Crowdfunding in the gaming industry : the influence of success factors in crowdfunding on the success of the final product

(1)

Crowdfunding in the gaming

industry

The influence of success factors in crowdfunding on the

success of the final product.

Submission date: 1 July 2018 MSc Entrepreneurship VU / UvA

Tom Marcus VU: 2528692 / UvA: 1137276 Thesis supervisor: dhr. dr. J. (Joeri) Sol

overuse of "states", which is unclear whether it concerns statements based on empirical evidence or theoretical reasoning

innovative data collection

poor referencing to tables, regression results only in appendix, no mention of typical limitations cross-sectional analysis

(2)

1

Preface

In front of you lay the master thesis “Crowdfunding in the gaming industry”. This thesis is established on behalf of the MSc program Entrepreneurship facilitated by Vrije Universiteit Amsterdam and University of Amsterdam. The universities cannot be held liable for the content of this thesis. I, Tom Marcus, am the only one that is responsible for the content of this master thesis, mistakes included.

This way I want to thank both universities for the support I have received during this master year. Especially Joeri Sol gave useful feedback, so a special thank you for dhr. Dr. J. Sol. I would also like to thank my girlfriend for supporting me during the writing of this thesis. She was the one encouraging me when I was not motivated enough.

I hope you will enjoy reading this thesis.

Tom Marcus

(3)

2

Abstract

Crowdfunding is a relatively new way for entrepreneurs to finance their business. Kickstarter is one of the biggest and oldest platforms that facilitate reward-based crowdfunding. To what extent does the success of a reward-based crowdfunding campaign influence the success of the final product? This thesis integrates a dataset collected from the Kickstarter platform with a dataset of video games on the Steam platform to explore which factors influence crowdfunding success and the success of the crowd funded product. This thesis tests the influence of the amount of backers (Backers), levels of rewards (Levels), updates placed by the founder (Updates), comments placed (Comments) and the duration of the campaign (Duration) on crowdfunding success. The same variables plus the variable success of the crowdfunding campaign (measured in % overfunding) are used testing the influence on new product success. The data reveals an influence of the amount of backers, amount of comments and amount of updates in a crowdfunding campaign on the success of a crowdfunding campaign. Finally, the amount of comments in a crowdfunding campaign and the amount of overfunding have an influence on the success of the new product. The results of this research add to a field of research towards crowdfunding which has not been explored a lot; the post-campaign phase of crowdfunding.

(4)

3

1. Introduction

Crowdfunding is a relatively new way for entrepreneurs to finance their business. Early research on this topic originates from the years 2007 until 2012. Nowadays crowdfunding is a booming business. While in 2012 the total funding volume was $2.7 billion, in 2015 this amount was grown to $34.4 billion. The trends show that this figure will keep on growing the next couple of years (Massolution, 2015). The successful launch of several businesses and products depend heavily on the money raised by making use of crowdfunding. Therefore it is of great importance for the entrepreneur to know how he/she can influence the success of a crowdfunding campaign and how to turn a successful crowdfunding campaign into a successful product.

According to Hemer (2011) crowdfunding emerged in creative industries like the music- and film industry. Another industry that is well suited for crowdfunding is the video game industry. Kickstarter is a well-known platform which facilitates reward-based crowdfunding campaigns for creative industries. They have divided all campaigns in certain categories of which the game category is the one where the most money is invested in. One sub-category of the game category is the video game category.

Kickstarter is one of the biggest and oldest platforms that facilitate reward-based crowdfunding (Kuppuswamy & Bayus, 2015). Of the over 350.000 launched projects slightly over 35% successfully raised the money pledged for (table 1). ("Kickstarter Stats — Kickstarter", 2017)

Table 1. Kickstarter statistics ("Kickstarter Stats — Kickstarter", 2017)

Launched projects Total dollars Successful dollars Unsuccessful dollars

(6)

5 The global video games market is worth $101.1Bn in 2016. The gaming industry is estimated to grow with a rate of 6.2% each year from 2015 until 2020. (Newzoo Games, 2017) The video game industry emerged in the early 1970’s whenthe first video games were released to the mass public. Back then video games were only being developed by huge technological companies. Zackariasson, Styhre & Wilson (2006) stated that the entry into the video game market is getting increasingly costly due to technological advancement of videogames. While this was only 11 years ago a lot has changed in the meantime. Due to the technological development of software and hardware and the knowledge available on the internet, nowadays the development of video games is accessible for everybody. Those developments also lead to a decrease in capital needed to develop a videogame. Crowdfunding lowers the entry barrier to this market even more. While earlier only the big development studios had access to the funding necessary for video game development, nowadays with crowdfunding everybody has a chance to raise the capital needed.

The pursuit of independency results in the video game developer to act like an entrepreneur. Besides developing a video game the developer has to do the marketing and publish his own game. Developing a video game is a time consuming process. This means there is little to no time to work at another job. There is a need for capital to be able to develop a video game. Nowadays, raising funds through the use of crowdfunding is more common than ever in the video game industry.

The gaming industry also profit from the digitalization. While 10 years ago most games were still bought on CD/DVD, nowadays most of the games are being downloaded through the internet. There is often no physical copy anymore. There are several platforms which facilitates this. The biggest platform which provides this is Steam. Steam is an online platform where a consumer can buy and download video games. Besides this, they also host a library

(7)

6 of all games bought. Steam facilitates a community hub for every single game where gamers can get in touch with each other.

Mollick (2014) states that most of the crowdfunding campaign founders try to deliver a product to the backers, but a few are succeeding in doing this in a timely manner. This problem gets bigger when projects are large or overfunded.

The research question of this thesis is as follow:

“To what extent does the success of a reward-based crowdfunding campaign influence the success of the final product?”

The sub questions are as follow:

 What is reward-based crowdfunding?

 What defines the success of a reward-based crowdfunding campaign?

 What defines the success of the final product?

The paper is of theoretical relevance because it will give new insights in the academic world of reward-based crowdfunding. Most research into reward-based crowdfunding focuses on the process of the crowdfunding campaign itself. This thesis will also take a look at what happens after the crowdfunding campaign is successful. By merging two datasets, one concerning crowdfunding campaigns and one concerning the final product, into one this thesis can do statistical tests not performed before in the crowdfunding literature by our knowledge. This master thesis is also one of the few papers that give an insight in the funding in the big video game industry.

The paper is of practical relevance for entrepreneurs, and more specific (indie) game developers, who are considering starting a reward-based crowdfunding campaign to raise capital. If the entrepreneur knows what factors influence the success of crowdfunding success

(8)

7 but also the success of the final product they can anticipate on those important factors so they can influence their success themselves.

