Are crowdfunded products valued more?

(1)

Are crowdfunded products valued more?

Thesis seminar Business Studies

Academic year 2013-2014

21 July 2014

Author:

Luuk Ballhaus

10273530

(2)

Abstract

This study aims to find to whether and under which conditions crowdfunded products are valued more by consumers. Crowdfunding is an alternative way of financing where

consumers can support the launch of new products by donating money to entrepreneurs. In return, the consumers usually receive a product instead of company shares. Many studies have focused on which type of crowdfunding projects are likely to be supported by consumers, but this study is the first the address the perceived value of products that are crowdfunded. I hypothesize that signaling that a product is crowdfunded increases the perceived value of consumers, but only when the signal is perceived credible. Furthermore, I also propose that the perceived value of consumers increases more when the intensity of the signals increases. This is tested by conducting a field experiment with consumers. The sample size used for analysis consists of 315 respondents. The results show that consumers have an increased perceived value of 19% for crowdfunded products. However, this is only the case when the signal is perceived credible. Furthermore, sending more intense signals about the success of the crowdfunding process has no significant influence on the perceived value of consumers. Implications for managers and theory are provided along with suggestions for future research.

(3)

1. Introduction

In the past decade, the planet is confronted with a worldwide financial crisis which brought large and small banks in danger. Especially in times like these, banks have to be more

cautious and getting a loan from a bank is becoming more difficult. Entrepreneurs suffer from this difficulty, since they often need a loan to finance their business. This situation stimulated them to come up with new ways of financing and one of them is crowdfunding.

Crowdfunding allows small inventors to pitch their ideas on internet platforms

(Pierrakis & Collins, 2013). Instead of taking a loan from the bank or selling an idea to a large company, crowdfunding connects investors and inventors directly with each other, making it possible for inventors to develop their own idea without the financial risks of a bank loan (Ward & Ramachandran, 2010; Pierrakis & Collins, 2013). The idea is that the ‘crowd’ can donate money to a project and thereby provides ‘funding’ to an entrepreneur. However, entrepreneurs maintain full ownership of the company when seeking funds through crowdfunding (Pierrakis & Collins, 2013). Donating money to a project does not give any ownership rights over the company, which is the case when buying shares from a company. Besides, crowdfunding has another advantage: the crowdfunding process is also a way to determine the demand of consumers.

Nowadays, consumers are confronted with high amounts of alternatives when shopping for a specific product, which creates uncertainty for sellers and consumers. Sellers find it difficult to estimate the demand for their products, while consumers are confronted with more choice than they need. Since consumers simply do not have the time to consider all alternatives, consumers are forced to rely on other mechanism like mental shortcuts (Griskevicius et al., 2009). One of these mental shortcuts is imitation.

Since a lot of consumers face the problem of uncertainty, whole groups of people are imitating each other. Besides, people have a psychological need to feel appreciated and to be part of a group, which results in the existence of social trends (Axelrod, 1986).

Psychologically, it is hard for consumers not to participate in social trends. ‘If everyone is using this product, then it is probably a good choice’. Looking at others to find out what is ‘the right thing to do’ is called social proof (Rao et al., 2001; Bandura & Mclelland, 1977, Griskevicius et al., 2009). Social proof is one of the factors that causes people to imitate one another, also in terms of what others purchase. This creates a situation in which products might benefit from popularity under consumers.

Although no research is done on the perceived value of consumers for crowdfunded products, one can expect that crowdfunded products signal popularity. A crowdfunded

(5)

product is already exposed to a relatively small amount of consumers on the respective

crowdfunding platform. This study aims to fill this gap in the literature. The research question is therefore set as the following: to what extent does signaling that a product is crowdfunded increase the perceived value of consumers?

This is tested by conducting a field experiment with consumers and a crowdfunded product. In this study, the willingness-to-pay of consumers is measured and the information about the crowdfunding process is being manipulated. Since no prior study focuses on the relationship between signaling that a product is crowdfunded and the perceived value of consumers, this study adds to signaling theory and to theory on social proof. Also, knowing the factors that drive willingness-to-pay is vital information for managers in order to

maximize the sales income. When a positive effect is found, it might be beneficial for marketers to communicate that this product is crowdfunded towards consumers, in order to increase their perceived value.

Firstly, the theoretical framework is discussed in section two along with the hypothesis that follow from reviewing the literature. Secondly, section three states the methodological choices, along with the sample frame. Subsequently, an analysis of the auction results is given in section four. And finally, the general findings are discussed in section five along with implications for managers and theory, limitations, suggestions for future research and a conclusion.

(6)

2. Theoretical Background

In this section, the literature on crowdfunding, signaling and social proof is reviewed to be able to clarify the research question. Firstly, literature concerning crowdfunding is reviewed to give a broad idea about this concept. Secondly, literature about the difficulty to interpret and determine a products quality is discussed, followed by ways to tackle this problem. Subsequently, literature about the psychological factors behind imitation and social proof is discussed, along with some implications. Finally, the hypothesis that follow from the theory are stated.

2.1 Crowdfunding

Crowdfunding is an alternative way of financing a project. Instead of taking a loan from the bank, crowdfunding connects entrepreneurs and investors directly, usually through internet platforms (such as Kickstarter.com). The concept allows small donations from many parties, instead of large sums from a few (Ward & Ramachandran, 2010; Pierrakis & Collins, 2013). To initiate crowdfunding, entrepreneurs can submit their product-idea at one of the platforms. Next, they have to establish certain targets which state the required funding and the timeframe to reach this amount, in order to succeed in the crowdfunding process (Belleflamme, Lambert & Schweinbacher, 2012).

As stated in the section above, donating money to a crowdfunding process does not give any ownership rights, as is with buying shares from a company(Pierrakis & Collins, 2013). Instead there are usually two possible outcomes in crowdfunding. One, the

crowdfunding process fails to collect the targeted amount of financing within the given time, which causes the project to be stopped and the donators receive their money back. Two, the crowdfunding process succeeds to collect the targeted amount of financing within the given time, which causes the project to be launched and the donators are the first consumers to receive the product (Belleflamme et al., 2012; Ward & Ramachandran, 2010).

