“What is the influence of social proof on the consumers’ interest to back a product in a crowdfunding campaign and how does perceived campaign risk play a role in this relationship?” S.A. Vermeij (10000573) University of Amsterdam MSc. Business Administration Entrepreneurship and Innovation Track Master Thesis Final Draft First Supervisor: dr. B. (Bram) Kuijken Second supervisor: dhr. L. (Liang) Zhao Word count: 12.532 Amsterdam, June 24, 2016
Statement of Originality
This document is written by Sanne Vermeij, who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.A word of thanks,
It is usual to thank some people in this section of the acknowledgement. And to say that this is a thesis to finish my Master of Science in Business Administration, in the direction of Entrepreneurship and Innovation. And finally realizing that those words are one of your lasts in a period of lifetime. So I thought to actually use this little part as well. Thank you Bram Kuijken for helping me out when I needed it the most. Thank you Saar, Nette and Allyson for helping during the writing process. And of course all the other friends who were there while writing, and just in general. During my whole study period there has been one man always standing in my back, pushing me further. Silently I know that one day, I will realize that all his effort of letting me obtain this degree, will mean the world to. For now, before that day comes, I already will thank him here for all the support and willing to get the best out of myself. Dad, the thesis is dedicated to you, 24 juni 2016, Sanne VermeijAbstract
More and more entrepreneurs are raising funds for new projects through crowdfunding. Many of them are successful, however many crowdfunding campaigns are prone to failure as well. Potential risks are most prevalent to those who are procuring the funds. Therefore, it is important to shed more light on the risks involved with crowdfunding. This paper aims to predict product interest in crowdfunding campaigns by using two different forms of social proof: peer social proof and expert social proof. It is researched whether the presence of social proof, taking the amount of campaign risk perceived into account as a mediator, determines the amount of product interest a backer (potential funder) has. Also, the role of risk propensity and perceived campaign trust have been taken into account. Hypotheses have been tested through an experiment, consisting of four treatments randomly shown to 230 respondents. The results show that perceived campaign risk is negatively related to product interest. The use of experts in crowdfunding campaigns decreases the amount of campaign risk people perceive. These findings contribute to understanding the role of risk in crowdfunding campaigns in that we have established a link between perceived campaign risk and interest in the crowdfunded product. This knowledge could serve as a foundation for further development regarding the role of perceived risk in crowdfunding campaigns. Keywords: Crowdfunding, risk, social proof, perceived campaign risk, risk propensity and perceived campaign trustTable of content
1. Introduction……….. 6 2. Literature Review…………. 9 2.1 Crowdfunding ………… 9 2.2 Social proof theory…………. 11 3. Theoretical Framework ……….... 13 3.1 Crowdfunding campaigns and social proof…………. 13 3.2 Perceived campaign risk ………... 18 3.3 Perceived campaign trust …………... 22 3.4 Risk propensity………….... 23 3.5 Conceptual model …………... 25 4. Methodology…………. 25 4.1 Research design………….. 25 4.2 Research population sample………….... 26 4.3 Measures ………...… 27 4.3.1 Control variables…………..… 27 4.3.1 Dependent variables…………... 28 5. Results…………...… 30 5.1 Descriptive statistics…………..…. 30 5.2 Treatments and product interest………….. 31 5.3 Normality, kurtosis and skewness…………..… 31 5.4 Correlations…………..31 5.4.1 General perceived CF risk………….... 32 5.4.2 Product risk and Financial risk…………... 33 5.4.3 Product interest & perceived campaign trust …………... 34 5.5 Hypotheses………….… 35 5.5.1 Mediating role of perceived campaign risk………… 35 5.5.2 Mediation role of perceived campaign trust …………... 38 5.5.3 Moderation of Risk propensity ………… 39 6. Discussion ………. 41 6.1 Social proof and product interest …………... 426.2 Perceived campaign risk …………... 43 6.3 Perceived campaign trust ………….. 44 6.4 Risk propensity………….... 45 6.5 Limitations…………... 45 6.6 Future research………… 47 6.7 Managerial implications …………48 7. Conclusion ………49 8. References ……….. 50 9. Appendix………… 57 9.1 Appendix A Questionnaire…………... 57 9.2 Appendix B The four different treatments………… 62 9.2.1 Treatment 1………….. 62 9.2.2 Treatment 2………… 63 9.2.3 Treatment 3…………. 64 9.2.4 Treatment 4………… 65
1. Introduction
Independent startups often have great difficulty obtaining external financing (Cosh, Cumming & Hughes, 2009). Crowdfunding is a new form of financing whereby many small businesses were able to get financing for their risky ideas (Schwienbacher & Larralde, 2010). A large number of unknown capital providers (“the crowd”) decides if a particular idea has potential by funding small amounts of money (Belleflamme, Lambert & Schwienbacher, 2014). Aside from getting money to execute the idea, crowdfunding offers other benefits, for example, introducing and testing an initial idea (Mollick, 2014). As crowdfunding grows, the amount of failed campaigns grows along with it. An example of such a campaign is the Kickstarter campaign for the small drone called the Zano. The campaign became a big success; more than 12.000 backers contributed, resulting in a total amount of $3.4 million raised. Nothing actually happened afterwards and one year later the company announced bankruptcy. The backers never regained the money they invested. The reason why this campaign failed, was that the creators of the campaign did not have the background, experience or knowledge which was needed to get the drone on the market. Until now, there haven’t been many court cases involving crowdfunding, legal experts in consumer issues say that this is because crowdfunding is a relatively new concept (Loria, June 2016). There are more examples of campaigns that failed, and this amount is still growing. Backers need to be protected as they are the ones who are most at risk, being that their money is utilized. Therefore, it is important to shed more light on the role of risks in crowdfunding campaigns. There are various risks assumed by potential backers, one of the main risks stems from the fact that a majority of the campaigns deliver the promised rewards later than expected, or not at all (Mollick, 2014). Another risk could be that the campaign creator uses the funds for different purposes or as aforementioned, the creator has no businessexpertise (Hui, Greenberg & Gerber, 2014). These risks represent a reason for backers to lose their interest in the campaign and, as a consequence, not fund the project. Therefore, it is most important for a creator to know whether the amount of perceived campaign risk influences the amount of interest in the product. This research focused only on reward based crowdfunding, as it is the most prevalent model (Mollick, 2014). This form of crowdfunding can be seen as a form of preselling (Ahlers et al, 2012), therefore it is comparable with online purchase transactions. The importance of researching online risk has been emphasized by Miyazaki and Fernandez (2001) who found that perceived risk mediates the relationship between internet experience and online purchase. Aside from providing creators with more information about the perception of risk from backers, legal issues might benefit as well because the position backers have in reference to the creators is strengthened. Even for the crowdfunding platforms it is interesting, because they serve to improve the usability of crowdfunding and therefore minimize the amount of risks. So, the entire crowdfunding community could profit from the results of this study. As crowdfunding is becoming more and more relevant, research is constantly conducted to further understand this concept. If people are uncertain about the value of a course of action, they tend to follow “the crowd” (Cialdini, 1993). So when the perceived risk in a crowdfunding campaign is high, then they might base the decision to fund or not, on the actions of others. In addition, this research shows that if the crowdfunding campaign provides information on previous backers, this could be a sign of quality for potential backers