This thesis is structured as follow: In chapter 2, this thesis will discuss the relevant literature of this thesis its topic; crowdfunding. Chapter 3 will discuss the methodology used in this thesis. The results of the research conducted, using the methodology discussed in chapter 3, are shown in chapter 4. In chapter 5 the results of chapter 4 will be discussed. Finally, in chapter 6, conclusions will be made with the help of earlier chapters. In chapter 7 limitations and future research on this topic will be discussed. At the end of this thesis are the appendices.

(9)

8

2. Literature

2.1 Crowdfunding

To define crowdfunding it is important to know where it originates from. According to Mollick (2014) crowdfunding is a combination of crowdsourcing and micro-financing. Crowdfunding emerged in creative industries like the music- and film industry (Hemer, 2011). Early work done on crowdfunding by Schwienbacher & Larralde (2010) defines crowdfunding as “an open call, essentially through the Internet, for the provision of financial resources either in form of donation or in exchange for some form of reward and/or voting rights in order to support initiatives for specific purposes” (p. 4). Important to add is that the contributions, of all individuals which are part of the crowd, are relatively small compared to other forms of investing. Also the time of a crowdfunding campaign is not endless. The time a crowdfunding campaign takes is often a couple of weeks, up to several months. Another important feature is that potential investors can see the support the campaign gets from other investors (Kuppuswamy & Bayus, 2015).

Crowdfunding thanks it success to three factors. First is the lack of investments made due to the financial crisis. Second is the emerging of web 2.0. Last is the success of crowdsourcing (Rossi, 2014). Crowdfunding spreads the risk, which comes with investing, over a large group of investors. It makes it possible to invest small amounts of money, but still being able to get high return rates.

Within crowdfunding there are 3 different parties which are necessary to make crowdfunding possible. First of all is the party that creates the crowdfunding campaign. This can be an entrepreneur but crowdfunding does not have to be business oriented. Mollick (2014) calls the creator of a campaign a founder. The second party consists of the investors in the crowdfunding project/campaign. One of the unique features about crowdfunding is the

(10)

9 numbers of investors possible. A crowdfunding campaign needs at least one investor but the number of investors can reach above 10.000 individual investors. Mollick (2014) refers to the investors as funders. Another term that is widely used is the term backer. The third party that makes crowdfunding work, is the platform which facilitates the crowdfunding. Some well-known examples of a crowdfunding platform are Kickstarter, Indiegogo and RocketHub. Kuppuswamy & Bayus (2015) refer to a crowdfunding platform as a crowdfunding community.

There are several types of crowdfunding to be distinguished. Barnett (2013) states there are 2 main models in crowdfunding; donation-based funding and investment crowdfunding. Ahlers, Cumming, Günther and Schweizer (2015) stated that there are four main types of crowdfunding: donation-based, reward-based, lending and equity. The four categories are based on the reward the investor receives for their contribution. Early work done on crowdfunding by Hemer (2011) concludes there are 5 forms of crowdfunding; donations, sponsoring, pre-ordering or pre-selling, lending and equity. Donations falls under donation-based funding and the other 4 types fall in the investment crowdfunding category. Figure 1 shows the difference in complexity between the different categories. Bradford (2012) divided crowdfunding into 5 different models; the donation model, the reward model, the pre-purchase model, the lending model and the equity model. He also states that there are crowdfunding sites that facilitate more than one of these models.

Lending-based crowdfunding is by far the most used category of crowdfunding. Of the estimated fundraising volume of crowdfunding ($34 billion), $25.1 billion is assigned to lending-based crowdfunding. The volume of reward-based crowdfunding is $2.68 billion (Massolution, 2015).

(11)

10 The only difference between both categorizations is between sponsoring and reward-based crowdfunding. Hemer (2011) states the crowdfunding type “sponsoring” involves a reward. Often this reward takes the shape of services like PR or marketing for the sponsor. Both types are pretty similar. Besides that the reward-model and the pre-purchase model are often mixed up. For example the popular crowdfunding platform “Kickstarter” uses both models at the same time. The projects can offer a list of different rewards dependent on the amount of money invested. If one of those rewards is the end-product we talk about pre-purchase. This paper will focus on reward-based crowdfunding.

Figure 1. The complexity of different crowdfunding categories (Hemer, 2011)

The research of Mollick (2014) made use of data from the crowdfunding platform Kickstarter. This platform is focused on reward-based crowdfunding. From the data Mollick concludes that crowdfunding projects often succeed by only a small margin. If a project fails this is often by a large margin. In reward-based crowdfunding the behavior of the funders is bathtub-shaped (Kuppuswamy & Bayus, 2015). This means that most funding of a crowdfunding campaign happens in the first and the last weeks of the campaign.

According to Hui, Gerber & Greenberg (2012) there are six different categories of crowdfunding work: Understand the opportunities and responsibilities, prepare the campaign material, test the campaign material and initial project prototypes, market the project, execute

(12)

11 the project goals and contribute knowledge back to the crowdfunding community. Understand the opportunities and responsibilities prepare the campaign material and test the campaign material and initial project prototypes are pre-project launch activities. Market the project is a pre-project and during project activity. Execute the project goals and contribute knowledge back to the crowdfunding community are post-project activities. If a crowdfunding campaign fails there is no need to carry out post-project activities. These post-project activities do not contribute to the success of a crowdfunding campaign but can contribute to the success of the final product.

2.2 Success of a crowdfunding campaign

Koch & Siering (2015) found that both founder-specific and campaign-specific aspects influence the success of a crowdfunding campaign. Cordova, Dolci & Gianfrate (2015) found there are three factors which have an impact on either success or overfunding of a crowdfunding project: the amount of investment requested by the founder, the duration of the campaign and the contribution frequency of the project. This confirms earlier research by Mollick (2014) and Kuppuswamy & Bayus (2015). However, according to Frydrych, Bock, Kinder & Koeck (2014) shorter duration signals legitimacy.

Gerber, Hui & Kuo (2012) define the success of a crowdfunding campaign as “reaching the funding goal”. Founders of a crowdfunding campaign do seek for funding at the first place. But this is not the only reason for founders to choose for crowdfunding to raise capital. Although reasons may vary per founder there are some other reasons, next to raising funds, for founders to engage in crowdfunding. Other reasons are: Establish relationships, receive validation, replicate successful experience of others and expanding awareness of work through social media (Gerber et al., 2012). Belleflamme, Lambert & Schwienbacher (2013) state that questionnaires sent to entrepreneurs show that crowdfunding facilitates the entrepreneur to attract attention onto their own company.

(13)

12 Potential funders of a reward-based crowdfunding campaign do not only evaluate the business idea. Potential funders also look at the team composition. Research shows that crowdfunding campaigns created by pairs and teams show higher crowdfunding success rates than individuals do. Another remarkable conclusion of the study conducted by Frydrych et al. (2014) is that crowdfunding campaigns created by women seem to have higher success rates than men.

