
“We have a few things for our first update so let’s get to it. First I
would like to show you …”
New Content new idea, new concept, new project, new
blog, new picture, new cover, new link,
video update, design update, brand new
“We would like to introduce all our backers to this great new
concept David presented us with today.”
“We were also looking forward to revealing some fun new ideas to
try to keep the momentum and enthusiasm going”
Reminder weeks remain, days remain, hour left,
final week, final countdown , crunch time,
clock ticking, last push, weeks counting,
finish line
“We are down to the final week of our campaign. Please take a
moment to pledge …”
“Less than 24 hours to go … one last push and one last reminder.”
Answer
Question
how began, why need, please read, what
happen, when project, when campaign,
answer question, feedback, explain, faq
section
“We added some new things to the FAQ section and explained
different ways to back the project.”
“We will only be online intermittently today to answer questions.”
New Reward new reward, change reward, better
rewards, additional reward, improved
reward, extra reward, incentive bonus,
reward level, pledge level, reward update
“We’re happy to announce two new reward levels with full HD
capability.”
For any contributions of $20 or more, here are [sic] your new
bonus…”
Appreciation thanks support, greatly appreciate,
everyone’s support , humbled, grateful,
excited, generosity, bottom heart, huge
thank, thanks (for) reading
“Thank you all for your support I feel so very blessed and loved”
“We are humbled and extremely appreciative for all your efforts to
help us…”
Table 1. The taxonomy of the types of updates created during crowdfunding campaigns.
Figure 3. Distribution of updates in the seven themes.
Session: Crowdfunding and Crowd Storage
CHI 2014, One of a CHInd, Toronto, ON, Canada
594
Both New Content and New Reward updates indicated that
project creators revised their projects during the campaign.
However, these revisions could be viewed from two
different perspectives. New Content updates emphasized
changes of the content of the project itself, while New
Reward updates indicated that creators offered new rewards
to attract funders. Metaphorically, New Reward can be
viewed as offering discounts on the product to attract
customers, while New Content can be viewed as improving
the product itself to attract customers. Both of the strategies
could influence the sales of the product; however, which
strategies would be more effective in promoting the project
in the context of crowdfunding campaigns? We will attempt
to compare the effects of these strategies when we address
the second research question (RQ2).
Kickstarter has provided a separate section for creators to
answer questions from funders and potential funders.
However, the Answer Questions theme in the updates
indicated that many creators still use project updates as a
platform to answers questions and explain their projects.
Timing of Updates
We investigated the temporal use of updates during the
campaigns. We evenly divided the duration of a campaign
into three intervals (phases): initial, middle, and final phase
and an update of the campaign can be assigned to one of
these phases based the posted time of the update. Figure 4
shows the distributions of updates in the three phases.
Interestingly, the distributions of the New Content and New
Reward updates in the three phases were significantly
different from a uniform distribution (χ2 = 285.18, p<.001;
χ2 = 70.10, p<.001). The number of New Content updates in
the initial phase was higher than the number of updates in
the middle and final phases, indicating that creators
adjusted their project content earlier rather than later during
a campaign. In contrast, the number of New Reward
updates increased in the final phase. These usage patterns
indicate that project creators initially focus on revising their
project content, but shift attention to adjusting rewards in
the final phase.
Similar to New Reward updates, the number of Social
Promotion, Appreciation, and Reminder updates were all
increased in the final phase compared to the middle phase
(p<.001). One interpretation of the results is that in the final
phase of campaigns, project creators were more likely to
utilize these types of updates to help them reach their
funding goal. In particular, the Reminder updates were
rarely used in the initial phase of the campaigns and they
were the least popular among all types of updates.
Conversely, Reminder updates were the second most
popular in the final phase and project creators were on
average three times more likely to use Reminder updates in
the final phase than in the initial phase.
In addition, both Figures 3 and 4 show that successful
campaigns had more updates than unsuccessful campaigns
across different update themes and different time phases. In
the next section, we will examine how updates were related
to the outcomes of the campaigns.
ASSOCIATION BETWEEN UPDATES & SUCCESS (RQ2)
To address the second research question, we used
hierarchical logistic regression to produce an analytical
model for the campaign outcomes. Hierarchical logistic
regression allows us to divide the predictor variables into
blocks such as update types, update representation, and
project representation, and compare the relative importance
of the blocks. Specifically, the regression starts with no
blocks in the model, tests the addition of each block using
Nagelkerke R², adds the block that improves the model the
most, and repeats this process for the remaining blocks.
This allowed us to study which blocks were more important
for the success of a campaign.
For the dependent variable we gave successful campaigns a
value of 1 and failed campaigns a value of 0. 4081 projects
were successfully funded, while 4448 projects were not.
The following explanatory variables related to updates and
projects were investigated. They were divided into one
block of control variables and the four analytical blocks:
Update Theme. This block included the ratios of the
update themes and the number of updates. For each
Figure 4. Distribution of updates in three phases. T1, T2, T3 refer to the initial, middle and final thirds of a campaign respectively.
Session: Crowdfunding and Crowd Storage
CHI 2014, One of a CHInd, Toronto, ON, Canada
595
campaign, we first applied the dictionaries to all of its
updates (an update belongs to a theme if it contained at
least two unigrams or bigrams from the corresponding
dictionary), counted the number of updates in each
update theme, and then computed the ratio of the number
of updates in each theme to the total number of updates.
For example, Progress Report in table 2 refers to the
ratio of Progress Report updates in a project to the total
number of updates in that same project.
Project Representation. Following prior work [12], we
measure the representation of a project by the attributes
related to the content of the project page: 1) Title Length
is the number of words in the project title. 2) Body
Length is the number of words in the project description.