(Herzenstein, Dholakia & Andrews, 2011; Zhang & Liu 2012; Burtch, Ghose & Wattal, 2013; Smith, Windmeijer & Wright, 2015). At present, no research has been conducted with regards to this peer backing behaviour in combination with perceived campaign risk and
perceived campaign trust, as far as the researches know. Only Zheng, Hung, Qi, and Xu (2016) conclude that trust explains the fundraising performance, however they stress the need for further research combining trust with other variables to increasingly explain the fundraising performance. Also, Kuppuswamy and Bayus (2015) conclude their research by stating that further research is required to examine whether goalgradient behaviour generalizes across platforms. This research aims to expand what is known about this topic by using the concept of social proof (Cialdini, 1993). Social proof is people's tendency to trust something more because other people do, too. The choice is made to split this theory up into two different forms: peer social proof (showing other backers contributions) and expert social proof (showing the brands of experts magazines in crowdfunding campaigns). In this paper, it is argued that the amount of perceived campaign risk mediates the relationship between social proof and product interest. Furthermore, risk propensity has been taken into account, this concept represents the degree to which an entity is willing to take chances with respect to risk of loss. Also perceived campaign trust, a concept closely related to perceived campaign risk in online purchases has been taken into account. This research makes use of an experiment. Four different versions of the same crowdfunding campaign are randomly shown to people who are interested in crowdfunding. Each of the versions shows a form of social proof, or no social proof at all. A survey questionnaire, showing the crowdfunding campaigns and questions related to the research, was administered digitally. This to provide a more clear and reliable validation of the model and its subvariables. The aim of this study is to expand literature about the role of social proof and the role of perceived risk of crowdfunding campaigns. The research question is:
“What is the influence of social proof on the consumers’ interest to back a product in a crowdfunding campaign and how does perceived campaign risk play a role in this relationship?” The paper is organized as follows. The next section deals with all the relevant literature around the main concepts used in this research. Next, the theoretical framework section explains what the hypotheses tested in this research are. After that, the methodology is accounted for. In the following section, the results are presented and discussed to concisely answer the research question. Lastly, the conclusion functions to sum up all findings.
2. Literature review
This chapter sheds light on the two main concepts this research revolves around; crowdfunding and the social proof theory. To clear up what these concepts mean, existing literature on the topic is reviewed. 2.1 Crowdfunding Crowdfunding is derived from the broader concept of crowdsourcing. Jeff Howe and Mark Robinson first used the term crowdsourcing in the June 2006 issue of Wired Magazine, an American magazine for high technology. Howe describes crowdsourcing as “the process by which the power of the many can be leveraged to accomplish feats that were once the province of the specialized few (Howe, 2008) .” According to Brabham (2008), the efficiency of crowds in solving problems of companies is related to its composition; the more diverse it is, the more efficient it can be. Earlier, Lévy (1997) has mentioned the process as “a collective intelligence”: ‘no one knows everything, everyone knows something, and all
knowledge resides in humanity’. According to Kleemann, Voß and Rieder (2008), companies make use of the crowd mainly for cost reduction reasons. Crowdfunding can be viewed as an element of crowdsourcing. In the case of crowdfunding, the objective is to collect money for investment, generally by using online social networks (Belleflamme, Lambert & Schwienbacher, 2014). There are many definitions of the concept crowdfunding. Kleemann et al. (2008) describe it 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”. Crowdfunding occurs without any intermediary; entrepreneurs ‘tap the crowd’ by raising the money directly from individuals. The typical mode of communication is through the internet (Belleflamme, et al, 2014). The type of reward that the crowd henceforth receives is not uniform, there are many business models for it. The study of Belleflamme, Lambert and Schwienbacher (2010) distinguishes between donations, passive investments and active investments by the crowd. Donations occur when backers finance the project without sharing the profits with the entrepreneur. Passive investment is a type where there is some form of a reward given to the crowd of investors. Those rewards can take various forms, such as a preorder of a CD. Lastly, active investment offers investors a chance to become active within the initiative, next to offering rewards to them. There can be several positive outcomes from this last type, entrepreneurs can expand their resources through their investors. Also, the project can get valuable feedback about what the market prefers the most. There are many factors that can predict whether a crowdfunding campaign will be a success or not. Mollick (2014) has done research about this and deduced that the goal size is negatively associated with success. Also, duration decreases the chances of success; the
reason for this might be that the creator has a lack of confidence. Schwienbacher and Larralde (2010) conducted research about the the successfactors; having an interesting project, the willingness to extend the skillset and knowhow concerning the controls of Web 2.0. On the other hand, the most common reason for failure for Kickstarter campaigns is that creators are not able to reach out to the right investors (An, Quercia & Crowcroft, 2014). Thus, managing interests is extremely important for the success of a crowdfunding campaign. Investors and creators have different motivations for participating in crowdfunding. Gerber and Hui (2013) found three main principles concerning these motivations: supporting resource exchange, supporting the community before, during and after and to provide transparency. For creators the main motivations were: raising funds, establish relationships, improve perceptions of their own abilities, replicate successful experiences of others, acquiring new skills and expand awareness. Backers are motivated to fund crowdfunding campaigns to collect rewards, support creators causes and engage and contribute to a trusting and creative community. 2.2 Social proof theory Imitation follows from the heuristic of social proof, that is, looking to the actions of others for clues as to what constitutes appropriate action (Cialdini, 1993). Cialdini (1993) found six principles of persuasion: reciprocity, commitment/consistency, authority, liking, scarcity and social proof. Cialdini (1993) defines the concept of social proof as follows: “if a lot of people are doing the same thing, they must know something we don't.” People tend to have an enormous amount of trust in others if they are uncertain. But often the crowd is mistaken because every single person in there is reacting to the principle of social proof.