Research by Belleflamme, Lambert & Schwienbacher (2014) points out the importance “to build a community that ultimately enjoys additional private benefits from participation” (p. 10). Schwienbacher & Larralde (2010) state that the entrepreneur engaging in crowdfunding should be open for other people’s opinion. The reason is that potential backers seek for projects where they can be of use and participate in. Hui, Greenberg and Gerber (2014) state that: “While both entrepreneurship and crowdfunding rely on collaboration to achieve project goals, crowdfunding inherently relies more on community support for project success” (p. 10).

The entrepreneur has to take care of his community because otherwise the community can disappear. There are three reasons which point out the importance of the community. In the first place each member of the community can become a lifelong fan of the firm. Secondly the community can generate ideas and become an important source for the firm. Thirdly the members of the community can become clients as well (Rossi, 2014).

Sometimes crowdfunding platforms can be seen as a special online community. This is the case when a group of contributors actively use a platform for a long period, communicating among other contributors and with the entrepreneurs behind the campaign. Another condition is that the communication should be facilitated by some sort of technological infrastructure. Examples are a forum or message board. (Galuszka, & Bystrov, 2014)

(14)

13 One way to keep in touch with the online community is by placing updates about the product development and the crowdfunding campaign. When the numbers of updates increase this has a positive effect on the success probability of the crowdfunding campaign. With the updates the founder of the campaign shows project commitment (Joenssen, Michaelis & Müllerleile, 2014). Xu, Yang, Rao, Fu, Huang, & Bailey (2014) did research into project updates during a crowdfunding campaign. They found that the way how the founder of a campaign communicates with (potential) backers during a crowdfunding campaign predicts the success of that campaign more than the project page representation.

Little research is done towards the effect of the reward-level structure and success in crowdfunding. Frydrych et al. (2014) did not provide clear implications for this relation. They do argue that “the reward-levels demonstrate an important factor that makes a project compelling to the audience” (p. 14). The success of a crowdfunding campaign is positively correlated when there are more pre-selling rewards or rewards that provide social reputation to the backer of the project. It is negatively correlated with rewards which are not product-related services (Crosetto & Regner, 2014).

Project success is positively correlated with a higher presence of pre-selling rewards and of rewards that provide social reputation to the pledger, while it is negatively correlated to the provision of not product-related services.

Crowdfunded video games are often still in development when the campaign launches. There are different stages to be distinguished in the development of video games. It all starts with the concept phase where the idea for the video game is created. The actual development starts with the pre-production of the video game. After this phase the actual production starts. When the video game is starting to look like an actual game the alpha phase starts. After the alpha

(15)

14 phase the beta phase starts. In the beta phase the game is almost done. This phase is being used for bug fixing.

The optimal crowdfunding campaign window for video games has changed a lot the last couple of years. In 2012 the crowdfunding campaign window was between the finishing of the concept phase until the pre-alpha stage. This is visualized in figure 2. Between 2012 and 2018 this campaign window shifted towards the launch date. In 2018 the crowdfunding campaign window lays between the pre-alpha stage and the end of the alpha stage. This is shown in figure 3. This shift towards the launch date should lead to a shorter time between the end of the crowdfunding campaign and the actual launch of the product (Bidaux, 2018).

(16)

15 Figure 3. the crowdfunding campaign window in 2018 (Bidaux, 2018)

2.3 Success of the final product

The success of the final product can also be seen as the success of a new product. Literature refers to this as new product development (NPD) success. New product performance is a multidimensional concept (Cooper & Kleinschmidt, 1995). There are 4 dimensions of new product success: The quality of the new product, financial new product success, the quality of the new product development process and the inexpensiveness of the new product ownership. Each of these dimensions measure a different aspect of NPD (Gruner & Homburg, 2000). This research also states that customer interaction while in early and late stages of the NPD can increase the success of the new product. If we look at the video game development cycle the early stage refers to the concept phase and the late stages refers to the alpha phase and beyond.

Mollick (2014) states that a lot of founders attempt to deliver a product but that most are not able to do this in a reasonable time. This problem gets bigger when the project is large or overfunded. If the launch date of the product is delayed this will negatively affect the credibility of the preannouncing company (Kohli, 1999). Company credibility influences the purchase intention of a customer (Lafferty, Goldsmith & Newell, 2002).

(17)

16

3. Methodology

This research wants to measure if there is a relation between the success of a crowdfunding campaign and the success of the product (partly) funded by the crowdfunding campaign. Literature shows there are variables which influence the success of a crowdfunding campaign. But will they also influence the success of the product?

Kickstarter is market leader in crowdfunding video games. Steam is market leader in facilitating the sales of video games. Therefore, this thesis will take a dataset of video games crowd funded at the platform Kickstarter and a dataset of video games on the Steam platform. This thesis will make use of a Kickstarter dataset which contains 861 crowdfunding campaigns in the category video games. The campaigns were created in the years 2009 until 2012. The reason for this is that most video games in this dataset had a planned release date which is before 2018. This means that all video games in this dataset that are successfully funded should be out on the market. If a more recent dataset would be used this would probably not be the case. It makes sense that the steam dataset is of a later date; 2018. This dataset contains 223 steam games which are crowd funded on Kickstarter.

The Kickstarter dataset is scraped from the Kickstarter website. The Steam dataset is scraped from SteamSpy, a website with all kinds of statistics concerning Steam. Both datasets are taken from kaggle.com and were edited and combined (by the author) into one dataset exclusively with video games crowd funded on Kickstarter.

Gerber et al. (2012) stated that a crowdfunding campaign is a success when the funding goal is reached. The independent variable therefore is “percentage of funding”. When the funding percentage reaches 100% the crowdfunding campaign is a success. The Kickstarter dataset limits this research in the choice of variables which influence the success of a crowdfunding

(18)

17 campaign. The dependent variables are as follow: Amount of backers, levels of rewards, updates, comments and duration.

Researchers like Rossi (2014) stressed the importance of building a community during a crowdfunding campaign. Joenssen et al. (2014) stressed the importance of updates. The dataset used for this thesis measures this for each campaign. The variables are the amount of updates placed by the founder of the campaign and the amount of comments placed by either the founder or the backers of the campaign. This leads to the following hypotheses:

H1: The amount of updates by the founder influences the percentage of funding.

H2: The amount of comments influences the percentage of funding.

Cordova et al. (2015) stated that the contribution frequency of the project has an impact on the success of a crowdfunding project. Therefore the amount of backers is a useful variable. This leads to the following hypothesis:

H3: The amount of backers influences the percentage of funding.

Research conducted by Cordova et al. (2015) showed the influence of the duration of a campaign on the success of a campaign and on overfunding. This leads to the following hypothesis:

H4: The duration of a crowdfunding project, influences the percentage of funding.