3) Number of URLs is the number of URLs referenced in
the project page. 4) Number of Images is the number of
images included on the project page. 5) Number of videos
is the number of videos included on the project page. 6)
Readability of the project description is measured by the
Flesch ease of reading score. These scores usually range
between 0 and 100.
Update Time. When we computed Update Theme block,
we had the updates in each theme for a campaign. We
further divided the updates in a theme into three phases:
initial (T1), middle (T2), and final (T3). We counted the
number of updates in each phase of a campaign and
computed the ratio of the number of updates in each
phase to the total number of updates in the theme.
In addition, we entered the control variables including
campaign category, campaign duration, and the funding
goal into the regression model as the initial block [23].
Except for the initial block of the control variables, each
analytic block of variables was incrementally added to the
regression model. The blocks were added to the model one
at a time, using the statistical criterion of maximizing the
Nagelkerke R² of the included blocks. For example, of the
remaining blocks, the next block added was the one that
maximized the prediction ability over the preceding model.
This process was repeated until all blocks were added.
Table 2 shows the parameters for the logistic regression and
the final analytical model of campaign outcomes based on
the procedure described above. The control variables in the
initial block was consistent with prior work [23]. Given that
we focus on project updates, we only report their effects
over and above the control variables for the sake of
simplicity. B is the estimated coefficient for each variable in
the model’s equation. If the Wald statistic is significant (p <
.05) then the parameter is useful to the model. Overall, the
final logistic regression model correctly classified 77.5% of
the campaigns, compared to the correct classification of
51.5% of campaigns in the base model. We should point out
that we do not intend to present our model as a ready-to-use
solution for a prediction system (more analysis on causality
is needed), but the accuracy indicates that our model
provided a reasonable fit to the data (True Positive Rate is
74.1%, True Negative Rate is 80.6%), providing validity to
our analysis on the association between updates and the
outcomes of the campaigns.
Importance of Update Themes
The Update Theme had the most predictive power in
predicting the outcome of campaigns. All the independent
variables in this block were positively correlated to the
success of a campaign (p < .001). Among the seven types of
updates, we found that Reminder updates offered the most
significant influence within our model (B = 2.000, p <
.001), followed closely by Progress Report (B = 1.818, p <
.001), New Reward (B = 1.690, p < .001), and Social
Promotion (B = 1.528, p < .001). The effects of Reminder
updates reflect the power of the ask in traditional charity
fundraising activities [2, 3] and social media systems [33].
Surprisingly, Answer Questions updates had the least
influence, though it was still predictive (B = 0.711, p <
.001). One explanation is that updates dedicated to
answering questions might have conflicting effects on
campaign outcomes. On the one hand, these updates can
help people better understand a campaign. On the other
hand, these updates reflect that people have difficulties in
understanding or appreciating the campaign.
Another interesting finding was that New Reward updates
(B = 1.690, p<.001) were more likely to increase the chance
of success of a campaign than New Content updates (B =
1.187, p<.001) (p<.001 for the difference, t-test). However,
as shown in Figure 3, there were more New Content updates
than New Reward updates during the campaigns (p<.001 for
the difference). These results indicate that project creators
currently spend more effort revising their project content
than revising their reward levels. Yet, after a campaign is
launched, the analysis shows that revising reward levels is a
more effective strategy than revising project content for
achieving campaign success.
Update Representation vs. Project Representation
Interestingly, we found that Update Representation was
more predictive of the campaign success than Project
Representation. We tested this by the following procedure
(not shown in Table 2): After the Update Theme was
entered into the model as the first block, both Update
Representation and Project Representation were entered
into the model as the second block, and both significantly
improved the regression model (p < .001 in each case).
However, adding Update Representation ( = 0.49, =
0.11) increased the predictive ability of the model more
than adding Project Representation ( = 0.41, = 0.03).
Though prior work has demonstrated the importance of the
project representation [12, 23], our results reveal that the
representation of updates serve an even larger role during a
campaign.