Cialdini (1993) organized an experiment on social proof, he held a doortodoor campaign and asked if residents wanted to support a certain campaign. Cialdini found that if the list with prior donors was longer, the next person who was asked to support, was more likely to donate as well. Also, if this list contained names of people (friends or neighbours) the supposed donator knew, then the effect was even more pronounced. The effect of the principle is more strongly pronounced in two scenarios. Firstly, when the decisionmakers are not certain about the value of a course of action. Secondly, when decisionmakers are able to observe what others did before they have to make that decision (Cialdini, 1993). Different experiments have been held by sociologists and psychologists, empirically proofing this behaviour (Allen, 1965; Asch,1956; Bearden & Etzel, 1982, in Huang & Chen, 2006). Also Park and Lessig (1977) found that consumers use the judgments of others as a sign of product quality. So indeed, people follow each other when purchasing. Interesting is that “previous research has shown that people imitate others out of a desire not only to be accepted but also to be safe” (Huang & Chen, 2006, p. 414). Huang and Chen (2006) found that sales volume and customer reviews influence the decision of the customer. They explained this with two forms of social influence; normative and informational influence. Normative influence is about the imitation of people to meet the expectations of others. Informal influence is taking information that is received from others as an indication of reality. Following the direction of others in decisionmaking can be very efficient, as one does not need to acquire a lot of information over a longer period of time by him or herself. However, Zhang and Liu (2012) explain that the outcome can be less optimal. When individuals do not inform themselves properly, but only rely on others, this could lead to informational cascades. They make the distinction between irrational and rational herding.
They say that there are different degrees in following others, one is more passively mimicking peers (irrational) and the other is that people actively and observationally learn from others (rational).
3. Theoretical Framework
In the theoretical framework, the main objective is to provide as much information on the research topic as possible, based on research that has already been conducted. The aforementioned main concepts, crowdfunding and social proof, are linked together. Perceived campaign risk, the role of risk propensity and perceived campaign trust are taken into account because of their mediating and moderating effects.
3.1 Crowdfunding campaigns and social proof Crowdfunding campaigns make often use of references to other backers who already invested. It is visible what other peer backers already funded and the amount of money already pledged is stated. This is why there has been chosen to explain this behaviour with the theory of social proof Cialdini (1993). As this theory says that the effect of the principle is most strong when people are not certain about the course of their actions. Also, people make use of other judgements as a sign of product quality (Park & Lessig, 1977). And sales volume and customer reviews influence the decision of the consumer (Huang & Chen, 2006). Kuppuswamy and Bayus (2015) did research to understand the backer dynamics over the project funding cycle. They found that backers are more likely to fund in the first week and the last week of the campaign. A Ushaped pattern of support is pervasive across projects (Kuppuswamy & Bayus, 2015), counting for both successfully and unsuccessfully funded projects, those with large and small goals, and projects in different categories. In the end
phase, the creators rapidly increase the amount of updates they post, this development results in a rapid increase in support as projects come to an end. Especially in the final stage of the crowdfunding project, have a positive influence on achieving the project goal because they awaken emotions and excitement from backers. Xu, Yang, Rao, Fu, Huang and Bailey (2014) and Antonenko, Lee and Kleinheksel (2014) also found that posting more updates, or solid communication with backers in general predictive in success. Gerber, Hui and Kuo (2012) found that funders are motivated to fund campaigns for different reasons, one of them is engaging and contributing to a trusting and creative community. Hui, Greenberg and Gerber (2014) did further research to the role of community in crowdfunding campaigns. They found that creators make use of community efforts to build an audience and to spread the word. Crowdfunding creates new social interactions and this motivates the crowd to participate in funding projects (Gerber, Hui & Kuo, 2012). Belleflamme et al. (2014) also conclude that the entrepreneur needs to build a community of individuals with whom it is very important to have interaction. Agrawal, Catalini and Goldfarb (2011) state that early investment serves as a signal of entrepreneurial commitment, like friends and family. Also this in line with the theory of social proof, as they found that people are more willing to fund if they know the people already funded. Investors in later stages use these early investments, as a signal to also fund. Additionally, Fishbach, Henderson, and Koo (2011) found that the identification with a group plays a big role. When the group identification is relatively weak, for example when the crowdfunding communities have many anonymous members, individuals decide to pursue a shared group goal if they think the aim of the project is worth it. Additionally, Colombo et al. (2015) empirically confirmed the importance of presence of other backers in the early stage of the project. Mostly when the quality of the product is