In 2012, two Dutch entrepreneurs invented the WakaWaka power (douwenkoren, 2013). This is a small device which charges an internal battery using solar power. This battery can be used to either charge your smartphone or to power a build-in lamp (douwenkoren, 2013). Besides serving smartphone users in the first world here, this product can also be used to replace the kerosene lamps that are often used in third world countries and which are harmful for the health of those people. The product idea was initiated at two different crowdfunding platforms and instantly became a great success (douwenkoren, 2013).

(7)

The quality of products that are in the crowdfunding process are hard to determine since these products are usually still in development (Ward & Ramachandran, 2010). Although the quality of crowdfunded products can be interpreted by looking at the specifications of the product, the question remains whether these products deliver what is expected of them (Ward & Ramachandran, 2010). Therefore, it is difficult for consumers to discriminate among crowdfunding projects and to see which of these projects have more potential than others (Belleflamme et al, 2013).

Although crowdfunding is a relatively new phenomena, it has already been the subject of many studies. Prior research mainly focused on the characteristics of crowdfunding and how a crowdfunding concept could signal quality to investors, causing them to choose this concept over another (Ward & Ramachandran, 2010). But what does a crowdfunded product signal towards consumers? Does signaling that a product is crowdfunded increase the

perceived value of consumers? The next paragraphs focus on the concepts signaling and social proof.

2.2 Signaling

When a product-idea succeeds in the crowdfunding process, this can be seen as a signal that tells consumers that this is a strong and popular product. A signal that helps consumers discriminate between high-quality sellers and low quality sellers is called a quality signal (Kirmani & Rao, 2000). This emerges from signaling theory, which was introduced in the 1970s from studies on asymmetric information between buyers and sellers in market interactions (Boulding & Kirmani, 1993). For example, buyers are usually less informed about the quality of a product than sellers are. Some sellers know that the products they are selling are of lesser quality. However, they might still advertise as if their products are of good quality, since they can ask higher prices then (Connelly et al., 2010). This causes all sellers to promote their products as being good, whether this is true or not. This problem is called adverse selection, and implies that it is difficult for buyers to discriminate among sellers, and to perceive which sellers or products are of better quality (Boulding & Kirmani, 1993; Connelly, Certo, Ireland & Reutzel, 2011; Kirmani & Rao, 2000).

One way to tackle this problem, as being a seller of good products, is by using signals which tell the consumers that you can be trusted (Boulding & Kirmani, 1993; Conelly et al, 2011). Although the use of signals can be quite costly, the gains for high-quality firms should be higher when they distinguish themselves from the lower-quality sellers (Kirmani & Rao,

(8)

2000; Zhao, 2000). For the lower-quality firms, the imitation of this signal would cause them to lose money. Either directly when the consumer discovers the true quality of the product, or indirectly because of the opportunity costs for strategies that could have been followed alternatively (Kirmani & Rao, 2000; Boulding & Kirmani, 1993).

Boulding and Kirmani (1993) studied the use of product warranties as a signal of credibility and their effect on the perceived product quality for consumers. By altering the characteristics of the warranty (size, scope and signaling conditions) the perceived quality should vary between different market equilibria. For this, they make a distinction between high and low bond credibility. High bond credibility implies that ‘high-quality firms can efficiently use a signaling mechanism which is too costly for low-quality firms to imitate’ (Boulding & Kirmani, 1993). Their results show that when bond credibility is high, bigger scope and length of the warranty would lead to a higher perceived quality. This is in line with the assumption that signaling is a way to communicate quality towards consumers. However, when bond credibility is low, increasing the scope and length of a warranty has an inverted effect. In this case the high-quality firms are not perceived as better than low-quality firms.

In another study, Chu & Chu (1994) studied the signaling quality by selling through a reputable seller when the original seller has no reputation of his own. This is a widely used signal which has not been studied much. They created a complex model in which they consider different market situations. The results show that signaling quality through a

reputable seller is beneficial for brands with no reputation yet (Chu & Chu, 1994). Especially in crowdfunding, most crowdfunding projects are initiated by small entrepreneurs that have an innovative idea, but usually not the funds and reputation to make their idea into a success. Selling a crowdfunded product through a reputable seller might therefore enhance the perceived value and perceived quality of consumers.

2.3 Signaling strategies

Many ways exist in how sellers can signal their quality to consumers. These signals can be related to the product itself or non-related to the product. For example, a product can be offered specifically through a reputable seller, or being marketed under a brand name that already has a big popularity (Chu & Chu, 1994; Rao, Qu & Ruekert, 1999). These signaling strategies have no relation to the product itself, but make use of the good reputation from others. However, sellers can also signal quality to their consumers by using signals that are related to the product. Product warranties and refund policies say something about the reliability and durability of a product, and guarantee the consumer that their newly bought

(9)

product is of a certain quality (Moorthy & Srinivasan, 1995; Boulding & Kirmani, 1993). Furthermore, sellers can use signals that signal quality by revealing the perceptions of others about the product. A movie award for instance shows that the movie is perceived as good by movie experts (Gemser, Leenders & Wijnberg, 2008).

However, not all signals have the same effects (Boulding & Kirmani, 1993). The results of the study discussed in the previous section by Boulding & Kirmani (1993) show that using signals is only effective when these signals are perceived credible by consumers. Warranties and refund policies are examples of credible signals. Sellers of good quality know that their products will probably last for the entire length of the warranty. However, sellers of bad quality know that they cannot make the same promise towards consumers since it is costly for them if their product fails to last the entire length of the warranty (Boulding & Kirmani, 1993; Moorthy & Srinivasan, 1995). Alternatively, sellers can also send signals with low credibility, although these signals are unlikely to give sellers a competitive

advantage(Boulding & Kirmani, 1993; Moorthy & Srinivasan, 1995). An example that is often used is a company catchphrase with: “We are the absolute number one in sportswear!”. There is no proof or evidence that this company actually is the seller of highest quality, and therefore offers no real advantage (Conelly et al., 2010).