Session: Crowdfunding and Crowd Storage
CHI 2014, One of a CHInd, Toronto, ON, Canada
596
B SE p Mean STD Distribution
Update Theme
R2 = 0.38
= 0.25
Social Promotion 1.528 0.111 <.001 0.107 0.220
Progress Report 1.818 0.128 <.001 0.086 0.196
New Content 1.187 0.151 <.001 0.055 0.155
Reminder 2.000 0.171 <.001 0.052 0.150
Answer Questions 0.711 0.160 <.001 0.049 0.145
New Reward 1.690 0.172 <.001 0.046 0.142
Appreciation 1.201 0.187 <.001 0.031 0.123
Num. of Updates 0.440 0.015 <.001 2.489 6.000
Update
Representation
R2 = 0.49
= 0.11
Title Length 0.331 0.016 <.001 2.999 2.673
Body Length 0.008 0.001 <.001 72.05 109.7
Num. of URLs -0.248 0.054 <.001 0.302 0.683
Num. of Images 0.010 0.044 0.81 0.116 1.207
Num. of Videos -1.522 2.836 0.59 0.000 0.012
Readability -0.059 0.009 <.001 3.995 7.313
Project
Representation
R2 = 0.53
= 0.04
Title Length -0.016 0.012 0.184 5.694 2.458
Body Length -0.001 0.000 <.001 910.3 703.8
Num. of URLs 0.077 0.007 <.001 10.77 7.248
Num. of Images -0.012 0.004 <.05 6.513 9.597
Num. of Videos 0.061 0.026 <.05 1.258 1.146
Readability -0.047 0.012 <.001 9.496 2.570
Update Time
R2 = 0.54
= 0.01
Social Promotion
T1 0.578 0.265 <.05 0.045 0.174
T2 0.015 0.262 0.953 0.041 0.164
T3 0.456 0.255 0.074 0.049 0.186
Progress Report
T1 -0.075 0.147 0.611 0.108 0.288
T2 0.583 0.171 <.001 0.071 0.228
T3 -0.431 0.167 0.05 0.063 0.219
New Content
T1 -0.163 0.239 0.496 0.029 0.151
T2 -0.102 0.243 0.674 0.030 0.153
T3 0.082 0.269 0.760 0.029 0.150
Reminder
T1 0.400 0.371 0.281 0.031 0.163
T2 0.228 0.356 0.522 0.049 0.198
T3 0.086 0.344 0.802 0.096 0.280
Answer Question
T1 -0.452 0.256 0.078 0.028 0.147
T2 -0.198 0.256 0.439 0.026 0.144
T3 0.208 0.260 0.423 0.028 0.150
New Reward
T1 0.231 0.260 0.373 0.023 0.137
T2 0.536 0.282 0.058 0.024 0.137
T3 1.180 0.267 <.001 0.032 0.164
Appreciation
T1 0.079 0.290 0.785 0.017 0.121
T2 -0.121 0.292 0.679 0.015 0.114
T3 -0.277 0.238 0.244 0.022 0.140
Table 2. Hierarchical logistic regression results for predicting the success of a campaign (Left columns). Nagelkerke R2 was reported with each
block and the contribution of each block was statistically significant (p<.001). Right columns show the descriptive statistics of the predictors.
Session: Crowdfunding and Crowd Storage
CHI 2014, One of a CHInd, Toronto, ON, Canada
597
The success of campaigns was positively correlated with
the length of the update content (title and body), while it
was negatively correlated with the length of project page
content. Also, the average length of the updates and project
page is about 70 words and 910 words respectively (see
Table 2). One interpretation is that updates of many
unsuccessful campaigns were too short to provide sufficient
information for potential funders, while the project page of
many unsuccessful campaigns were too lengthy for
potential funders. Similarly, URLs in the updates and
project page had the opposite effect on campaign outcomes.
The readability of both updates and project page had an
overall effect of decreasing the likelihood of the campaign
being successful. This indicated that funders appreciated
sophisticated content rather than simple or naive content.
Also, the effect of the presence of a video in the project
representation was consistent with prior work [12, 23].
Timing of Updates
The ratio of the number of Social Promotion updates in the
initial phase to the total number of Social Promotion
updates during a campaign is positively correlated to the
success of the campaign (B = 0.578, p < .05). This positive
correlation suggests that if a project creator wants to
socially promote their project in updates, more promotion
in the initial phase is more likely to increase the chance of
success.
The high ratio of Progress Report updates in the middle
phase increased a campaign’s chance of success (B = 0.583,
p<.001). However, Figure 4 shows that the number of
Progress Report updates decreased from the initial phase to
the middle phase. Specifically, the ratios of Progress
Report updates in the initial and middle phases to the total
number of Progress Report updates were 45.6% and 30.1%
respectively. One explanation of these results is that many
campaigns provided progress reports in the initial phase and
this may have increased funders’ expectations for progress
reports in the later phases of the campaign.
The ratio of New Reward updates in the final phase had a
positive correlation with the probability of campaign
success (B = 1.180, p<.001). Metaphorically speaking, if we
think of adding new reward levels during a campaign as
changing the price of a product to increase sales, discounts
in the final phase are more effective than discounts in the
initial phase for crowdfunding campaigns.
In addition, we found that the posted time of Reminder,
New Content, Answer Question, and Appreciation updates
did not have significant effects on the success of a
campaign. This indicates that the effects of these types of
updates are not related to the time of posting.
DESIGN IMPLICATIONS
We foresee many opportunities for systems to better
support a campaign during crowdfunding activities based
on the findings from our data analysis.
Support Various Uses of Updates
Seven different uses of updates were identified in our study,
and a majority of the updates were not used to report
project progress. One explanation is that the platform
studied does not offer appropriate tools for project creators
to perform the desired activities (e.g. social promotion) and
the project creators had to compensate by using updates. In
addition, the update types are mixed together and displayed
in reverse chronological order on the platform. It may
therefore be difficult for project creators to highlight certain
types of updates or convey specific messages effectively.
System designers may consider the various uses of updates
and design tools to support these uses accordingly [29]. For
example, in order to help project creators communicate a
message more effectively during crowdfunding campaigns,
future systems can allow the creators to assign tags to their
updates, and the system can map different visual attributes
such as color schemes to each type of update. Also, systems
could provide templates for and examples of each type of
project update, creating awareness of the types of updates
available and making it easier to create effective updates.
Improve Update Representation
In our dataset, we found that the update representation was
more important than the project representation in predicting
the success of a campaign. Although project representation
has received considerable attention from the research
community, relatively little attention has been directed at
improving the representation of project updates. One way
crowdfunding platforms can help is to allow creators to
learn from prior successful examples. For instance, many
crowdfunding sites [37] provide a comprehensive taxonomy
to classify various prior projects, and project creators can
easily navigate the projects and find relevant examples [17].
This usage echoes research findings in education and
cognitive psychology, such as the results of LeFevre and
Dixon [20] who found that examples are important in the
learning process and example-driven approaches are often
more effective than instructions without examples.