unclear, backers find themselves relying on others to encourage themselves to donate. This is in line with the expectations of the theory of social proof (Cialdini, 1993). The information about the amount of backers and the money pledged are indicators of interest and therefore highlighted by the platforms. An interesting phenomenon with Kickstarter is one that involves social influence within the community. Commonly known as the “Kickstarter Effect,” as a project approaches its goal the activity around the campaign is increasing a lot and this pushes the campaign over its target (Galinsky, 2010; Nelson, 2013 in Kuppuswamy & Bayus, 2015). Also Herzenstein, et al. (2011) found that the contributions of others are influencing the decisions of a backer. They did research in lending crowdfunding, they called it strategic herding behaviour by lenders. Lenders are more willing to bid on an auction with a greater number of bids. Social proof, meaning following a crowd because people perceive them to know more things than they do themselves, has been researched on two different scales. This research focuses on two different forms of social proof; ‘Peer Social Proof’ and ‘Expert Social proof’. This distinction has been made based on the research of Wijnberg and Gemser (2000). They did research about the selection systems in visual art and described different types of selection, so also peer selection and expert selection. They describe it as follows: “Peer selection is when the selectors and the selected are part of the same group” (Debackere et al. 1994 in Wijnberg & Gemser, 2000 ). In this paper this distinction is also utilized, however not in the purpose of selecting visual art but selecting a specific crowdfunding campaign; to fund or not. This theory suits the research topic because crowdfunding is also very competitive, with many creators and many backers available. In crowdfunding campaigns, intensive use is made of tools to display backers that are already involved in the campaign, this will be referred to as “peer social
proof”. The actual effect of peer social proof is measured through the variable ‘product interest’. The following hypothesis is stated: Hypothesis 1a: Peer social proof has a positive relationship with consumers’ product interest. The second form of social proof is “Expert Social Proof”. Wijnberg and Gemser (2000) described expert selection as when: “the selectors are not producers or consumers, but experts who have the power to shape selection by virtue of specialized knowledge and distinctive abilities.” This could also be a form of social proof as the many are following one expert, who they perceive as knowing more than they do, and they just blindly follow. Research has been conducted concerning linking brands to other brands to ensure the credibility of it. Also in crowdfunding, creators use references to brands to increase people’s interest in the products offered in the campaign. Most of the time, these brands are correlated with online magazines who write articles to recommend certain products. The brands of those magazines have a distinguished reputation and consistently offer an ‘expert point of view’ that many people may feel close associations with. This can be seen as a form of social proof, as external arguments get valued over one’s own opinion. When consumers are uncertain about product attributes, firms make use of brands (Erdem & Swait, 1998). A brand can be defined as "a name, term, sign, symbol, or design, or combination of them which is intended to identify the goods and services of one seller or group of sellers and to differentiate them from those of competitors" (Kotler 1991; p. 442). Showing a brand name could signal intangible product properties, this functions as a shortcut, since otherwise the only way to learn is through experience or just by accepting in faith
(Erdem & Swait, 1998). When offering a product in a highly competitive market, marketers must often link their product or brand to other entities to improve brand equity (Keller, 2003). Brand equity is defined as the added value a brand gives a product (Farquhar, 1989). Entities a brand is linked to can be places (country of origin), things (events), people (endorsers) and, most relevant for this paper, to other brands (alliances) (Keller, 2005). It is important to make the right decisions concerning what brand elements are the most ideal to present your product in the best way in the minds of the consumers. Therefore many marketers are employed to create the best brand equity (Keller, 2005). In crowdfunding, the process of making the best decisions about which brand entities to use is sometimes difficult. Many creators of campaigns have great ideas but the execution of those ideas and bringing them to market is difficult. For example, marketing skills to spread the word of the product to the right target group (Hui et al, 2014). The choice the creator is making about which expert recommendation they show is important. It is the question whether it is the right expert to increase the interest in your product. The competitive level is high, as there are many campaigns available on crowdfunding platforms. Linking to other brands by showing expert magazines might help campaigns to differentiate themselves from others (Keller, 2003). Therefore the effect of showing other experts, referred to as experts in this case, might improve the interest backers have in the products offered in crowdfunding campaign. The following hypothesis is stated: Hypothesis 1b: Expert social proof has a positive relationship with consumers’ product interest.