Looking at the prior section, it seems likely to think that signaling, as being a seller of products with good quality, is without limitations. It would imply that the optimal strategy is to send as much signals as possible and also as intense as possible. However, sending too many signals might create suspicion, and could cause the advertising efforts to backfire (Gefen & Straub, 2004; Sirdeshmukh, Singh & Sabol, 2002). The challenge for sellers is to find the right balance of signaling, and remaining trustworthy and credible in the eyes of the consumer.

2.4 Social proof and imitation

On a daily basis, consumers are confronted with too many signals to process properly. This causes consumers to also rely on alternative methods, for example mental shortcuts like looking at others, to help them make decisions (Griskevicius et al., 2009). Imitation is one of these mental shortcuts. People imitate each other because they need social proof, which can be defined as “determining the appropriate behavior for yourself in a situation by examining the behavior of others” (Cialdini et al., 1999). Furthermore, multiple social studies also stated that imitation is “a characteristic response to uncertainty in decision making” (Bandura &

(10)

Mclelland, 1977; Rao, Greve & Davis, 2001). By imitating role models, people reduce their search costs and avoids the cost of experimentation (Rao et al., 2001).

People look for social proof in almost every situation during the day, most of the time without even realizing it. “Whether the question is what to do with an empty popcorn box in a

movie theater, how fast to drive on a certain stretch of highway, or how to eat chicken at a dinner party, the actions of those around us will be important in defining the answer”

(Axelrod, 1986). If everyone is doing things in a certain way, than that way is probably efficient (Boulding & Kirmani, 1993).

Shang, Basil and Wymer (2009) investigated whether people are willing to participate in a towel and linen reuse program in hotels, under different types of circumstances. In this study, the authors manipulated the information about what they would do with the savings from the reuse program, combined with situations where social proof is evident or not. Three messages about what would happen with the savings were communicated with the hotel guests, namely (1) no information, (2) reduced room rates and (3) charity donations. Three scenarios are used here since hotel guests could be there for business, which would imply that they did not personally pay for their stay and would not be personally effected by reduced room rates, and guests that stay for vacation, which are personally effected by reduced room rates. The results showed that when social proof was not evident, the hotel guests were more likely to participate in the reuse program when they would personally benefit from the savings. However, when social proof was evident, no difference was found between the different situations. This implies that social proof is a strong determinant of human actions, and that we might be influenced more by the actions of others than we think.

Especially if situations are new to us, people look at peers for valuable information on how to behave, without knowing the reasons or arguments behind that behavior (Axelrod, 1986). This also helps fulfill another psychological need, being part of a group (Axelrod, 1986). When there is uncertainty on how to behave in a new social environment, people tend to rely on experts. People with expert knowledge in a certain field are considered as more reliable in situations regarding that particular field than people who are novel to the situation (Shang et al., 2009).

One of the biggest implications of social proof in advertising is found in social trends. Looking at dermatological products for instance, in the past decade the amount of anti-aging products introduced has increased exponentially (Kreyden, 2005). This is mainly caused by the media suggesting how ‘the perfect female’ would look like, together with the need for consumers to participate in this trend. When almost everyone participates in a social trend, it

(11)

becomes very hard to stay out of that since people have the psychological need to be part of a group and to feel appreciated (Axelrod, 1986; Shang et al, 2009).

When a crowdfunding project is a success, this implies popularity under the

consumers this product is already offered to. The question remains whether this popularity enhances the perceived value of other consumers. As stated in the prior paragraph, a signal is only likely to enhance perceived value when the signal is credible. This study aims to find what is needed to signal towards consumers in order to make the crowdfunding process a credible signal.

2.5 Contribution

Although the theories on signaling and social proof are studied widely, theories on

crowdfunding are relatively new and often still in development. The existing literature mainly focuses on how consumers can pick the most promising crowdfunding projects, while this study is the first to look at how crowdfunding can increase the perceived value of consumers. Therefore, this study adds to theory on signaling, social proof and crowdfunding.

This study aims to find whether and how crowdfunding is perceived as a credible quality signal by consumers. Furthermore, this study also aims to find whether the

relationship between popularity and perceived value is proportional. This is vital information for managers in order to market their products and also for researchers since this study might be used for further research.

2.6 Hypotheses

This section focuses on the hypotheses behind the research question of this study: to what

extent does signaling that a product is crowdfunded increase the perceived value of consumers? The hypotheses are developed using the literature described in the previous

paragraphs.

2.6.1 Crowdfunding as a signal to consumers

As stated earlier, although crowdfunding is a relatively new phenomena, there has already been a significant amount of research done on this subject. Prior research mainly focused on the characteristics of crowdfunding and how a crowdfunding concept could signal quality to consumers, but this study tries to focus on the signaling capabilities of a crowdfunded product towards consumers. Since a lot of product-concepts are offered at crowdfunding websites,

(12)

consumers have to decide which projects they consider the most promising (Belleflamme et al., 2012). This means that only the most popular concepts in the eyes of these consumers will generate enough funds to enter the next stage in the development process (Belleflamme et al, 2012).

Knowing that people validate their actions by looking at the actions of others, a crowdfunded product can work as a signal towards consumers. Crowdfunded products signal that this product is popular under the consumers that are willing to help in crowdfunding processes and therefore is likely to be of good quality. However, prior studies suggest that products can benefit from quality signals if these signals are perceived credible (Boulding & Kirmani, 1993). When the signal is perceived incredible, no real advantage should be present (Boulding & Kirmani, 1993; Moorthy & Srinivasan, 1995). This leads to the first hypothesis:

Hypothesis 1 [H1]: Signaling that a product is crowdfunded increases the willingness-to-pay of consumers, but only if the signal is perceived credible.