Crosetto & Regner (2014) state there are different kinds of rewards. For some rewards they found a positive correlation and for others they found a negative correlation. Frydrych et al. (2014) do argue that reward-level structures are important in reaching crowdfunding success. Because the literature is not clear about this variable we will

(19)

18 After checking if our findings are in line with earlier research on factors influencing crowdfunding success we will focus on the influence of crowdfunding success factors on new product success. As stated before there are 4 dimensions in measuring new product success. With the dependent variables “Userscore” and “Owners” we tend to measure new product success through the factor “quality of the new product” (Gruner & Homburg, 2000).

Gruner & Homgburg (2000) found that customer interaction influences the new product success in the early and late stages of NPD. Schwienbacher & Larralde (2010) found that potential backers seek for projects where they can be of use and participate in. If there are more backers, there is a bigger online community. Factors concerning the online community of a crowdfunding campaign could also concern the online community in NPD because this community will not just disappear after the crowdfunding campaign is successful. This leads to the following hypotheses:

H6: The amount of updates by the founder influences the success of the final product.

H7: The amount of comments influences the percentage of the final product.

H8: The amount of backers influences the success of the final product.

There is no literature supporting that there is an influence of the variables “Duration” and “levels” on new product success. If these variables seem to influence the success of a crowdfunding campaign this thesis still wants to test the following hypotheses:

H9: The duration of a crowdfunding project, influences the success of the final product.

(20)

19 At last it should be tested if the percentage of funding influences the success of the final product. With this variable we can test the influence of overfunding on the success of the final product.

H11: The percentage funded influences the success of the final product.

This thesis will perform quantitative tests to answer this question. Regressions will be performed using the stepwise method. First regressions are performed to try to understand which factors influence the success of a crowdfunding campaign. The goal of the second regressions is to see if the factors which influence the funded percentage also influence the success of the final video game.

(21)

20

4. Results

In this chapter the results from testing of the hypotheses are shown. The output of SPSS is available in appendix 1.

From all 861 video games campaigns only 216 were successful (25.2%), 489 failed (57.1%), 8 were cancelled (0.9%) and 144 campaigns were still live (16.8%). If the cancelled and live campaigns were removed the successful percentage would be 30.6% against 69.4% failed campaigns. While only 30.6% of the campaigns were successful, the mean of the funding percentage is 67.1%. When the funding percentage reaches 100% the crowdfunding campaign is successful. The reason for the percentage of 67.1% is that some campaigns are highly overfunded. The maximum funding percentage measured in this dataset is 3248.30%. The average backed amount is $53.72. It is unknown how much of the projects that were still live were successful and how much campaigns failed. This thesis will exclude the canceled and live campaigns out of this research.

Table 2. Descriptive statistics campaign status

Frequency Percent Valid Percent Cumulative Percent Valid canceled 8 .9 .9 .9 failed 489 57.1 57.1 58.0 live 144 16.8 16.8 74.8 successful 216 25.2 25.2 100.0 Total 857 100.0 100.0

There are a total of 17087 video games published on the Steam platform. Of all these video games 223 are (partly) funded by Kickstarter. This is 1.3% of all the video games published on Steam. Of the 216 successful campaigns only 37 are released on the Steam platform. 7 campaigns of the 144 which were live made it to a release on Steam.

First a regression is performed to try to understand which factors influence the success of a crowdfunding campaign. Only successful and failed campaigns are included. The variable

(22)

21 “Fundedpercentage” says something about the pledged amount compared to the goal of the campaign. The funded percentage of a campaign is regressed on the following variables:

 Amount of backers (Backers)

 Levels of rewards (Levels)

 Updates placed by the founder (Updates)

 Comments placed (Comments)

 Duration of the campaign (Duration) Table 3. Descriptive statistics variables

N Minimum Maximum Mean Std. Deviation

Backers 705 0 87142 526.02 4440.486 Levels 705 0 37 8.30 3.809 Updates 705 0 45 5.44 7.217 Comments 705 0 13850 129.32 1012.146 Duration 705 4.0 91.0 38.810 15.8903 Fundedpercentage 705 0.00% 3248.30% 67.1354% 182.63509% Valid N (listwise) 705

Two separate standard multiple regressions were performed using the stepwise method. At first, a regression which contains all successful and failed campaigns. The second model only includes successful and failed campaigns with a funding goal above $5.000. Mollick (2014) states that crowdfunding campaings have big differences in funding goals. It makes a big difference if the goal is $500 or $50.000. Therefore Mollick adopted a threshold of $5000. This makes it able to compare it with seed financing by venture capitalists and angel investors. It also removes some outliers.

In the first model a significant regression equation was found (F(3,701) = 41.823, p< .000), with an R2 of .152. The predicted funded percentage is equal to 22.481 + 7.510 (Updates) + 0.14 (Backers) - 0.028 (Comments). The funded percentage increased 7.510% for each extra update, 0.014% for each extra backer and decreases -0.028 for each extra comment. Updates, backers and comments are significant predictors of the funded percentage.

(23)

22 In the second model a significant regression equation was found (F(3,446) = 78.691, p <.000), with an R2 of .348. The funded percentage increased 0.014 for each extra backer, 6.632% for each extra update, and decreases -0.024 for each extra comment. Backers, updates and comments are significant predictors of the funded percentage.

The goal of the second regression is to see if the factors which influence the funded percentage also influence the success of the final video game. Only successful campaigns with a funding goal above $5.000 are included. The success of the product is measured by the amount of owners (Owners) of the video game and the user score (Userscore) of that same game. The user score of a game is the percentage of positive ratings which owners of the video game have given for the video game. Owners can give a positive or negative rating for the video game. Unfortunately the revenue made by the game is not available.

The amount of owners of the video game is right-skewed with a skewness of 11.15 and a standard error of 0.92. According to Cramer & Howitt (2004) the skewness has to be divided by the standard error to see if the skewness is statistical significant. If the outcome is bigger than 1.96 the skewness is significant. In this case the skewness of the owners is significant. To work with a right-skewed variable the variable has to be logged. This results in the independent variable “LnOwners” which has a better normal distribution. The descriptive statistics of the dependent variables are shown in table 4.

Table 4. Descriptive Statistics of successful Kickstarter campaigns on Steam (>$5000)

Owners 30 6558 2849140 542808.20 772602.486

LnOwners 30 8.79 14.86 12.1712 1.59196

Userscore 30 21 96 74.43 19.138

(24)

23 Table 5 shows the descriptive statistics of the different variables used in regression model three and four. The success of the final product is regressed on the following variables:

 Amount of backers (Backers)

 Levels of rewards (Levels)

 Updates (Updates)

 Comments (Comments)

 Duration (Duration)

 Funded percentage (Fundedpercentage)

Table 5. Descriptive Statistics of successful Kickstarter campaigns (>$5000)

LnOwners 34 7.87 14.86 11.9589 1.69660 Userscore 118 0 96 18.92 33.915 Owners 118 0 2849140 140523.10 451645.625 Fundedpercentage 118 100.21% 2005.43% 185.4740% 206.63645% Backers 118 15 87142 2953.81 10557.277 Levels 118 4 37 11.42 4.924 Updates 118 1 41 14.88 8.839 Comments 118 0 13850 741.15 2389.125 Duration 118 7.0 91.0 38.127 13.7339 Valid N (listwise) 34

A significant regression equation was found (F(2,115) = 22.144, p< .000), with an R2 .265. The predicted user score increases with 0.005 for each extra comment and 0.051 for each extra percentage of funding. Both comments and the funded percentage are significant predictors of the user score.