3.2 Perceived campaign risk The amount of risk people perceive is widely recognized as the main source of negative influence in the behaviour of consumers in ecommerce (Featherman & Pavlou, 2003; Pavlou, 2003; Kim, Ferrin & Rao, 2008). An increasing risk for backers could lead them to rescind interest in backing the project. Thus, it is important for a crowdfunding creator to know what risks the backer perceives, in order to encourage future backers by making them feel more secure about investing in the project. Half a century ago, the concept perceived risk was already introduced to consumer behaviour research by Bauer (1967, p.24): “Consumer behaviour involves risk in the sense that any action of a consumer will produce consequences which he cannot anticipate with anything approximating certainty, and some of which at least are likely to be unpleasant”. This definition has been used consistently by other researchers to explain the concept. Based on risks, consumers anticipate and make decisions on whether to move forward with a purchase or not. Interesting is that, when used in online purchase literature, the concept perceived risk is often related to the concept trust (Kim et al, 2008; Van der Heijden, Verhagen & Creemers, 2003; Chang & Chen, 2008). Deutsch (1958) was the first researcher who identified that risk is a precondition for the need of trust. Trust is only applicable in situations where risk is involved. If there are no doubts about the outcome, there is no risk, and therefore no need for trust. Van der Heijden et al. (2003) also found this result, but in the area of ecommerce; perceived risk is an antecedent for trust in online stores. Also, Chang and Chen (2008) found that trust and perceived risk are reciprocal. Ferrin and Rao (2008) have studied the concept perceived risk in relation to trust and perceived benefit. The perceived benefit is what drives the consumer to purchase a product
online (Peter & Tarpey, 1975). Peter and Tarpey (1975) explain this with the valence framework, which describes three different strategies in terms of how consumers make decisions. These are minimization of expected negative utility (perceived risk), maximization of expected positive utility (perceived return) and maximization of expected net utility (net perceived return). Consumers make decisions based on these strategies to maximize the net valence resulting from either the negative or positive outcome of their observation. In this research, based on some of the conclusions of researches revolving around online purchases, theory is now transferred to the behaviour of backers in crowdfunding campaigns. Crowdfunding purchases can be compared to online purchases in the sense that both practices involve risks that can be exchangeable. 3.2.1 Perceived crowdfunding campaign risk There are many different risks to be experienced in crowdfunding campaigns. The focus in this paper is the risk a backer of a crowdfund campaign might harm. There are different risks which can be perceived in crowdfunding campaigns, in this paper lack of physical contact with the creator, creditworthiness of the creator, product is not delivered on time, information asymmetry, payment risk and last the social risk perceived are further explained. The first risk a backer could perceive in a crowdfunding campaign is that there is a lack of physical contact with the creator. Most research refers contact with the creator to the concept trust, as creating trust online is very difficult because facetoface contact is impossible and it is not easy to replace this in an electronic environment (Papadopoulou, Andreou, Kanellis & Martakos, 2001). In crowdfunding campaigns, direct contact between the creator and the backer is important. Various researches have been conducted about this. Zheng et al. (2016) found that the more entrepreneursponsor interaction exists, the more
trust in a crowdfunding project is generated. Antonenko et al. (2014) point out that intensive communication positively impacts successful projects on the project website, as well as reacting promptly to questions, posting own questions, and providing frequent status updates. Hui et al. (2014) state that this personal contact with the backers is something that often goes wrong, as creators find themselves not able to react to all the hundreds, sometimes thousands of backers. The researchers state that it is overwhelming and that creators cannot engage in as much contact with backers as they would like.
Another perceived campaign risk which can be found is that backers find creators not creditworthy. Marom, Robb, and Sade (2015) and Gorbatai and Nelson (2015) have found that the characteristics of creators of crowdfunding campaigns is determining the success of the campaign. According to Boeuf, Darveau and Legoux (2014) announcements of personal information involving the entrepreneur (project owner), along with personal pictures, are considered positive due to the higher trust and serious support from backers this results in. Additionally, Colombo, Franzoni and RossiLamastra (2015) found that a picture of the project creator can boost the probability of successful projects. To conclude and to explain this perceived campaign risk, Herzenstein, et al. (2011) and Zhang and Liu (2012) have found that one of the success factors of crowdfunding campaigns is creditworthiness.
Another perceived campaign risk that is perceived a lot in crowdfunding campaigns is that the product is not delivered on time. Mollick (2014) finds that over 75% of the products are delivered later than expected. As stated before, Hui et al. (2014) held interviews with creators and this revealed the crowdfunding process to be overwhelming. They find that many creators underestimate the work involved and find themselves overwhelmed with the responsibilities of coordinating the whole process. Some creators have less business, marketing and management experience (Hui et al. 2014). Shane, Khurana and Hall (1999)
found that the amount of experience an entrepreneur has does have an effect on the success of the started firm. This makes even more sense if a creator is for example only good at
conceiving of a great idea but not at executing it in the right way.
Information asymmetry concerning the transaction process has also been highlighted as a campaign risk perceived by a backer. As the creator of the campaign has more
information than the backer has (Pavlou & Gefen, 2004). Crowdfunding initiatives are about the introduction of a nonexistent product or service; people do not know what to expect. Thus, the perceived risk is greater when the information asymmetry is very prevalent (Belleflamme et al, 2014). Most of the time the entrepreneur only offers a short description and promises a certain product in the final stage (Nocke, Peitz, & Rosar, 2011; Belleflamme et al, 2014), it is not known whether it is possible to actually produce the product promised. The true quality of the product is only revealed later, so the backer deals with a lot of uncertainty and therefore risk. When the backer has chosen a product in the end, the payment is often executed by credit card. Multitudes of research has been conducted and gathered about internet purchasing, as it presents numerous risks for consumers over and above the transaction process itself being perceived as risky (Einwiller, Geissler, & Will, 2000; Einwiller & Will, 2001; GrabnerKrauter & Kaluscha, 2003). Shannon (1998) found that providing credit card information to an online business, which has no physical location increases the amount of risk a customer perceives. As crowdfunding presents the opportunity of reaching out to many different backers all over the world, the geographical distance between creator and the backer is most of the time big (Agrawal et al, 2011), and thus the currencies will also differ. So in crowdfunding, the amount of payment risk perceived might be even bigger than in normal online purchasing decisions.