2.6.2 The impact of social proof on willingness-‐to-‐pay

The mechanism behind the first hypothesis is based on the notion that consumers validate their own actions by looking at the actions of others (Cialdini et al., 1999). Since consumers are constantly looking for social proof, popularity of a product is an important determinant of the perceived value of consumers for a product. This is due to the fact that popular products are efficient to imitate, and therefore reduces search costs and the cost of experimentation (Boulding & Kirmani, 1993; Rao et al., 2001). A credible crowdfunding signal should therefore provide consumers with social proof (Axelrod, 1986).

The question remains when crowdfunding becomes a credible signal. Signaling that a product is crowdfunded might create an ambiguous situation for consumers, since no

information concerning the success of the crowdfunding process is communicated. The second hypothesis is therefore set as the following:

Hypothesis 2 [H2]: Crowdfunding as a signal is perceived credible if the success of the crowdfunding process is evident.

2.6.3 The impact of the intensity of social proof on willingness-‐to-‐pay

Moreover, a crowdfunded product also shows that it is popular under a large group of people. Crowdfunding platforms only accept small donations, so a large group of people is needed in order to reach the funding objective (Belleflamme et al., 2012). Knowing that a

(13)

group of people like this product will create more social proof than knowing that only one people likes this product (Axelrod, 1986; Shang et al., 2009). This implies that the intensity of the signal a crowdfunded product sends to consumers is higher when more people are

involved. And since a crowdfunded product is supported by multiple people (Pierrakis & Collins, 2013), it can be assumed that this will also give more social proof to consumers. Thus, the third hypothesis is set as:

Hypothesis 3 [H3]: Increasing the intensity of signaling social proof to consumers increases their willingness-to-pay.

(14)

3. Methodology

After explaining the various theories and concepts concerning what a crowdfunded product signals to consumers, this next section focusses on the methodology and research design. Firstly, I clarify and justify the chosen research design, a second price auction with sealed bids. Secondly, some methodological choices are discussed.

3.1 Research design

To be able to test the hypothesis stated in the prior section, the following research design with four treatment groups is tested:

Treatment 1: Just the product (Baseline)

Treatment 2 : The product + mention crowdfunded

Treatment 3 : The product + mention crowdfunded + 100% target financing acquired Treatment 4 : The product + mention crowdfunded + percentage of target financing acquired : 2349%

First of all, a baseline or control group is needed to see whether the manipulations brought an effect. Then, the other three treatments are created to be able to manipulate the intensity of the crowdfunding signal. Treatment 2 and treatment 3 both give the same information, since a crowdfunded product always has 100% target financing acquired. The fact that these treatments are, information wise, the same makes it possible to use this as a manipulation check and to know if the consumers fully understand the concept crowdfunding. The fourth treatment sends a more intense signal, namely that the project collected 23.49 times the targeted amount of financing. This implies that this product is popular under a large group of consumers, since each donator can only make a small payment towards this concept. This indicates that this is a popular product and therefore more social proof should be provided to consumers.

The hypotheses are tested using a quantitative approach. This study aims to find causal relationships between a set of variables. Saunders, Lewis & Thornhill (2009) state that in these situations, an experiment is most suitable. For this study, a field experiment with consumers is used. A crowdfunded product is auctioned using a Vickrey auction, which means that all bidders submit one sealed bid and the highest bidder wins the auction. However, the winning bidder pays the price of the second highest bid. This study uses real auctions, with real products and real consumers. This creates an incentive to bid up to your

(15)

maximum willingness-to-pay, since the bidders are confronted with a real situation, and will always receive a discount in case they win the auction, due to the second-price system (Lusk, Alexander & Rousu, 2007; Noussair, Robin & Ruffieux, 2004). So instead of measuring what people say, this method measures what people will actually do, in terms of willingness-to-pay.

For this study, an innovative and crowdfunded product was auctioned, namely the StickrTrackr. This is a small tracking device, which can be located using an app on a

smartphone. For example, when the StickrTrackr is attached to a set of keys, your smartphone can tell where they are when they are lost. This product was chosen because it is (1) a

crowdfunded product and (2) was a successfully crowdfunded project with 2349% of the target financing collected.

Since the auction is online, a lot of people can participate in the bidding using their computer or smartphone. For this study, this is a great advantage compared to ‘normal’ auctions because it is only a small effort for the respondents to place a bid, and all the data can be gathered relatively easy and within one day. Finally, after the respondent places his/her bid, five questions are asked for control purposes.

3.2 Control questions

As stated above, five short questions are asked after the participants place their bids. These questions grant the opportunity to add more variables to the research design, and can be used for control purposes. The full questions and an overview of the main goals of the questions with their argumentation is provided .

Q1: Do you consider the StickrTrackr to be a popular product? yes/no

Knowing whether the participants in the auction perceive this product as popular is viable information for this study. Products that are perceived as popular provide social proof. Since I stated that products that provide social proof increase the willingness-to-pay of consumers, this question is needed to be able to test this hypothesis. If people who consider this product as popular are also willing to pay more than people who do not consider this product as popular, than the only factor causing this rise in willingness-to-pay is the popularity of the product (i.e. the social proof it provides).

Q2: Are you familiar with crowdfunding? yes/no

This question has the purpose to filter out those respondents that are unfamiliar with

(16)

the data from people that are unfamiliar with crowdfunding might be affected by ambiguous factors other than this study is aiming for.

Q3: Do you think others were willing to buy this product for 10 euros before launch? yes/no

The main goal of question 3 is to perform a manipulation check. Treatments 2, 3 and 4 stated that the StickrTrackr is a crowdfunded product. Since treatment 2 did not state information about whether or not the crowdfunding process was a success, the participants in this treatment might not be sure whether this is the case or not when answering this question. 63.2% of the participants answered yes to this question. The results show that there is a significant difference between the mean bid amounts of 325.07 eurocents between treatment two and treatment three and four combined (P=0.002, F=9.959).