A significant regression equation was found (F(3,114) = 14.765, p< .000), with an R2 .280. The predicted LnOwners increases with 0.001 for each extra comment, 0.008 for each extra percentage of funding and 0.106 for each update. Comments, the funded percentage and the amount of updates are significant predictors of the LnOwners.

refers to which table?

(25)

24

5. Discussion and conclusion

5.1 Discussion

What we see a lot with video games that are (partly) financed by crowdfunding is that the developers are not able to deliver the promised product (the game) in time. Out of the 216 successful campaigns, only 37 were able to release their video game on the Steam platform. This is in line with the study of Mollick (2014) and Kohli (1999).

In chapter 3 the hypotheses of this thesis were explained. The regressions done in chapter 4 result in the rejection of the following hypotheses: H4, H5, H8, H9, H10.

The first two regression models are to examine the influence of different variables on the success of crowdfunding. The hypotheses H4 (the duration of a crowdfunding project, influences the percentage of funding) and H5 (the amount of levels of rewards influences the percentage of funding) are rejected. The rejection of H4 is interesting. Multiple scientific papers claim that there is an influence. Authors like Cordova et al. (2015) and Mollick (2014) have found evidence for this claim. However, according to Frydrych et al. (2014) shorter duration signals legitimacy. Another explanation for this could be that the influence of the duration could differ between different categories within Kickstarter. The reward-level construction has been examined by different researchers and they could not find a relationship either. This does not mean that setting the right reward-levels is not important. Frydrych et al. (2014) argue that setting the right reward-levels do make a difference toward the success of a successful crowdfunding campaign.

The results shown in chapter 4 fail to reject a number of hypotheses. The first two regression models results in three statistical significant dependent variables namely; “Backers”, “Updates” and “Comments”. These variables are statistical significant for all projects and for projects larger than $5000. It is notable that in both models the amount of comments has a

(26)

25 small negative influence on the success of the crowdfunding success where a positive influence was expected. Therefore H2 (the amount of comments influences the percentage of funding) has to be rejected. An explanation for this is that a comment could be positive or negative. Negative comments could influence the campaign in a negative way. Further research is necessary. The positive influence of the amount of updates on crowdfunding success is in line with research conducted by Joenssen et al. (2014). The influence of the amount of backers on crowdfunding success makes sense. If there are more backers, there is more money funded. One assumption is that all backers spend the average backing amount of $53.27.

The third and fourth regression models are to examine the influence of the variables, which possibly influence crowdfunding success, on new product success. The hypotheses H8 (the amount of backers influences the success of the final product), H9 (the duration of a crowdfunding campaign, influences the success of the final product) and H10 (the amount of levels of rewards influences the success of the final product) are rejected. The duration of a campaign and the amount of reward-levels did not have an influence on crowdfunding success either.

The third and fourth regression models fail to reject a number of hypotheses. The variables “Percentagefunded” and “Comments” are statistical significant for the independent variable “Userscore”. The amount of overfunding and the amount of comments have a positive influence on the user score of the video game. The variables “Percentagefunded”, “Comments” and “Updates” are statistical significant for the independent variable “Owners”. The amount of overfunding, the amount of comments and the amount of Kickstarter updates have a positive influence on the amount of owners of a video game (partly) funded by crowdfunding.

(27)

26 This thesis used 2 dependent variables to measure new product success. Gruner & Homburg (2000) showed there are four dimensions in measuring new product success. The user score is the best way to measure the dimension “quality of the new product”. The amount of owners is not only a result of the quality of the product. Not every owner has bought the game and stands for a success in revenue. Besides that, not every owner has paid the same price for the product. For this reason this thesis chooses the variable “Userscore” for the measurement of new product success. Therefore this thesis will also reject H6 (the amount of updates by the founder influences the success of the final product). There is no evidence found to reject H7 (the amount of comments influences the percentage of the final product) and H11 (the percentage funded influences the success of the final product).

Mollick (2014) stated that crowdfunding projects from Kickstarter usually deliver concrete products as a reward. Overfunding will result in more backers and products as a reward which will lead to more owners of the product. While the comments had a small negative influence on the success of crowdfunding, it has a small positive effect on the success of the new product. Like stated earlier comments can be positive or negative. The explanation of a positive influence is because successful campaigns probably have more positive comments. Further research is required to confirm this statement.

5.2 Conclusion

This thesis aims to bring an insight into what the influence of a successful crowdfunding campaign has on the success of the product asked funding for. At first this research tries to validate what other literature states about the success of a crowdfunding campaign.

Secondly this research explores the influence of a successful crowdfunding campaign on the success of the product asked funding for. There is statistical evidence that the success of a crowdfunding campaign (extend of overfunding) and the amount of comments during a

(28)

27 campaign, influences the user ratings of the final product. There is also evidence that more updates, comments and a higher funding percentage lead to a higher number of product owners. Finally we can conclude that crowdfunding success has a positive influence on the success of the final product. A higher funding percentage leads to a higher user rating.

Entrepreneurs should use the findings of this thesis to their advantage. The results stress the importance of building an online community when seeking funds through crowdfunding. By facilitating and building an online community and by placing regular updates for this community the entrepreneur can influence customer interaction which leads to more comments. At the same time the customer interaction in the development process influences the new product success. If this is done right, the entrepreneur can get more out of crowdfunding than just funding which makes crowdfunding even more appealing than it already is.

(29)

28

6. Limitations & future research

Like in all scientific papers there are limitations to the research conducted in this master thesis. This research only takes in account successfully crowd funded campaigns from Kickstarter and video games released on Steam. It could be possible that failed Kickstarter campaigns still got funded through another crowdfunding platform. Another possibility is that successfully funded campaigns did not intend to release their video game on the Steam platform but did release a finished game in another way.

Another limitation is the limited amount of time available to conduct this research. While most researches are conducted in more than one year this master thesis had to be written in less than half a year. When there was more time available a more in-depth research had to be conducted.

With the Kickstarter dataset this thesis could not test all factors influencing the success of a crowdfunding campaign. For example, the influence of team composition and gender could not be measured with the available data in this dataset. It is interesting to see if the results will hold when other variables are used concerning product success. While the results rejected the hypothesis about the influence on the user score, the influence on the number of product owners could not be rejected for the variables “Percentagefunded” and “Updates”.