Next to these campaign risks as explained above, there is also perceived social risk. This is explained by Featherman and Pavlou (2003, p. 455) and is “the loss of status in one’s social group as a result of adopting a product or service, looking foolish or untrendy”. In crowdfunding this might also occur, as it much related to social media, and people are able to see each other when funding a project or not. Crowdfunding campaigns are many times also available on social media, and Lu, Xie, Kong and Yu (2014) even found that the use of social media plays a positive role in crowdfunding campaigns. So this social risk might play a big role as well in this case. There are many different kind or risks to perceive in crowdfunding. As Miyazaki and Fernandez (2001) have found that perceived risk mediates the relationship between internet experience and online purchase, it is expected that it might also be the case in crowdfunding. This perceived campaign risk will be measured on campaign level. The existing research about the role of perceived risk in crowdfunding campaigns is not extensive. Therefore, it is taken as an overall feeling, instead of already specifying it to much. The following hypothesis is expected: Hypothesis 2: The positive relationship between social proof and product interest is mediated by perceived campaign risk 3.3 Perceived campaign trust As before mentioned and demonstrated by literature, trust plays a big role in online purchases in combination with perceived risk (Kim et al, 2008; Van der Heijden et al, 2003; Chang & Chen, 2008). Therefore, it has also been checked whether perceived campaign trust mediates the relationship of social proof and the amount of interest a backer has in the product. Trust in
and of itself is studied and there it has been established that trust is a common basis for monetary transactions (Gefen, 2000) and it is one of the most crucial factors for success of an online enterprise (Beldad, De Jong & Steehouder, 2010). Risk is a precondition of the need for trust (Deutsch, 1958). Van der Heijden et al. (2003) confirm this relation with their finding that trust has a negative relation with perceived risk. Both concepts directly influence the attitude towards purchasing online (Van der Heijden et al, 2003). So these two closeby concepts might influence the online purchase process of consumers in different ways. Creating loyalty is important in the branche of online purchases as there are so many choices available; there is a lot of competition in the online world. Trust in online purchases appears to be especially important for creating loyalty when the perceived level of risk is high (Anderson & Srinivasan, 2003). Zheng et al. (2016) researched the role of trust management in reward based crowdfunding campaigns and established that trust management significantly positively increases the fundraising performance. So trust might play a role in the model proposed in this paper as well, and in specific combination with perceived campaign risk. Therefore, it is expected that perceived campaign trust plays a role in the amount of risk someone might perceive from a crowdfunding campaign. The following hypothesis is stated as follows: Hypothesis 3: The positive relationship between social proof and product interest is mediated by perceived campaign trust. 3.4 Risk propensity The way people are able to react to risk is interesting as it might influence the way people perceive risk. Therefore, this is also measured in this paper. People differ in the way they deal
with risks (Lion, Meertens & Bot, 2002). Some are more willing to take risks than others, and thus influences decision making. Weber (2001) developed a riskvalue framework, this framework is based on domains. There are five contained domains where risk taking can play a role: financial decisions, health/safety, recreational, ethical and social decisions (Weber, 2001). The amount of risk individuals perceive is highly domain specific. Another interesting detail is that women appear to be more riskaverse in domains except in the social risk (Weber, 2001). Kahneman and Tversky (1979) studied risk taking behaviour and developed the prospect theory, which is a theory that describes the decision making process under risk. They say that risk taking depends on the way the risky situation is actually framed. It explains that individuals who show big risk seeking behaviour are restricted by the domain of gains and that risk aversion is restricted by the domain of losses. As explained earlier, making use of the valence framework of Peter and Tarpey (1975) is interesting to see what role risk propensity plays in how much campaign risk people perceive. On the other hand, the valence framework also helps determine how much benefit they perceive, with the outcome maximizing the net valence. Here, it could be expected that the ones who perceive more risk in campaigns are also the ones who avoid risks. Thus, if a person finds crowdfunding risky and this person is an avid risk avoider, so very low on risk propensity, this might explain why they perceive the campaign as very precarious. Therefore the following hypothesis is also tested: Hypothesis 4: The positive relationship between social proof and perceived campaign risk is moderated by risk propensity, so that this relationship is stronger for lower values of risk propensity.
3.5 Conceptual model Figure 1: Conceptual model
4. Methodology
In this chapter the research design, research population sample and measures are discussed. The chapter of measures is divided in two different chapters, namely control variables and dependent variables. Per variable there is explained why and how they are asked. 4.1 Research design The aim of the paper is to see whether social proof determines the amount of interest a funder has in a product, and what role perceived campaign risk plays in this relationship. In order to test this, research following the design of a quasi experiment was carried out. This experiment was chosen because it enables the researcher to manipulate certain aspects that might directly affect the product interest. As crowdfunding is a relatively new concept mostly qualitative researches have been conducted (Hui et al, 2014; Xu et al, 2014; Gerber et al,2012). This study aims to provide more insights in the actual decision making process by conducting an experiment and using different crowdfunding models. Hopefully, this study contributes in such a way that other researchers can use it as a building block. The data of the experiment is collected via Qualtrics (Qualtrics.com). Qualtrics is an online tool for creating and distributing surveys. There were four treatments in which respondents got to see four versions of one randomly chosen crowdfunding campaign (see appendix B). The product represented in the crowdfunding campaign is an organic sun paste. This product was chosen due to the expectation that it might be a product most people are equally interested in. Peer social proof has been measured by showing a version of the campaign that displayed no other involved backers involved, another version showed that 445 backers already funded. Expert social proof has been measured by showing experts in one version, and erasing them in the other version (see appendix B). 4.2 Research population sample The population interesting for this study is every person who has ever funded or thought about funding a crowdfunding campaign. The research is conducted using a nonprobability convenience sample, as the population is big and the sampling frame is unknown. Respondents have been contacted through personal email and Facebook. The respondents who joined the survey were able to win certain prizes (e.g. surf lessons or breakfast in bed) in order to encourage them to submit their answers to the questionnaire. To create a good sample the demographic information (age, gender and educational background) were kept in mind. The amount of respondents ultimately reached was 230, this amount facilitated a valid analysis. The dataset was checked on the completeness. Individuals that did not fill in most of the survey were removed from the database because these would
not provide accurate representative data. After removing these individuals, the sample size was 192 respondents. There were four different treatments and all respondents were randomly assigned to them (see figure 2). Table 1: Descriptive table of the sample 4.3 Measures An online questionnaire was issued to collect the data needed to test the hypotheses. All measures used in this research involved individual responses to the questions of the surveys. 4.3.1 Control variables The questionnaire asks respondents for their demographics: age, gender (man/ woman), country of residence and education background. Those demographics are requested because these could also determine outcomes, as Weber (2001) found for example, that women
appear to be more riskaverse than men. Respondents were asked for their experiences with crowdfunding in the past, as An et al. (2014) found that the past crowdfunding history matter in the funding behaviour of backers. The question also asked was, what the amount of risk backers perceived in general with crowdfunding, as it was expected to be thought of as a potential reason why people perceive risk in this specific crowdfunding campaign.