Furthermore, treatments 3 and 4 did state that the crowdfunding process was a success and this should imply that others were in fact willing to buy this product before the launch. People that were in treatment 3 or 4 and answered ‘no’ to this question might not fully

understand the concept crowdfunding. Furthermore, if participants that answered ‘yes’ to this question are also willing to pay more, this would imply that the fact that this product was popular under the first consumers increases other consumers’ willingness-to-pay.

Q4: Is it likely for you to pay more for a product if you perceive that as a popular product? yes/no

What makes this study different than traditional WTP-studies is that it measures what people actually do, instead of measuring what people say they do. By asking this question, the ability to check whether people act upon their words arises.

Q5: The more popular a product is, the more I would be willing to pay for it. true/false

Finally, the fifth question has a goal similar to that of the fourth question. Again, this question is used to check whether people act upon their words. As the intensity of social proof

provided increases, I hypothesized that the willingness-to-pay should also increase.

3.3 Product characteristics

As stated in the section above, the product that is auctioned in the experiment is the

StickrTrackr. First of all, since this research focuses on whether crowdfunded products are valued more, the auctioned product has to be a crowdfunded product. Subsequently, this product has something new and unique to offer, which creates a situation where the participants cannot make a comparison with another product. This makes it harder for

(17)

consumers to estimate the value of the product. This is a great advantage, because if the bidders already knew a reference price, then they would only bid up until that price because they are able to buy it for that price somewhere else. Crowdfunded products are also products that come from relatively small developers (Ward & Ramachandran, 2010). This means that the actual value of the product can be measured, where if a product of for example Philips was sold, it might be ambiguous whether the actual value of the product or the value of the brand is measured.

3.4 Sample characteristics

For this study, the existing panel owned by Veylinx is invited to participate in the auction. In this particular auction, a total of 471 respondents placed a bid on the StickrTrackr.

Furthermore, 45 respondents did not fully answer the control questions and are filtered out of the data. This leaves a sample of 426 respondents. To be able to generalize the results of this sample to the whole population of Dutch citizens, the distributions of age, gender and education of the sample are compared to their respective distributions of the actual population. As stated earlier in this paragraph, participants that were unfamiliar with crowdfunding have been removed from the sample since this study is aiming to find results about crowdfunding and the effect on willingness-to-pay. After removing this data, a check is made to see whether the remaining participants were still evenly distributed over the 4

treatments. The distribution in the treatments varied from 22.5% till 28.1 %, and are shown in table 1. The remaining sample consists of 324 participants, of which 58.6% males and 41.4% females. This differs a little from the entire Dutch population, which consists of 50.5%

females and 49.5% males (Centraal Bureau Statistiek, 2013). Table 2 and 3 compare the age and education level distributions of the sample and the population. Some major differences in the age and educational level of the sample and the population exist. In the sample, the majority is between 20 and 40 years and almost 50% of the participants has been to university. The main reason for these differences is that the panel Veylinx owns primarily consists of young students and researchers and their networks.

Table 1: Participants in the 4 treatments after filtering data

Treatment 1 2 3 4 Sample (N=324) 24.1% 25.3% 22.5% 28.1%

(18)

Table 2: Age sample compared to age of Dutch population

Age in years <20 20-40 40-65 65-80 >80 Sample (N=324) 3.7% 57.5% 29.8% 9% 0% Population 23.1% 24.6% 35.5% 12.6% 4.2%

Table 3: Education compared to education Dutch population

Highest education Primary Secondary MBO HBO WO school school

Sample (N=324) 2.3% 6.8% 11.3% 31.1% 48.6% Population 8.4% 23% 40.1% 19.3% 9.2%

(19)

4. Results

In this next section the results of the data-analysis are provided. Firstly, some descriptive statistics are discussed. Secondly, the hypothesis are tested by comparing the different treatment groups. Subsequently, some other influences are discussed.

4.1 Descriptive statisctics

Although the sample characteristics differ significantly from the population, this does not mean that the results are not useful. The product that was auctioned for this study was only usable in combination with a smartphone app that only runs on the newest smartphones. It therefore seems expected that people under 40 are the biggest group of potential customers. Furthermore, 73.1% of the people that were unfamiliar with crowdfunding were over 40. This indicates that the sample includes more people that are interesting for this study than a sample similar to the Dutch population.

First of all, many participants place a bid of zero eurocents or a bid just above zero eurocents. Because this study aims to find differences in willingness-to-pay, one way to tackle this problem is by looking at the bids in the top 50 percentile only. The bids in the top 50 percentile range from 200 eurocents till 4500 eurocents. These bids included 9 outliers, which were removed from the data set. After removing the outliers, the mean bid of the top 50 percentile is 962,11 (SD=509,868). In the top 50 percentile, the mean bid in treatment one is 862.49 eurocents (N=39, SD=448,386), for treatment two is 785,76 eurocents (N=37,

SD=380,768), for treatment three the mean bid is 1078,37 eurocents (N=30, SD=470,666) and for treatment four is 1135,18 (N=40, SD=625,733). Table 4 shows an overview of the mean bid amounts of the auction bids in eurocents in their respective treatments within the reduced sample, and within the bids of the top 50 percentile only. Table 5 shows the distribution of the control questions. The treatments used are:

Treatment 1: Just the product (Baseline)

Treatment 2 : The product + mention crowdfunded

Treatment 3 : The product + mention crowdfunded + 100% target financing acquired Treatment 4 : The product + mention crowdfunded + percentage of target financing acquired : 2349%

(20)

Table 4: Mean bid amounts per treatment in whole sample and in top 50 percentile Whole sample Treatment N Mean SD 1 77 468.01 516.320 2 78 411.19 450.909 3 70 505.66 595.792 4 90 566.02 673.669 Top 50 percentile Treatment N Mean SD 1 39 862.49 448.386 2 37 785.76 380.786 3 30 1078.37 470.666 4 40 1135.18 625.733