Future research could conduct the same tests as this thesis but with recent data concerning crowdfunding campaigns. The shift of the crowdfunding campaign window for video games towards the release date could affect the chance of success of a campaign and even the success of the final product. The shift towards the release date is a conscious choice made by entrepreneurs in order to increase the chances of a successful crowdfunding campaign.

(30)

29

References

Ahlers, G. K., Cumming, D., Günther, C., & Schweizer, D. (2015). Signaling in equity crowdfunding. Entrepreneurship Theory and Practice.

Barnett, C. (2013). Top 10 crowdfunding sites for fundraising. New York: Forbes.

Belleflamme, P., Lambert, T., & Schwienbacher, A. (2013). Individual crowdfunding practices. Venture Capital, 15(4), 313-333.

Belleflamme, P., Lambert, T., & Schwienbacher, A. (2014). Crowdfunding: Tapping the right crowd. Journal of business venturing, 29(5), 585-609.

Bidaux, T. (2018). Kickstarter in 2017 – In depth look at the Games category [Blog]. Retrieved from http://icopartners.com/2018/02/kickstarter-2017-depth-look-games-category/

Cooper, R. G., & Kleinschmidt, E. J. (1995). Benchmarking the firm's critical success factors in new product development. Journal of product innovation management, 12(5), 374-391.

Cordova, A., Dolci, J., & Gianfrate, G. (2015). The Determinants of Crowdfunding Success: Evidence from Technology Projects. Procedia - Social And Behavioral Sciences, 181, 115-124. doi: 10.1016/j.sbspro.2015.04.872

Cramer, D., & Howitt, D. L. (2004). The Sage dictionary of statistics: a practical resource for

students in the social sciences. Sage.

Crosetto, P., & Regner, T. (2014). Crowdfunding: Determinants of success and funding

dynamics (No. 2014-035). Jena Economic Research Papers.

Frydrych, D., Bock, A. J., Kinder, T., & Koeck, B. (2014). Exploring entrepreneurial legitimacy in reward-based crowdfunding. Venture Capital, 16(3), 247-269.

(31)

30 Galuszka, P., & Bystrov, V. (2014). The rise of fanvestors: A study of a crowdfunding community. First Monday, 19(5). http://dx.doi.org/10.5210/fm.v19i5.4117

Gerber, E. M., Hui, J. S., & Kuo, P. Y. (2012, February). Crowdfunding: Why people are motivated to post and fund projects on crowdfunding platforms. In Proceedings of the

International Workshop on Design, Influence, and Social Technologies: Techniques, Impacts and Ethics (Vol. 2, p. 11).

Gruner, K. E., & Homburg, C. (2000). Does customer interaction enhance new product success?. Journal of business research, 49(1), 1-14.

Hemer, J. (2011). A snapshot on crowdfunding (No. R2/2011). Fraunhofer Institute for Systems and Innovation Research (ISI).

Hui, J. S., Gerber, E., & Greenberg, M. (2012). Easy money? The demands of crowdfunding work. Northwestern University, Segal Design Institute, 1-11.

Hui, J. S., Greenberg, M. D., & Gerber, E. M. (2014, February). Understanding the role of community in crowdfunding work. In Proceedings of the 17th ACM conference on Computer

supported cooperative work & social computing (pp. 62-74). ACM.

Joenssen, D., Michaelis, A., & Müllerleile, T. (2014). A Link to New Product Preannouncement: Success Factors in Crowdfunding. SSRN Electronic Journal. doi: 10.2139/ssrn.2476841

Kickstarter Stats — Kickstarter. (2017). Kickstarter.com. Retrieved 14 June 2017, from

https://www.kickstarter.com/help/stats

Koch, J. A., & Siering, M. (2015). Crowdfunding success factors: the characteristics of successfully funded projects on crowdfunding platforms.

(32)

31 Kohli, C. (1999). Signaling new product introductions: A framework explaining the timing of preannouncements. Journal of Business Research, 46(1), 45-56.

Kuppuswamy, V., & Bayus, B. L. (2015). Crowdfunding creative ideas: The dynamics of project backers in Kickstarter.

Lafferty, B., Goldsmith, R., & Newell, S. (2002). The Dual Credibility Model: The Influence of Corporate and Endorser Credibility on Attitudes and Purchase Intentions. Journal Of

Marketing Theory And Practice, 10(3), 1-11. doi: 10.1080/10696679.2002.11501916

Massolution. (2015). 2015CF Crowdfunding Industry Report.

Mollick, E. (2014). The dynamics of crowdfunding: An exploratory study. Journal of

business venturing, 29(1), 1-16.

Newzoo Games. (2017). 2017 Global Games Market Report. Newzoo Games. Retrieved from https://newzoo.com/insights/articles/the-global-games-market-will-reach-108-9-billion-in-2017-with-mobile-taking-42/

Rossi, M. (2014). The new ways to raise capital: an exploratory study of crowdfunding. International Journal of Financial Research, 5(2), p8.

Schwienbacher, A., & Larralde, B. (2010). Crowdfunding of Small Entrepreneurial Ventures. SSRN Electronic Journal. http://dx.doi.org/10.2139/ssrn.1699183

Xu, A., Yang, X., Rao, H., Fu, W. T., Huang, S. W., & Bailey, B. P. (2014, April). Show me the money!: An analysis of project updates during crowdfunding campaigns. In Proceedings

of the SIGCHI conference on human factors in computing systems (pp. 591-600). ACM.

Zackariasson, P., Styhre, A., & Wilson, T. L. (2006). Phronesis and creativity: Knowledge work in video game development. Creativity and Innovation Management, 15(4), 419-429.

(33)

32

Appendix 1

The output given by SPSS.

Descriptive statistics Kickstarter campaign status Status

Frequency Percent Valid Percent Cumulative Percent Valid canceled 8 ,9 ,9 ,9 failed 489 57,1 57,1 58,0 live 144 16,8 16,8 74,8 successful 216 25,2 25,2 100,0 Total 857 100,0 100,0

Descriptive statistics Kickstarter funded video games published on Steam Steam

Frequency Percent Valid Percent Cumulative Percent

Valid

No Kickstarter 16864 98,7 98,7 98,7

Kickstarter 223 1,3 1,3 100,0

Total 17087 100,0 100,0

Descriptive statistics funded percentage Statistics fundedpercentage N Valid 705 Missing 0 Mean 67,1354% Median 8,8750% Std. Deviation 182,63509% Range 3248,30% Minimum 0,00% Maximum 3248,30%

(34)