Based on the specific risks crowdfunding campaigns can involve, specific questions pertaining to risk were asked. The first questions was about product risk, as some creators have minimal business, marketing and management experience (Hui et al, 2014). The question was asked how much product risk (i.e. not working, defective products) they perceived. And the second question was about financial risk, as Shannon (1998) found that providing credit card information to an online business, which has no physical location increases the amount of risk a customer perceives. It was asked how much financial risk (i.e. fraud, hard to return the money) they perceived. 4.3.2 Dependent variables The following scales have been created to measure the variables from the theoretical framework:
Perceived campaign risk. Based on Kim et al. (2008) see Appendix A. The scale used to measure the construct is the existing and validated Likert scale, used on a 7point scale (completely disagree – completely agree) at interval level. Cronbachs alpha = 0.88 (8 items)
Product Interest. Based on Gerber et al. (2013), see Appendix A. This is used to establish whether the respondent is motivated to fund the project or not. The scale used to measure the construct is also the existing and validated Likert scale, used on a
7point scale (completely disagree – completely agree) at interval level. Cronbachs alpha = 0.86 (3 items)
Risk propensity. Based on Meertens and Lion (2008). Measuring an individual's tendency to take risks. The scale used to measure the construct is also the existing and validated Likert scale, used on a 7point scale (completely disagree – completely agree) at interval level. Cronbachs alpha = 0.78 (7 items)
Perceived campaign trust. Based on Zheng et al. (2016), see Appendix A. As
crowdfunding is quite a new concept, it is hard to use already validated scales for the dependent variables “perceived campaign trust” and “perceived campaign risk”. Much research is available about online purchases (Kim et al, 2008), the statements used are based on this research. The scale used to measure the construct is the existing and validated Likert scale, used on a 7point scale (completely disagree – completely agree) at interval level. Cronbachs alpha = 0.76 (3 items).
5. Results
In this section, the results of the empirical study are presented. This paper aims to find a difference in interest shown towards a crowdfunding product by randomly showing four different conditions of social proof. It is expected that social proof increases the interest in the product, as people follow the crowd, and have more interest in a product, if other people also have this. 5.1 Descriptive statistics In this experiment, 230 respondents were asked to scan a crowdfunding campaign and see whether they were interested in product or not. Based on the excluding criteria, one participant was excluded based on age, 38 respondents were excluded due to missing data. 191 respondents were left, 87 female (45,5%) and 104 male (54,5%). The population of the Netherlands where there are 49,5% male and 50.5% female (CBS, 2015). Comparing those results, you can say that the gender distribution is fairly coherent. Only looking at the age of the respondents, and finding an average age of 25 years old, this is not a good representation of the Dutch society as the average age is 39 (CBS, 2015). Also, there are 174 Dutch respondents (91.1%) and only 17 from other countries (8,9%). The explanation of this sample is that the respondents were recruited through a snowball sampling, among Dutch students, so this way of recruiting might explain the descriptives of the sample. However, PWC published some demographics of crowdfunding users and this says that 64% are men and 36% are female users. Also the average age of the crowdfunder is between 25 and 34 (Source: ChoiceLoans). So these demographics will explain why the low age of 25 is still kind of generalizable and the higher amount of men involved in this sample also.5.2 Treatments and interest in the product After controlling for the demographics of the sample, it is needed to review the descriptive statistics of the data collected. The 191 respondents were randomly assigned to the four treatments: Treatment 1 (no peers/no experts): N=48, Treatment 2 (no peers/experts): N=53, Treatment 3 (peers/no experts): N=46 and Treatment 4 (peers/experts): N=44. The minimum respondents per treatments was around the 50, to make it applicable for statistical research (Saunders, 2011). The following table will present the descriptives per treatment. Table 2: Descriptive statistics per treatment When reviewing those numbers, some things were noticed. Treatment 2 has a fairly high minimum compared to the other variables of 2.33. On average the Mean of all the variables are quite high on a scale of 17, of 5.23 compared to 3.5. So this might imply that all the respondents who saw the different treatments were kind of interested in the product. 5.3 Normality, kurtosis and skewness Exploration of the data was needed to check whether the data is normal distributed. The
It was found that the data was negatively skewed (skewness value = .952, SE=.176) and the kurtosis was slightly high (kurtosis value = .825, SE=.35) which confirms the values and the shape of the histogram. The histogram for perceived campaign risk was approximately normally distributed (skewness value = .44, SE=.176) and the kurtosis was slightly high (kurtosis value = .231, SE=.35) which confirms the values and the shape of the histogram. The histogram for trust in the product was skewed to the left. Analyzing the data further, there was found that the data was negatively skewed (skewness value = 1.238, SE=.176) and the kurtosis was extremely high (kurtosis value = 2.275, SE=.35) which confirms the values and the shape of the histogram. The histogram for risk propensity of the campaign was approximately normally distributed (skewness value = .11, SE=.176) and the kurtosis was slightly low (kurtosis value = .537, SE=.35) which confirms the values and the shape of the histogram. 5.4 Correlations There were many control variables in this experimental design, so this is why there was found some significant correlations. They are all put in the table below and a descriptive of the significant correlations are explained in the text below. 5.4.1 General perceived CF risk
The factor “General perceived CF risk” correlated with many other variables. So there was a positive significant effect on the both variable product risk (r = .23, p =.002) and financial risk (r = .23, p =.001). This means that the higher people in general perceive CF risk the higher the product and financial risk is perceived. This also accounts for the perceived
campaign risk (r = .17, p =.018), this effect was also positive significant, so the higher the campaign risk was perceived the higher people in general perceive CF as risky.