However, current platforms do not offer an efficient way
for project creators to learn and improve the presentation of
their updates. One solution is to provide specific guidelines
such as encouraging project creators to be more specific
and detailed in their updates [16]. Another solution is to
help creators find useful examples to follow. For instance,
based on the categorized list of update themes and the
associated words, future systems can offer a sitemap or an
index tool to help users navigate through the updates and
the corresponding projects, and find relevant examples from
which to base the creation of their own updates.
Connect Updates with Social Media
Our results showed that Social Promotion updates were the
largest proportion of all update activities. This highlights
the importance of social media for promoting crowdfunding
projects. Prior work has also recognized the increasingly
important role of social media for businesses [5, 13, 24, 27].
Session: Crowdfunding and Crowd Storage
CHI 2014, One of a CHInd, Toronto, ON, Canada
598
For crowdfunding, many sites allow project creators to
embed a link of their social media profiles (e.g. Facebook
page) to their project page and updates. Beyond supporting
links to social media content, systems should help creators
better harness the speed and reach of social media platforms
to more effectively promote their campaigns. For example,
tools could be designed to help creators develop and
execute strategies for advertising their campaigns via social
media (e.g. what to say, when to say it, and to whom).
New Reward and New Content
New Reward updates were more predictive than New
Content updates for the outcome of campaigns. This does
not mean that project creators do not need to revise their
product ideas, but rather that creators also need to pay more
attention to the adjustment of their rewards during the
campaign. Based on literature in cognitive science and
business strategies [8, 26, 30], the relatively high predictive
power of New Reward updates is likely to occur in low-
involvement choices among a large number of alternatives.
In other words, when competition increases, the price
difference (the reward that people receive) becomes a
differentiator in the market. This phenomenon reflects the
highly competitive nature of crowdfunding campaigns.
We also found that the high ratio of New Reward updates in
the final phase of a campaign is positively correlated to the
success of the campaign. One possible explanation is that
the initial reward offered by the campaign may tend to
serve as a contrasting reference point and the additional
rewards change funders or potential funders’ perceptions on
the campaign and thus affect their pledge decisions [9, 28].
Providing attractive rewards during a campaign is a
challenging task and how systems can offer assistance
constitutes an interesting area for further research. For
example, a system can generate aids based on the content of
project pages and prior updates to help creators brainstorm
diverse ideas about rewards [31].
Prompt for an Update
Creating updates during a crowdfunding campaign can
increase its chance of success; however, a significant
portion of the campaigns did not have any updates. System
designers may therefore consider adding prompts for
project creators to create updates. Also, our results revealed
that the effects of Social Promotion, Progress Report, and
New Reward updates were more significant during certain
phases of the campaign. Systems could incorporate these
findings into the design of the prompts to guide project
creators as to when to post an update to maximize its effect.
Frequent updates used to solicit support from funders
reflect frequent communication used in social media to
bridge social capital or gain other benefits [7]. Future work
should examine how different communication strategies in
updates (beyond its theme) affects support from potential
funders, similar to how the language used in online posts
affects the likelihood of a reply [6].
LIMITATION
This is a quantitative study based on data collected from a
single crowdfunding platform. Our approach is useful for
describing what happens, but a more comprehensive study
is needed to increase the generalizability of the results and
more systematically reveal the causal relations between the
various dynamic events that occur during the campaign and
the campaigns’ success. Qualitative studies can be applied
to understand the expectations and perceptions of the
updates from both creators and funders’ perspectives. On
the other hand, although we only studied one crowdfunding
platform, the results of our work may be used to guide and
compare the analysis of campaigns on other sites [38, 39].
Such comparisons can lead to more generalizable
knowledge, which can be applied to further improve the
effectiveness of crowdfunding.
CONCLUSION
Project updates are an important part of managing a
campaign on a crowdfunding platform, but how updates are
leveraged in practice and how they relate to the success of
the campaign is unknown. This paper has made three
contributions to closing this knowledge gap. First, we
identified seven types of project updates made during
crowdfunding campaigns and found differences between
the stated design intent and the actual uses of updates (e.g.
for social promotion). Second, we reported the statistical
relations between the different types of updates and the
outcomes of the campaigns. One significant result was that
how project creators communicate with potential funders
during a campaign is more predictive of success than the
representation of the project page. Finally, the results were
formulated into design implications for improving
crowdfunding platforms and tools. Most importantly,
designers should consider the functionality of project
updates and how to better support their various uses in a
campaign. The outcomes of this work can help creators
better manage their crowdfunding campaigns and lead to
better tools for guiding and reducing the effort.
REFERENCES
1. Agrawal, A., Catalini, C., & Goldfarb, A. The
Geography of Crowdfunding. NBER Working Paper,
16820, (2011).
2. Andreoni, J. Philanthropy. In Kolm, S.-C. and Ythier,
J.M. eds. Handbook of Giving, Reciprocity and
Altruism, North Holland, 2006, 1201-1269.
3. Andreoni, J. and Payne, A.A. Government Grants to
Private Charities: Do They Crowd Out Giving or
Fundraising? American Economic Review, 93, 3, (2003),
792–812.
4. Belleflamme, P., Lambert, T., & Schwienbacher, A.,
Crowdfunding: Tapping the Right Crowd. In AFFI,
(2011).
5. Besley, T. and Coates, S. Group Lending, Repayment
Incentives And Social Collateral. Journal of
Development Economics, 46, 1, (1995), 1-18.
Session: Crowdfunding and Crowd Storage
CHI 2014, One of a CHInd, Toronto, ON, Canada
599
6. Burke, M., Joyce, E., Kim, T., Anand, V., & Kraut, R.,
Introductions and Requests: Rhetorical Strategies That
Elicit Response in Online Communities In Communities
and Technologies, (2007), 21-39.