Table 5: Distribution yes/no for the control questions

Question Percent yes

Q1: Do you consider the StickrTrackr to be a popular product? yes/no 38.7 Q2: Are you familiar with crowdfunding? yes/no 100

Q3: Do you think others were willing to buy this product for 10 euros 63.2

before launch? yes/no

Q4: Is it likely for you to pay more for a product if you perceive that as 34.3 a popular product? yes/no

Q5: The more popular a product is, the more I would be willing to pay 22.9 for it. true/false

4.2 Hypotheses testing

The aim of this study was to find to what extent crowdfunded products are valued more by consumers. I hypothesized that crowdfunded products offer social proof, and that social proof increases willingness-to-pay. This is tested by incrementally adding more signals about the popularity of this product under other consumers. The first hypothesis is set as: [H1]: signaling that a product is crowdfunded increases the willingness-to-pay of consumers, but only if the signal is perceived credible. To test the first hypothesis, a dummy variable is

(21)

created to separate the participants from the baseline treatment (1) and the rest of the participants. As stated in the section above, for this analysis only the bids from the top 50 percentile are used to filter out the high amount of zero bids. This way, the bids are more similar to a normal distribution. A one-way ANOVA was performed to test whether the first hypothesis is supported. The mean bid for treatment 1 is 862.49 eurocents (SD=448.386) and for the other treatments combined the mean bid is 998.42 eurocents (SD=527.784), but the results show no significance (P=0.155, F= 2.046).

In this first analysis, The only manipulation was that the auctioned product is

mentioned to be crowdfunded or not. Whether the crowdfunding process was a success or not was not communicated and could cause ambiguity. To test the first and second hypothesis, another dummy variable is created to test the impact of communicating the success of the crowdfunding process. The second hypothesis is set as: [H2]: Crowdfunding as a signal is perceived credible if the success of the crowdfunding process is evident.

For this, treatment 2 is compared to treatment 3 and 4 combined. Again, only bids in the top 50 percentile are considered and a one-way ANOVA is performed to test whether mentioning the success of the crowdfunding process has an effect on the willingness-to-pay of consumers. The mean bid amount in treatment 2 is 785.76 eurocents (SD=380.786) and for the treatments 3 and 4 combined the mean bid amount is 1110.83 eurocents (SD=561.483). Other than the first ANOVA, this analysis does show significant results (P=0.002, F=9.959).

Finally, the third and last hypothesis states that: [H3]: Increasing the intensity of signaling social proof to consumers increases their willingness-to-pay. This implies that by incrementally adding more signals, the willingness-to-pay should also incrementally increase. Applied to this auction, treatment 4 should be significantly higher than treatment 3, since the only difference in information provided in these treatments is the intensity of the financing acquired. To test whether the treatments significantly differ from each other, a one-way ANOVA is performed using the mean bid amounts per treatment. To filter out the high amount of zero bids, only the bids in the top 50 percentile are considered in this analysis. The results show that the mean bid amounts for treatments 3 and 4 do not significantly differ from each other (P=0.678, F=0.173).

4.3 Testing the self-‐image of the participants and other influences

As stated in the methods section, five control questions are used in the research design. Questions 4 and 5 have the main goal to check whether the participants would act upon their

(22)

words. The participants were asked whether they thought they would be willing to pay more for popular products than for unpopular ones with question 4. A majority of 65.7% of the reduced sample answered ‘no’. To test this, only the participants are selected that answered ‘no’ to this question. Then, a one-way ANOVA is performed using mean bid amounts per treatment, with a post hoc LSD test. Again, for this test only the bids in the top 50 percentile are considered. The post hoc test shows that treatments 3 and 4 both are significantly higher than treatment 2. Treatment 3 with a mean difference of 343.597 eurocents (P=0.035) and treatment 4 with a mean difference of 463.975 eurocents (P=0.002). To check whether the people who answered ‘yes’ to this question are right, the same test is performed as before, only this time only the participants that answered ‘yes’ are selected. The post hoc test shows no significant differences between the treatment groups, although a majority of 64,1% of these people consider the StickrTrackr as a popular product.

Furthermore, the fifth question that was asked to the participants was whether they think that their willingness-to-pay is proportional to the popularity of the product (i.e. the more popular the product, the higher the willingness-to-pay). Here, 77,1% of the participants answered ‘no’ to this question. Again, the participants that answered ‘no’ to this question are selected and the same statistical test as above is used. The post hoc test shows no significant results here. In the same test with only the people that answered ‘yes’, a significant difference between the first and the third treatment was found, with a mean difference of 511,038

eurocents (p=0,026).

Subsequently, a regression analysis is performed to test whether some of the control questions influence the willingness-to-pay of consumers. For this analysis, the bids from the top 50 percentile only were considered. Also, the natural logarithm of the bid amounts is used since this gives results in percentages. The results are summarized in table 6. Question 2 is missing since this question was used to filter out the participants that were unfamiliar with crowdfunding.

Table 6. Results of regression analysis control questions

Control question B Significance

Q1: Do you consider the StickrTrackr to be a popular product? yes/no 0.206 0.019 Q3: Do you think others were willing to buy this product for 10 euros 0.441 0.000

before launch? yes/no

Q4: Is it likely for you to pay more for a product if you perceive that as -0.079 0.411 a popular product? yes/no

(23)

Q5: The more popular a product is, the more I would be willing to pay 0.138 0.180 for it. true/false

(24)

5. Discussion

In the previous section the results of the auction and the steps taken in the analysis of the data are discussed. This next section focuses on the interpretation of the results. Firstly, the general findings are discussed. Subsequently, the theoretical implications and the implications for managers are stated. Finally, the limitations of this study plus recommendations for further research are given.

5.1 General findings

The study aims to find whether and when crowdfunded products are valued more by consumers. I proposed a conceptual framework in which crowdfunding signals popularity under other consumers and therefore provides social proof to other consumers. Next, I proposed that this social proof has a positive proportional effect on the consumers’ willingness-to-pay. Firstly, the differences in the auction bids between the treatments are reflected and interpreted, followed by the other findings.