There was also a significant negative relationship with interest in the product (r = .15, p =.044) and the amount of in general perceived CF risk. So this means that the higher the interest is in the product is, the lower people in general perceive CF risk.
Perceived campaign trust and General perceived CF risk were found to have a
negative significant relationship (r = .18, p =.013). So this means that the higher the amount of perceived campaign trust is, the lower will be the general perceived CF risk. 5.4.2 Product risk and financial risk Financial risk and the amount product risk perceived were positively significant related to each other (r = .34, p <.001). So this means that the higher the financial risk is perceived, the higher people perceive product risk. A negative relationship between the amount of product interest and the amount of product risk perceived was found (r = .16, p =.032). This means that the higher the amount of product interest is perceived, the lower people perceive product risk. There is a significant positive relationship between the amount of perceived campaign risk and the amount product risk (r = .31, p <.001) and financial risk (r = .26, p <.001). So this means that the higher the campaign risk is perceived, the higher people perceive product risk and financial risk.
5.4.3 Product interest and perceived campaign trust
Product interest is positively related to the amount of perceived campaign trust (r = .25, p <.001). So, this means that the higher the amount of perceived campaign trust, the more people have product interest.
Product interest was negatively related to the amount of general perceived CF risk (r = .38, p <.001). So this means that the higher the amount of general perceived CF risk, the less people have interest in the product. A significant negative relationship was found between the amount of perceived campaign trust and the amount of perceived campaign risk (r = .38, p <.001). The less people perceive campaign risk the more perceived campaign trust they have. Table 3: Means, standard deviations and correlations
5.5 Hypotheses The hypotheses were tested by the PROCESS macro for SPSS from Hayes (2013). There was first done a mediation analysis of the role of perceived campaign risk. And thereafter two times moderation analyses for risk propensity and perceived campaign trust. 5.5.1 Mediating role of perceived campaign risk The multiple regression model explains 19% (R2 =.19) of the variance in product interest and is significant (F(12,178)=3.57, p<.001). This means that social proof, perceived campaign risk and all the control variables (Age, Gender, Education, Ever Backed Before, General perceived CF risk, Product risk and Financial risk) are explaining 19,4% of the variance of product interest. Results are presented in Table 4. Comparing the expert group to the control group there was not found a significant difference on perceived campaign risk (b=.16, t(178)=.66, p=.51). The expert group is lower on perceived campaign risk than the control group. Comparing the peer group to the control group there was not found a significant difference on perceived campaign risk (b=.16, t(178)=.62, p=.54). The peer group is not lower on perceived campaign risk than the control group. Comparing the expert and peer group to the control group there was also not found any significant difference on perceived campaign risk (b=.12, t(178)=.48, p=.63). The expert and peer group is not lower on perceived campaign risk than the control group. perceived campaign risk has a significant influence on product interest (b=.53, t(178)=4.87, p<.001). If perceived campaign risk is increasing then product interest will decrease. This means that perceived campaign risk is negatively related to product interest so if perceived campaign risk is increasing with 1 then product interest will decrease with (.54).
So, hypothesis 2 is rejected, the positive relationship between social proof and product interest is not mediated by perceived campaign risk. To conclude this, the total effect of social proof on product interest is not significant (R 2 =0.01, F(3,179)=.54, p=.66). The conclusion is that there is no direct or indirect effect of social proof on product interest. Therefore perceived campaign risk is not a mediator. So hypothesis 1a and hypothesis 1b are also rejected, there is no significant relationship between social proof and product interest. Table 4: Process Analysis Mediation perceived campaign risk After finding the significant relationship between perceived campaign risk and product interest, there has been done a extra check whether social proof moderates the significant negative relationship between perceived campaign risk and product interest. The control variables have been taken in account and therefore the PROCESS macro for SPSS from Hayes (2003) is the best way to test this moderation.
The multiple regression model explains 22% (R2 =.22) of the variance of product interest and is significant (F(15,175)=3.24, p<.001). This means that perceived campaign risk, social proof and all control variables (Age, Gender, Education, Ever Backed Before, General perceived CF risk, Product risk and Financial risk) are explaining 22% of the variance of product interest. Comparing the expert group to the control group there was not found a significant difference on product interest (b=.14, t(175)=.59, p=.56). Expert group is not significantly different in product interest than the control group. Comparing the peer group to the control group there was not found a significant difference on product interest (b=.20, t(175)=.79, p=.43). The peer group is not significantly different in product interest than the control group. Comparing the expert and peer group to the control group there was also not found any significant difference on product interest (b=.15, t(175)=.60, p=.55). The expert and peer group is not significantly different in product interest than the control group. Looking at the interactions between perceived campaign risk and the treatments of social proof there has not been found a significant effect. The interaction of Expert_SP with perceived campaign risk on product interest was significant (b=.62, t(175)=2.07, p=.04). This means that individuals from the expert group who perceive lower campaign risk will significantly have more interest in the product. The interaction of Peer_SP with perceived campaign risk on product interest was not significant (b=.45, t(175)=1.62, p=.11) and last the interaction of Expert_Peer_SP also did not significantly interact with perceived campaign risk on product interest (b=.35, t(175)=1.33, p=.19). Perceived campaign risk has no significant influence on product interest in this model (b=.24 t(175)=1.35, p=.18). To conclude this, the treatments with different versions of