7. Burke, M. and Kraut, R., Using Facebook after Losing a
Job: Differential Benefits of Strong and Weak Ties. In
CSCW, (2013), 1419-1430.
8. Campbell, J.Y., Lo, A.W., & MacKinlay, A.C. The
Econometrics of Financial Markets. Princeton
University Press, 1997.
9. Clark, E.M., Brock, T.C., & Stewart, D.W. Attention,
Attitude, and Affect in Response to Advertising.
Psychology Press, 1993.
10. Gerber, E. and Hui, J.S. Motivations and Deterrents for
Participation. TOCHI, (In Press).
11. Gerber, E.M., Hui, J.S., & Kuo, P.-Y., Crowdfunding:
Why People Are Motivated to Post and Fund Projects on
Crowdfunding Platforms. In CSCW (Workshop), (2012).
12. Greenberg, M.D., Pardo, B., Hariharan, K., & Gerber,
E., Crowdfunding Support Tools: Predicting Success &
Failure. In CHI Extended Abstracts, (2013), 1815-1820.
13. Hui, J.S. and Gerber, E., Easy Money? The Demands of
Crowdfunding Work. In Segal Technical Report: 12 -
04, (2012).
14. Hui, J.S., Greenberg, M.D., & Gerber, E.M.,
Understanding the Role of Community in Crowdfunding
Work. In CSCW (2014).
15. Jo, Y. and Oh, A., Aspect and Sentiment Unification
Model for Online Review Analysis. In WSDM, (2011),
815-824.
16. Kittur, A., Nickerson, J.V., Bernstein, M.S., Gerber,
E.M., Shaw, A., Zimmerman, J., Lease, M., & Horto,
J.J., The Future of Crowd Work. In CSCW, (2013),
1301-1318.
17. Kuo, P.-Y. and Gerber, E., Design Principles:
Crowdfunding as a Creativity Support Tool. In CHI
Extended Abstracts, (2012).
18. Kuppuswamy, V. and Bayus, B.L. Crowdfunding
Creative Ideas: the Dynamics of Projects Backers in
Kickstarter. SSRN Electronic Journal, (2013).
19. Larrimore, L., Jiang, L., Larrimore, J., Markowitz, D., &
Gorski, S. Peer to Peer Lending: The Relationship
Between Language Features, Trustworthiness, and
Persuasion Succes. Journal of Applied Communication
Research, 39, 1, (2011), 19-37.
20. LeFevre, J.-A. and Dixon, P. Do Written Instructions
Need Examples? Cognition and Instruction, 3, 1,
(1986), 1-30.
21. Manning, C.D. and Schütze, H. Foundations of
Statistical Natural Language Processing. MIT press,
1999.
22. Mitra, T. and Gilbert, E., The Language that Gets
People to Give: Phrases that Predict Success on
Kickstarter. In CSCW, (2014).
23. Mollick, E. The Dynamics of Crowdfunding: An
Exploratory Study. Journal of Business Venturing, 29, 1,
(2013), 1-16.
24. Muller, M., Ehrlich, K., Matthews, T., Perer, A., Ronen,
I., & Guy, I., Diversity among Enterprise Online
Communities: Collaborating, Teaming, and Innovating
through Social Media. In CHI, (2012), 2815-2824.
25. Muller, M., Geyer, W., Soule, T., Daniels, S., & Cheng,
L.-T., Crowdfunding inside the Enterprise: Employee-
Initiatives for Innovation and Collaboration In CHI,
(2013), 503-512.
26. Puccinelli, N.M., Goodstein, R.C., Grewal, D., Price, R.,
Raghubir, P., & Stewart, D. Customer Experience
Management in Retailing: Understanding the Buying
Process. Journal of Retailing, 85, 1, (2009), 15-30.
27. Qualman, E. Socialnomics: How Social Media
Transforms The Way We Live and Do Business. John
Wiley, 1994.
28. Schindler, R.M. Consequences of Perceiving Oneself as
Responsible for Obtaining a Discount: Evidence for
Smart-Shopper Feelings. Journal of Consumer
Psychology, 7, 4, (1998), 371-392.
29. Shneiderman, B. Designing The User Interface:
Strategies for Effective Human-Computer. Addison-
Wesle Reading, 1998.
30. Spence, M. Cost Reduction, Competition, and Industry
Performance. Journal of the Econometric Society, 52, 1,
(1984), 101-122.
31. Wang, H.C., Cosley, D., & Fussell, S.R., Idea Expander:
Supporting Group Brainstorming with. Conversationally
Triggered Visual Thinking Stimuli. In CSCW, (2010),
103-106.
32. Wang, Y.-C., Burke, M., & Kraut, R.E., Gender, Topic,
and Audience Response: An Analysis of User-
Generated Content on Facebook. In CHI, (2013), 31-34.
33. Wash, R. and Lampe, C., The Power of the Ask in
Social Media. In CSCW, (2012), 1187-1190.
34. Kickstarter FAQ,
http://www.kickstarter.com/help/faq/creator+questions.
35. Kickstarter School,
http://www.kickstarter.com/help/school.
36. Kickstarter Stats, http://www.kickstarter.com/help/stats.
37. The Kick Back Machine,
http://www.thekickbackmachine.com.