Firstly, the results show that signaling that a product is crowdfunded does not result in an increase in the perceived value of consumers for this product. Only when the success of the crowdfunding process is evident, an increase in the perceived value is found. When reverting back to the literature, a signal is only likely to have a positive effect if the signal is perceived credible (Boulding & Kirmani, 1993). When the signal is perceived incredible, no real

advantage should be present (Boulding & Kirmani, 1993; Moorthy & Srinivasan, 1995). This is in line with the first hypothesis: [H1]: signaling that a product is crowdfunded increases the willingness-to-pay of consumers, but only if the signal is perceived credible. This means that the first hypothesis is supported.

Furthermore, the only manipulation that is made when comparing treatment 2 to treatments 3 and 4 is that those treatments also signal the success of the crowdfunding process. This means that signaling the success of the crowdfunding process makes crowdfunding as a signal credible. This is in line with the assumption that a credible crowdfunding signal provides consumers with social proof, and is therefore popular and efficient to imitate (Boulding & Kirmani, 1993; Rao et al., 2001). This causes the second hypothesis to be supported also. [H2]: Crowdfunding as a signal is perceived credible if the success of the crowdfunding process is evident.

However, the results do not show a significant difference between treatments 3 and 4. Although the theory suggests that a more intense signal of the success of the crowdfunding

(25)

process would give consumers more social proof (Axelrod, 1986), this does not result in a higher perceived value in this study. This causes the third and final hypothesis to remain unsupported. [H3]: Increasing the intensity of signaling social proof to consumers increases their willingness-to-pay.

Finally, the regression analysis with the control questions shows some significant results. Participants that consider the StickrTrackr as a popular product are willing to pay 20.6% more than the people that do not consider it as a popular product. This is in line with the theory which states that popular products should provide social proof and therefore increases the perceived value of consumers (Axelrod, 1986; Rao et al., 2001). The regression analysis also showed that participants who thought that others were willing to pay for this product before launch had an increased willingness-to-pay of 44.1%. This implies that consumers value a product higher when it is popular under other consumers. This is also in line with the theory on social proof and imitation.

5.2 Theoretical implications

Literature on crowdfunding mainly focusses on how consumers can discriminate between crowdfunding projects. However, very little is known about what a successfully crowdfunded product can signal towards consumers. This study is the first to research to what extent a crowdfunded product can benefit from popularity under other consumers. The results show that signaling that a product is crowdfunded is perceived as an incredible signal for quality. The signal becomes credible when the success of the crowdfunding process is being signaled towards consumers. This is in line with the existing literature on social proof, signaling and imitation.

However, the benefits a crowdfunded product gains from popularity only works up to a certain point. Therefore, this study not only supports the existing theories on social proof and signaling, but also adds to these theories. The research design grants the chance to measure not just what people say, but also what they actually do in terms of willingness-to-pay. This resulted in a surprising situation in which the participants did the opposite of what they said they would do.

Finally, the results supported the literature on social proof once more by showing that the people who considered the StickrTrackr as a popular product had an increased

willingness-to-pay of 20.6%. Furthermore, people who considered this product as popular under other consumers had an increased willingness-to-pay of 44.1%.

(26)

5.3 Managerial implications

This study aimed to find a new indicator on what is for managers arguably the most important variable, the willingness-to-pay of consumers. Knowing as many factors as possible

influencing the willingness-to-pay is essential to maximize the profits from sales. This study shows that signaling that a product is crowdfunded towards consumers increases their willingness-to-pay. Although, just mentioning that a product is crowdfunded is not enough, managers should also state that the crowdfunding process was a success, which implies that many other consumers favor their product. Managers should also know that the effect is not proportional and works up to a certain point only. When that point is reached, other factors influencing willingness-to-pay must be addressed. It is always useful for managers to know which group of consumers values their product most, since they can adjust their advertisement campaigns based on this information.

5.4 Limitations and suggestions for further research

Several suggestions for future research are made when looking at the limitations of this study. Firstly, this study only considers the signaling of social proof towards consumers. It might be the case that a crowdfunded product also signals other things towards consumers, which have not been investigated in this specific study. Secondly, some limitations about the research design are also present. This study is conducted using a Dutch panel, therefore making it not possible to generalize the results to other countries. It is recommended that Veylinx expands the panel it uses in order to be able to make statements concerning bigger populations. Also, the intensity of social proof is a concept, which is hard to measure. It remains unclear whether more intense signals also lead to more intense social proof.

Also, participants in the treatments where the success of the crowdfunding process is mentioned and answered ‘no’ to the question whether they thought others would be willing to buy this product before launch, could have been filtered out of the sample since they might not fully understand the concept of crowdfunding. It is a fact that people were willing to pay for this product before launch if the success of the crowdfunding process is communicated. It was chosen not to do this since this would have created an uneven distribution of the sample over the different treatments. Besides, there is also a chance that some of the participants said that they were familiar with crowdfunding, but in fact are not.

(27)

Furthermore, it remains unclear whether the same results are found when a different product is auctioned. Further research should focus on whether the results are different when using different types of crowdfunded products. Further research should also focus on whether the results differ when comparing relatively cheap products versus more expensive products.

(28)

6. Conclusion

The study aims to find whether and when crowdfunded products are valued more by consumers. I proposed a conceptual framework in which crowdfunding signals popularity under other consumers and therefore provides social proof to other consumers. Next, I proposed that this social proof has a positive proportional effect on the consumers’ willingness-to-pay. The results show that the willingness-to-pay of consumers for a crowdfunded product is higher, but only under certain conditions.