33 Descriptive statistics of the variables. (Model 1: Regression model with all successful and failed campaigns)

Descriptive Statistics

backers 705 0 87142 526,02 4440,486 levels 705 0 37 8,30 3,809 updates 705 0 45 5,44 7,217 comments 705 0 13850 129,32 1012,146 duration 705 4,0 91,0 38,810 15,8903 fundedpercentage 705 0,00% 3248,30% 67,1354% 182,63509% Valid N (listwise) 705 Correlations (Model 1) Correlations

fundedpercentage backers levels comments updates duration

Pearson Correlation fundedpercentage 1,000 ,260 ,095 ,193 ,314 -,043 backers ,260 1,000 ,147 ,820 ,154 -,040 levels ,095 ,147 1,000 ,228 ,323 -,008 comments ,193 ,820 ,228 1,000 ,228 -,043 updates ,314 ,154 ,323 ,228 1,000 ,067 duration -,043 -,040 -,008 -,043 ,067 1,000 Sig. (1-tailed) fundedpercentage . ,000 ,006 ,000 ,000 ,126 backers ,000 . ,000 ,000 ,000 ,142 levels ,006 ,000 . ,000 ,000 ,417 comments ,000 ,000 ,000 . ,000 ,125 updates ,000 ,000 ,000 ,000 . ,037 duration ,126 ,142 ,417 ,125 ,037 . N fundedpercentage 705 705 705 705 705 705 backers 705 705 705 705 705 705 levels 705 705 705 705 705 705 comments 705 705 705 705 705 705 updates 705 705 705 705 705 705 duration 705 705 705 705 705 705

(35)

34 Regression model summary. (Model 1)

Model Summaryd Model R R Square Adjusted R Square Std. Error of the Estimate

Change Statistics

Durbin-Watson R Square Change F Change df1 df2 Sig. F Change 1 ,314a ,099 ,097 173,51647% ,099 76,937 1 703 ,000 2 ,380b ,144 ,142 169,18203% ,046 37,483 1 702 ,000 3 ,390c ,152 ,148 168,56112% ,007 6,181 1 701 ,013 1,963

a. Predictors: (Constant), updates

b. Predictors: (Constant), updates, backers

c. Predictors: (Constant), updates, backers, comments d. Dependent Variable: fundedpercentage

ANOVA (Model 1)

ANOVAa

Model Sum of Squares df Mean Square F Sig.

1 Regression 2316424,723 1 2316424,723 76,937 ,000b Residual 21165900,823 703 30107,967 Total 23482325,546 704 2 Regression 3389289,770 2 1694644,885 59,207 ,000c Residual 20093035,776 702 28622,558 Total 23482325,546 704 3 Regression 3564917,166 3 1188305,722 41,823 ,000d Residual 19917408,380 701 28412,851 Total 23482325,546 704

a. Dependent Variable: fundedpercentage b. Predictors: (Constant), updates

c. Predictors: (Constant), updates, backers

(36)

35 Regression model 1 Coefficientsa Model Unstandardized Coefficients Standardized Coefficients

t Sig. 95,0% Confidence Interval for B

B Std. Error Beta Lower Bound Upper

Bound 1 (Constant) 23,864 8,188 2,914 ,004 7,788 39,940 updates 7,949 ,906 ,314 8,771 ,000 6,169 9,728 2 (Constant) 23,772 7,983 2,978 ,003 8,097 39,446 updates 7,106 ,894 ,281 7,946 ,000 5,350 8,861 backers ,009 ,001 ,216 6,122 ,000 ,006 ,012 3 (Constant) 22,481 7,971 2,820 ,005 6,831 38,131 updates 7,510 ,906 ,297 8,293 ,000 5,732 9,289 backers ,014 ,003 ,340 5,581 ,000 ,009 ,019 comments -,028 ,011 -,154 -2,486 ,013 -,050 -,006

a. Dependent Variable: fundedpercentage

Descriptive statistics of the variables (model 2: Projects above $5000) Descriptive Statistics Mean Std. Deviation N fundedpercentage 54,1350% 132,27642% 447 backers 806,96 5558,967 447 levels 9,13 3,998 447 comments 200,71 1266,089 447 updates 5,88 7,722 447 duration 39,150 14,3261 447

(37)

36 Correlations (Model 2)

Correlations

fundedpercentage backers levels comments updates duration

Pearson Correlation fundedpercentage 1,000 ,454 ,200 ,342 ,427 -,065 backers ,454 1,000 ,146 ,819 ,171 -,059 levels ,200 ,146 1,000 ,241 ,343 -,062 comments ,342 ,819 ,241 1,000 ,258 -,064 updates ,427 ,171 ,343 ,258 1,000 ,034 duration -,065 -,059 -,062 -,064 ,034 1,000 Sig. (1-tailed) fundedpercentage . ,000 ,000 ,000 ,000 ,086 backers ,000 . ,001 ,000 ,000 ,107 levels ,000 ,001 . ,000 ,000 ,096 comments ,000 ,000 ,000 . ,000 ,088 updates ,000 ,000 ,000 ,000 . ,234 duration ,086 ,107 ,096 ,088 ,234 . N fundedpercentage 447 447 447 447 447 447 backers 447 447 447 447 447 447 levels 447 447 447 447 447 447 comments 447 447 447 447 447 447 updates 447 447 447 447 447 447 duration 447 447 447 447 447 447

Regression model summary (Model 2)

Change Statistics

Durbin-Watson R Square Change F Change df1 df2 Sig. F Change 1 ,454a ,206 ,204 118,01216% ,206 115,333 1 445 ,000 2 ,576b ,331 ,328 108,40533% ,126 83,366 1 444 ,000 3 ,590c ,348 ,343 107,19917% ,016 11,048 1 443 ,001 1,873

a. Predictors: (Constant), backers

b. Predictors: (Constant), backers, updates

c. Predictors: (Constant), backers, updates, comments d. Dependent Variable: fundedpercentage

(38)

37 ANOVA (Model 2)

ANOVAa

a. Dependent Variable: fundedpercentage b. Predictors: (Constant), backers

c. Predictors: (Constant), backers, updates

d. Predictors: (Constant), backers, updates, comments

Regression model 2 Coefficientsa Model Unstandardized Coefficients Standardized Coefficients

t Sig. 95,0% Confidence Interval for B

B Std. Error Beta Lower Bound Upper

Bound 1 (Constant) 45,423 5,640 8,053 ,000 34,338 56,509 backers ,011 ,001 ,454 10,739 ,000 ,009 ,013 2 (Constant) 10,362 6,449 1,607 ,109 -2,312 23,037 backers ,009 ,001 ,392 9,956 ,000 ,007 ,011 updates 6,160 ,675 ,360 9,131 ,000 4,834 7,486 3 (Constant) 8,866 6,393 1,387 ,166 -3,699 21,431 backers ,014 ,002 ,573 8,557 ,000 ,011 ,017 updates 6,632 ,682 ,387 9,723 ,000 5,292 7,973 comments -,024 ,007 -,227 -3,324 ,001 -,038 -,010

(39)