38. Indiegogo. http://www.indiegogo.com.
39. RocketHub. http://www.rockethub.com.
Session: Crowdfunding and Crowd Storage
CHI 2014, One of a CHInd, Toronto, ON, Canada
600
... Both the method and frequency of communication are essential to increase the likelihood of success. Xu et al. (2014) attribute a far more significant increase in the probability of doing well to thorough interaction with the crowd rather than to the design of the project site (Xu et al., 2014). The critical role of communication via social networks for a crowdfunding project is also highlighted by Mollick (2014), who adds that geography can impact the project's success, such as the business founder's closeness to the crowd and potential financial backers and his/her efforts to make the campaign appeal to them as much as possible. ...
... Both the method and frequency of communication are essential to increase the likelihood of success. Xu et al. (2014) attribute a far more significant increase in the probability of doing well to thorough interaction with the crowd rather than to the design of the project site (Xu et al., 2014). The critical role of communication via social networks for a crowdfunding project is also highlighted by Mollick (2014), who adds that geography can impact the project's success, such as the business founder's closeness to the crowd and potential financial backers and his/her efforts to make the campaign appeal to them as much as possible. ...
... To differentiate their impact on the success of a crowdfunding campaign, the following hypotheses arise: Startnext, like many other crowdfunding platforms, allows the inclusion of a link to social media platforms on the campaign page. In turn, the link to the campaign page can be integrated on each platform (Xu et al., 2014). The characteristics of social media, including the fact that it is highly dynamic and has an extensive reach, can be actively used to achieve benefits for one's own campaign (Xu et al., 2014). ...
Article
Full-text available
Crowdfunding is a popular approach for entrepreneurs to obtain financial support and has therefore gained importance in the literature and practice in recent years. The communication behavior of starters that lead to the success of a campaign will be summarized and expanded by new findings. As a theoretical added value, this article offers the first attempt to transfer communication theories to the field of reward-based crowdfunding to reveal success determinants in crowdfunders user behavior. The purpose of this paper is to give practical recommendations to starters on how to run a campaign successfully and, more generally, on how they can make full use of social media. By analyzing the dataset of crowdfunding campaigns from Startnext, the largest German-speaking reward-based crowdfunding platform, new success factors are identified using multivariate analysis methods. The results show that social media can offer significant value for businesspeople in terms of marketing if it is used strategically and frequently.
... Greenberg et al. discovered that at the time of launch, success or failure of a project can be predicted with 68% accuracy by using SVM [18]. It has been found that features like having quality video [16], quality and consistent progress updates and comments [18][19][20][21][22], creator's backing history [23], trust relationship with backers [24], creator's personality traits [25,26], play an essential role in leading to successful funding. Mitra et al. [27] showed that language or certain phrases used by project creators have an impact in driving the crowd to invest in a project. ...
... Notably, scams tend to contain more images in their Campaign section; on average scams contain 17.1 images in a Campaign, whereas non-scams have 13.5 ones. It has been known that information contained in the Campaign section is less predictive of the success of crowdfunding campaigns as well [20,27], whereas those extracted from the Updates and Comments sections often serve as good predictors of success [20,22], which is in line with our results. ...
... Notably, scams tend to contain more images in their Campaign section; on average scams contain 17.1 images in a Campaign, whereas non-scams have 13.5 ones. It has been known that information contained in the Campaign section is less predictive of the success of crowdfunding campaigns as well [20,27], whereas those extracted from the Updates and Comments sections often serve as good predictors of success [20,22], which is in line with our results. ...
Article
Full-text available
Crowdfunding has seen an enormous rise, becoming a new alternative funding source for emerging companies or new startups in recent years. As crowdfunding prevails, it is also under substantial risk of the occurrence of fraud. Though a growing number of articles indicate that crowdfunding scams are a new imminent threat to investors, little is known about them primarily due to the lack of measurement data collected from real scam cases. This paper fills the gap by collecting, labeling, and analyzing publicly available data of a hundred fraudulent campaigns on a crowdfunding platform. In order to find and understand distinguishing characteristics of crowdfunding scams, we propose to use a broad range of traits including project-based traits, project creator-based ones, and content-based ones such as linguistic cues and Named Entity Recognition features, etc. We then propose to use the feature selection method called Forward Stepwise Logistic Regression, through which 17 key discriminating features (including six original and hitherto unused ones) of scam campaigns are discovered. Based on the selected 17 key features, we present and discuss our findings and insights on distinguishing characteristics of crowdfunding scams, and build our scam detection model with 87.3% accuracy. We also explore the feasibility of early scam detection, building a model with 70.2% of classification accuracy right at the time of project launch. We discuss what features from which sections are more helpful for early scam detection on day 0 and thereafter.
... Crowdfunding platforms can be exploited as social media tools in which the entrepreneurs interact with the potential investors and present the campaign as a showcase. More specifically, they can present and promote their project or company using texts, photos, or videos (Lai & Turban, 2008;Xu et al., 2014). Moreover, crowdfunding platforms, as social networks, allow the project's promoters to keep in touch with investors by publishing updates on their projects (Beier & Wagner, 2016;Mollick, 2014;Xu et al., 2014). ...
... More specifically, they can present and promote their project or company using texts, photos, or videos (Lai & Turban, 2008;Xu et al., 2014). Moreover, crowdfunding platforms, as social networks, allow the project's promoters to keep in touch with investors by publishing updates on their projects (Beier & Wagner, 2016;Mollick, 2014;Xu et al., 2014). Furthermore, in a crowdfunding campaign, especially the reward-based model, the fundamental of e-commerce resides. ...
... The existing crowdfunding platforms are somehow social media platforms where entrepreneurs can upload texts, photos, videos to promote their project or activity (Lai & Turban, 2008;Xu et al., 2014). Also, it allows them to interact with the crowd through publishing updates on their projects, answering questions of the crowd, and gathering comments from potential customers (Beier & Wagner, 2016;Mollick, 2014;Xu et al., 2014). ...