This study shows that the willingness-to-pay of the participants only increases when the success of the crowdfunding process is mentioned. Furthermore, this effect is not

proportional, as was hypothesized, which means that more intense signals of social proof do not further increase the willingness-to-pay of consumers. Since the auctions in this study are real, the ability arises to measure and compare what people say they do, and what they

actually do, in terms of willingness-to-pay. Unexpectedly, the participants that answered to be unwilling to pay more for popular products actually turned out to be the ones that are willing to pay more. And surprisingly, no difference in the bid amount between the different

treatments is found with the people that answered to be willing to pay more for popular products.

Furthermore, this study contributes to existing literature on signaling, social proof and crowdfunding. This study focuses on whether popularity under investors is beneficial for a product, which is still not studied much. A suggestion for further research is to replicate this study and use a sample that is also representative outside of the Netherlands.

(29)

References

Axelrod, R. (1986). An evolutionary approach to norms. American political science review, 80(04),1095-1111.

Bandura, A. & Mclelland, D. C. (1977). Social learning theory. 305-316.

Belleflamme, P., Lambert, T., & Schwienbacher, A. (2013). Crowdfunding: Tapping the right crowd. Journal of Business Venturing.

Boulding, W., & Kirmani, A. (1993). A consumer-side experimental examination of signaling theory: do consumers perceive warranties as signals of quality?.Journal of Consumer Research, 111-123.

CBS (2013). Bevolking; kerncijfers. Accessed on June 23, 2014.

(http://statline.cbs.nl/StatWeb/publication/?VW=T&DM=SLNL&PA=37296ned&D1 =a&D2=0,10,20,30,40,50,60,(l-1),l&HD=130605-0924&HDR=G1&STB=T) CBS (2013). Beroepsbevolking; Behaalde onderwijs. Accessed on June 23, 2014.

(http://statline.cbs.nl/StatWeb/publication/?VW=T&DM=SLNL&PA=71822NED&D

1=0-1&D2=0&D3=0&D4=0-4&D5=a&D6=0&D7=0,4-l&HD=110405-1452&HDR=T,G2,G1,G5,G6&STB=G4,G3)

Chu, W., & Chu, W. (1994). Signaling quality by selling through a reputable retailer: An example of renting the reputation of another agent. Marketing Science, 13(2), 177-189.

Cialdini, R. B., Wosinska, W., Barrett, D. W., Butner, J., & Gornik-Durose, M. (1999). Compliance with a request in two cultures: The differential influence of social proof and commitment/consistency on collectivists and individualists. Personality and Social Psychology Bulletin, 25(10), 1242-1253.

Connelly, B. L., Certo, S. T., Ireland, R. D., & Reutzel, C. R. (2011). Signaling theory: A review and assessment. Journal of Management, 37(1), 39-67.

Douwenkoren (2013). De 15 toonaangevende crowdfunding projecten van 2012. Accessed on July 15, 2014. http://www.douwenkoren.nl/de-15-toonaangevende-crowdfunding-projecten-van-2012/

Gefen, D., & Straub, D. W. (2004). Consumer trust in B2C e-Commerce and the importance of social presence: experiments in e-Products and e-Services. Omega 32 (6) 407-424. Gemser, G., Leenders, M.A.M. and Wijnberg, N. (2008). Why some awards are more

effective signals of quality than others: A study of movie awards. Journal of Management, 34(1): 25-54.

(30)

Griskevicius, V., Goldstein, N. J., Mortensen, C. R., Sundie, J. M., Cialdini, R. B., & Kenrick, D. T. (2009). Fear and loving in Las Vegas: Evolution, emotion, and

persuasion. Journal of Marketing Research, 46(3), 384-395.

Kirmani, A., & Rao, A. R. (2000). No pain, no gain: A critical review of the literature on signaling unobservable product quality. Journal of marketing, 64(2), 66-79.

Kreyden, O. P. (2005). Antiaging–a scientific topic or just a social trend?. Journal of cosmetic dermatology, 4(4), 228-229.

Lusk, J. L., Alexander, C., & Rousu, M. C. (2007). Designing experimental auctions for marketing research: The effect of values, distributions, and mechanisms on incentives for truthful bidding. Review of Marketing Science,5(1).

Moorthy, S., & Srinivasan, K. (1995). Signaling quality with a money-back guarantee: The role of transaction costs. Marketing Science, 14(4), 442-466.

Noussair, C., Robin, S., & Ruffieux, B. (2004). Revealing consumers' willingness-to-pay: A comparison of the BDM mechanism and the Vickrey auction. Journal of economic psychology, 25(6), 725-741.

Pierrakis, Y., & Collins, L. (2013). Crowdfunding: A New Innovative Model of Providing Funding to Projects and Businesses. Available at SSRN 2395226.

Rao, H., Greve, H. R., & Davis, G. F. (2001). Fool's gold: Social proof in the initiation and abandonment of coverage by Wall Street analysts. Administrative Science

Quarterly, 46(3), 502-526.

Rao, A. R., Qu, L., & Ruekert, R. W. (1999). Signaling unobservable product quality through a brand ally. Journal of Marketing Research, 258-268.

Saunders, M., Lewis, P. & Thornhill, A. (2009). Research methods for business students. Pearson Education Limited.

Shang, J., Basil, D. Z., & Wymer, W. (2010). Using social marketing to enhance hotel reuse programs. Journal of Business Research, 63(2), 166-172.

Sirdeshmukh, D., Singh, J., & Sabol, B. (2002). Consumer trust, value, and loyalty in relational exchanges. Journal of marketing, 66(1), 15-37.

Ward, C., & Ramachandran, V. (2010). Crowdfunding the next hit: Microfunding online experience goods. In Workshop on Computational Social Science and the Wisdom of Crowds at NIPS2010.

Zhao, H. (2000). Raising awareness and signaling quality to uninformed consumers: A price-advertising model. Marketing Science, 19(4), 390-396.

(31)

Appendix

A. Treatment ads

Treatment 1 Treatment 2

Baseline Mention crowdfunding

Treatment 3 Treatment 4

Mention crowdfunding + success Mention crowdfunding + financing amount

Are crowdfunded products valued more?