38 Descriptive statistics of the variables (model 3: Successful projects above $5000)

Descriptive Statistics Mean Std. Deviation N userscore 18,92 33,915 118 backers 2953,81 10557,277 118 levels 11,42 4,924 118 comments 741,15 2389,125 118 updates 14,88 8,839 118 duration 38,127 13,7339 118 fundedpercentage 185,4740% 206,63645% 118 Correlations (model 3) Correlations

userscore backers levels comments updates duration fundedpercentage

Pearson Correlation userscore 1,000 ,394 ,179 ,434 ,173 -,124 ,396 backers ,394 1,000 ,105 ,808 ,012 -,100 ,405 levels ,179 ,105 1,000 ,247 ,094 -,005 -,034 comments ,434 ,808 ,247 1,000 ,136 -,111 ,245 updates ,173 ,012 ,094 ,136 1,000 ,117 -,006 duration -,124 -,100 -,005 -,111 ,117 1,000 -,096 fundedpercentage ,396 ,405 -,034 ,245 -,006 -,096 1,000 Sig. (1-tailed) userscore . ,000 ,026 ,000 ,031 ,090 ,000 backers ,000 . ,130 ,000 ,449 ,140 ,000 levels ,026 ,130 . ,003 ,156 ,478 ,357 comments ,000 ,000 ,003 . ,071 ,116 ,004 updates ,031 ,449 ,156 ,071 . ,103 ,474 duration ,090 ,140 ,478 ,116 ,103 . ,151 fundedpercentage ,000 ,000 ,357 ,004 ,474 ,151 . N userscore 118 118 118 118 118 118 118 backers 118 118 118 118 118 118 118 levels 118 118 118 118 118 118 118 comments 118 118 118 118 118 118 118 updates 118 118 118 118 118 118 118 duration 118 118 118 118 118 118 118 fundedpercentage 118 118 118 118 118 118 118

(40)

39 Model summary (model 3)

Model Summaryc Model R R Square Adjusted R Square Std. Error of the Estimate

Change Statistics

Durbin-Watson R Square Change F Change df1 df2 Sig. F Change 1 ,434a ,189 ,182 30,678 ,189 26,988 1 116 ,000 2 ,527b ,278 ,265 29,066 ,089 14,224 1 115 ,000 ,594

a. Predictors: (Constant), comments

b. Predictors: (Constant), comments, fundedpercentage c. Dependent Variable: userscore

ANOVA (model 3)

ANOVAa

1 Regression 25399,638 1 25399,638 26,988 ,000b Residual 109174,675 116 941,161 Total 134574,314 117 2 Regression 37416,735 2 18708,367 22,144 ,000c Residual 97157,579 115 844,849 Total 134574,314 117

a. Dependent Variable: userscore b. Predictors: (Constant), comments

c. Predictors: (Constant), comments, fundedpercentage

Regression model 3 Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. 95,0% Confidence Interval for B

B Std. Error Beta Lower

Bound Upper Bound 1 (Constant) 14,353 2,958 4,852 ,000 8,494 20,212 comments ,006 ,001 ,434 5,195 ,000 ,004 ,009 2 (Constant) 5,765 3,611 1,596 ,113 -1,388 12,918 comments ,005 ,001 ,359 4,392 ,000 ,003 ,007 fundedpercentage ,051 ,013 ,308 3,771 ,000 ,024 ,077

(41)

40 Descriptive statistics of the variables (model 4: Successful projects above $5000)

Descriptive Statistics Mean Std. Deviation N LnOwners 3,4458 5,51337 118 backers 2953,81 10557,277 118 levels 11,42 4,924 118 comments 741,15 2389,125 118 updates 14,88 8,839 118 duration 38,127 13,7339 118 fundedpercentage 185,4740% 206,63645% 118 Correlations (model 4) Correlations LnOwner s backer s level s comment s update s duratio n fundedpercentag e Pearson Correlatio n LnOwners 1,000 ,384 ,161 ,409 ,211 -,103 ,381 backers ,384 1,000 ,105 ,808 ,012 -,100 ,405 levels ,161 ,105 1,000 ,247 ,094 -,005 -,034 comments ,409 ,808 ,247 1,000 ,136 -,111 ,245 updates ,211 ,012 ,094 ,136 1,000 ,117 -,006 duration -,103 -,100 -,005 -,111 ,117 1,000 -,096 fundedpercentag e ,381 ,405 -,034 ,245 -,006 -,096 1,000 Sig. (1-tailed) LnOwners . ,000 ,041 ,000 ,011 ,134 ,000 backers ,000 . ,130 ,000 ,449 ,140 ,000 levels ,041 ,130 . ,003 ,156 ,478 ,357 comments ,000 ,000 ,003 . ,071 ,116 ,004 updates ,011 ,449 ,156 ,071 . ,103 ,474 duration ,134 ,140 ,478 ,116 ,103 . ,151 fundedpercentag e ,000 ,000 ,357 ,004 ,474 ,151 . N LnOwners 118 118 118 118 118 118 118 backers 118 118 118 118 118 118 118 levels 118 118 118 118 118 118 118 comments 118 118 118 118 118 118 118 updates 118 118 118 118 118 118 118 duration 118 118 118 118 118 118 118 fundedpercentag e 118 118 118 118 118 118 118

(42)

41 Model summary (model 4)

Change Statistics

Durbin-Watson R Square Change F Change df1 df2 Sig. F Change 1 ,409a ,167 ,160 5,05333 ,167 23,272 1 116 ,000 2 ,501b ,251 ,238 4,81192 ,084 12,931 1 115 ,000 3 ,529c ,280 ,261 4,73996 ,029 4,518 1 114 ,036 ,363

a. Predictors: (Constant), comments

b. Predictors: (Constant), comments, fundedpercentage

c. Predictors: (Constant), comments, fundedpercentage, updates d. Dependent Variable: LnOwners

ANOVA (model 4)

ANOVAa

a. Dependent Variable: LnOwners b. Predictors: (Constant), comments

c. Predictors: (Constant), comments, fundedpercentage

(43)

42 Regression model 4 Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. 95,0% Confidence Interval for B

B Std. Error Beta Lower

Bound Upper Bound 1 (Constant) 2,747 ,487 5,637 ,000 1,782 3,712 comments ,001 ,000 ,409 4,824 ,000 ,001 ,001 2 (Constant) 1,391 ,598 2,327 ,022 ,207 2,575 comments ,001 ,000 ,335 4,031 ,000 ,000 ,001 fundedpercentage ,008 ,002 ,299 3,596 ,000 ,004 ,012 3 (Constant) -,186 ,947 -,196 ,845 -2,062 1,690 comments ,001 ,000 ,311 3,750 ,000 ,000 ,001 fundedpercentage ,008 ,002 ,306 3,735 ,000 ,004 ,013 updates ,106 ,050 ,171 2,126 ,036 ,007 ,206

Crowdfunding in the gaming industry : the influence of success factors in crowdfunding on the success of the final product