Article
Full-text available
The widespread use of the Internet and Web 2.0 emergence have promoted crowdfunding worldwide. In particular, crowdfunding exhibits significant growth potential in developing countries. It also constitutes a novel investment and financing mode that overtakes funding institution capabilities by providing practical marketing tools. Despite its considerable progress reported in developed countries, the deployment of this multidisciplinary economic model remains limited and restricted in developing countries. Turkey represents a promising emerging market for crowdfunding activities with its recent regulations. The potential of such an economic model is subject to further awareness of the crowd and investors (especially entrepreneurs) of crowdfunding concepts. This paper provides a narrative review where the authors aim to survey, synthesize, and discuss crowdfunding as a marketing tool and its relevant principles. Specifically, it depicts the bases and objectives of crowdfunding activities. Moreover, it introduces the marketing foundations relevant to this funding model to motivate the use of crowdfunding in developing countries such as Turkey.
... The appearance of continuous project management can give a favorable signal in terms of reliability. However, it was shown that the effect is different depending on the purpose of the updated post (Xu et al., 2014), or on frequency and timing (Block et al., 2018). Because, on the other hand, too much change signals rather low reliability at idea selection stage. ...
... Indeed, we observed a statistically significant effect of updates on crowd participation. However, the correlation showed non-linear results which were different from those seen in previous studies (Xu et al., 2014). We noticed that updates that deliver information might be valued highly by crowds (Block et al., 2018), but this tendency could also have unwanted effects. ...
Article
Whether the novelty of an idea is a factor that directly influences crowdfunding success remains an area of ambiguity. We hypothesize that target funder diversification is effective with incremental ideas. However, focused business proposals are better suited to assert radical ideas. We also hypothesize the impact of two different strategic actions that founders can take during fundraising campaigns, agile information update and communication, on crowdfunding success. A deep-learning-based novelty detection model combined with statistical analysis is used to empirically test 7406 crowdfunding projects crawled from online platform. Our results support our hypotheses and reveal that information updates from startup founders show non-linear quadratic relationships with fundraising performance, whereas two-sided communication helps stimulate investors. We also revealed that novelty level can influence strategic choice, indicating that a project with a higher novelty should have a focused target. Our finding suggests a solution to the conflicting conclusions in previous studies on the direct impact of novelty level and target diversification, by explaining the process of novelty-dependent behavioral strategies based on signaling theory.
... Authors such as Thies et al. (2014) have considered the influence of social media activity. Xu et al. (2014) argued that updating content in this type of media has become critical. Therefore, the social media activity is related to key business development outcomes, and the social media presence of crowdfunding projects is a key component of their overall campaign strategy (Wolfe et al., 2021). ...
Article
Full-text available
Crowdlending platforms have gained importance in recent years due to digitalization. The Covid-19 pandemic has severely impacted the entire economy, including repercussions on digital transactions. Investors considering transactions on these platforms require confidence in both the platform and the project in order to make the right decision. For this reason, this study analyzes the links between the parties involved in the process. A survey was sent to 135 investors on the Colectual platform in January 2022, and the research method adopted is Fuzzy Set Qualitative Comparative Analysis (fsQCA). The study examines the connections between the platform, the investors, and the project developers, to assess the influence exerted by the platform on the confidence of the parties seeking and providing investment. The main result is that Covid-19 has been a key factor in the confidence of investors in the platform, and it has had a significant influence on the investments made.
... Previous research on financial crowdfunding shows that project attributes, such as the target, duration, and the number of Facebook friends of the initiator, affect the success of the project (Greenberg et al., 2013;Li and Duan, 2014;Mollick, 2014;Xu et al., 2014;Kuppuswamy and Bayus, 2015). Meanwhile, in the humanitarian communication side, based on Text-analytical approaches scholars explored the role of mission statement, donation appeals and humanitarian advertising for charitable organizations. ...
Article
Full-text available
The COVID-19 outbreak has been a public health crisis of international concern, causing huge impact on people’s lives. As an important part of social public crisis management, how to quickly and effectively raise resources to participate in emergency relief in the era of self-media is a common challenge faced by global charitable organizations. This article attempts to use empirical evidence from Tencent charitable crowdfunding platform, the largest charitable crowdfunding platform in China, to answer this question. We consider 205 COVID-19 charitable projects and 11,177,249 donors to assess the process by which non-profit organizations raise funds through the information about project descriptions. Based on the effects of information and emotional framing, we explore the effects of the readability (i.e., complexity and understandability) and negative tone of the project description on fundraising amount. We then investigate the mediating role of forwarding times, as affective response to the text might explain forwarding times, which in turn affects money raised by increasing the visibility of the campaign. On this basis, the moderating role of recipient’s crisis involvement is tested during this process. The empirical results indicate that the complexity of the description will reduce the fundraising amount, while understandability and negative tone help to improve it. Furthermore, we found that forwarding times played an important mediating role in this process. Then the buffer effect of crisis involvement on the negative effect of complexity was validated, and its amplification on the positive effects of understandability was also verified.
Article
A sense of community is important in encouraging people to contribute to a variety of causes and the communities that support them. Researchers have identified website design features that can engender a sense of community on sites to promote contributions. However, most findings about design features are based on observational empirical research testing single features at a time or on standard practice and rarely use integrated theories to provide rationale for their design suggestions. This work investigates ways to re-design an entire website---with a simulated medical crowdfunding interface entitled Community Journey---informed by Sense of Community Theory to increase site visitors' sense of community and contributions. A between-subjects experiment revealed that the Community Journey interface increased potential supporters' sense of community and their overall willingness to contribute via monetary donations, campaign shares, personal messages, and offline support. Think-aloud interviews identified the interface features responsible for the overall increase in willingness to contribute. Finally, we suggest theory driven design implications for creating websites to build a strong support community and to encourage various contributions.
Article
This article investigates multimodal elements—images, links, gifs, videos, and galleries—of crowdfunding campaigns on the platform Kickstarter to develop an understanding of characteristics of successful campaigns. The authors scraped 327,586 campaign pages, analyzing the multimodal elements of successful and unsuccessful campaigns. They found that successful campaigns featured more images, links, and gifs and more frequently included a project video than did unsuccessful campaigns. Images, links, and the presence of a project video had a positive impact on success while gifs and project galleries did not. These findings give business communicators practical guidance, develop theoretical aspects of Kickstarter research, and validate previous findings with a larger data set.
Article
Full-text available
PURPOSE: While there is abundant literature on the key determinants of reward-based crowdfunding success, little research is dedicated to crowdfunding projects that are not only successful but receive significantly more funds than initially targeted through the defined funding goal. This study seeks to shed light on this vastly neglected topic in crowdfunding research. METHODOLOGY: Drawing on a rich dataset of 338 reward-based crowdfunding projects, this study applied a two-step statistical analysis. First, regression analyses to determine relevant crowdfunding success factors were conducted in order to corroborate extant literature and to highlight that the data properly reflects the already identified key findings on crowdfunding success. In a second step, the very same factors were investigated for the case of overfunded projects, utilizing logistic regression analyses and a Blinder-Oaxaca Decomposition. FINDINGS: Although this study confirmed the findings of previous research considering the factors that increase the success probability of crowdfunding projects, the very same factors turned out to not explain the emergence of project overfunding. For instance, while project founders can provide updates, a higher number of different rewards, or utilize social media pages to increase the likelihood for success, these factors do not contribute to explain the phenomenon of project overfunding. IMPLICATIONS: The results of this study emphasize that in order to understand overfunding of crowdfunding projects, future research must go beyond the basic crowdfunding success factors. Building on the notion of the Two-Factor Theory, the findings suggest that the factors contributing to success can be considered hygiene factors that are required to succeed in the first place. However, these factors do not motivate the crowd to provide further funding to an already successful project. Hence the motivating factors remain yet unobserved in extant literature. In practice, this means that project teams achieving their funding goal cannot rely on the same factors that were helpful to succeed to encourage further funding from the crowd. The differentiation of hygiene and motivating factors for overfunding in reward-based crowdfunding offers rich opportunities for future research. More subjective factors, such as the individual perception of crowd members towards crowdfunding projects, are suggested to play an important role for the occurrence of project overfunding. ORIGINALITY/VALUE: By investigating project overfunding, this study addresses the research gap concerning the factors contributing to the emergence of project overfunding. There is little evidence on the characteristics of overfunded crowdfunding projects, and thus this study provides essential theoretical and empirical groundwork for future research to build upon this study's results.
Article
Full-text available
Crowdfunding is changing how, why, and which ideas are brought into existence. With the increasing number of crowdfunded projects, it is important to understand what drives people to either create or fund these projects. To shed light on this new social phenomenon, we present a grounded theory of motivation informed by the first cross-platform qualitative study of the crowdfunding community. By performing 83 semistruc-tured interviews, we uncover creator motivations, which include the desire to raise funds, expand awareness of work, connect with others, gain approval, maintain control, and learn; and supporter motivations, which include the desire to collect rewards, help others, support causes, and be part of a community. We also explore deterrents to crowdfunding participation, including, among creators, fear of failure, and, for supporters, lack of trust. Based on these findings, we provide three emergent design principles to inform the design of effective crowdfunding platforms and support tools.
Article
Full-text available
This study examined the relationship between language use and persuasion success in the Peer-to-Peer (P2P) lending environment where unaffiliated individuals borrow money directly from each other using a textual description to justify the loan. Over 200,000 loan requests were analyzed with Linguistic Inquiry and Word Count (LIWC) software. The use of extended narratives, concrete descriptions and quantitative words that are likely related to one's financial situation had positive associations with funding success which was considered to be an indicator of trust. Humanizing personal details or justifications for one's current financial situation were negatively associated with funding success. These results offer insights into how individuals can optimize their persuasiveness by monitoring their language use in online environments.
Article
Crowdfunding is changing how, why, and which ideas are brought into existence. With the increasing number of crowdfunded projects, it is important to understand what drives people to either create or fund these projects. To shed light on this new social phenomenon, we present a grounded theory of motivation informed by the first cross-platform qualitative study of the crowdfunding community. By performing 83 semistructured interviews, we uncover creator motivations, which include the desire to raise funds, expand awareness of work, connect with others, gain approval, maintain control, and learn; and supporter motivations, which include the desire to collect rewards, help others, support causes, and be part of a community. We also explore deterrents to crowdfunding participation, including, among creators, fear of failure, and, for supporters, lack of trust. Based on these findings, we provide three emergent design principles to inform the design of effective crowdfunding platforms and support tools.
Conference Paper
Although both men and women communicate frequently on Facebook, we know little about what they talk about, whether their topics differ and how their network responds. Using Latent Dirichlet Allocation (LDA), we identify topics from more than half a million Facebook status updates and determine which topics are more likely to receive feedback, such as likes and comments. Women tend to share more personal topics (e.g., family matters), while men discuss more public ones (e.g., politics and sports). Generally, women receive more feedback than men, but "male" topics (those more often posted by men) receive more feedback, especially when posted by women.