Consumers’ perceptions towards crowdfunded products

(1)

Consumers’ perceptions towards

crowdfunded products

Universiteit van Amsterdam

Vrije Universiteit Amsterdam

Daan Glandorf

UvA: 10899235

VU: 2125927

Amsterdam, June 30, 2015

Master thesis Entrepreneurship

Supervisor: B. Kuijken

(2)

Abstract

Crowdfunding steadily becomes a more important source of finance. Hence it’s important to get a better understanding of this relatively new phenomenon. Very little prior research is done to examine consumers’ perceptions towards crowdfunded products. The aim of this study was to examine whether signaling that a product is successfully crowdfunded influences consumers’ willingness-to-pay and perceived quality. In addition, this research examines whether consumers’ willingness-to-pay and the perceived quality for crowdfunded products are affected by herding behavior. The author hypothesized that consumers are willing to pay more for successfully crowdfunded products. Moreover, it was hypothesized that consumers perceive that a product is of a higher quality once they know it’s crowdfunded. Finally, it was tested whether signaling the popularity of a crowdfunding project in terms of amount of backers and funding has an effect on consumers’ perceived quality and willingness-to-pay for a product. The findings of this research show that there is no increase in willingness-to-pay and perceived quality for the product once it’s signaled that a product is crowdfunded. Next to that, it’s found that consumers’ perceived quality is not affected by herding behavior. However, consumers are willing to pay more once they know that the crowdfunding project is overfunded. A contrary effect is found for willingness-to-pay once consumers know how much backers supported the crowdfunding project.

(3)

1. Introduction

Crowdfunding is an emerging phenomenon that enables ventures and individuals to raise funds from many individuals. This paper adopts the following definition of Belleflamme et al. (2014): “Crowdfunding involves an open call, mostly through the Internet, for the provision of financial resources either in the form of donation or in exchange for the future product or some form of reward to support initiatives for specific purposes.” According to a report from the World Bank (2013) the crowdfunding market has the potential to become one of the most important sources of capital. It is estimated that in 2025 the amount that is globally raised due to crowdfunding, will be between $90 billion and $96 billion per year, which is nearly two times the size of the global venture capital industry today. Since crowdfunding is expected to become a common source of finance for entrepreneurs in the future, it’s important to examine what implications might follow for ventures that apply crowdfunding.

Prior studies have already noted that besides the financial support that is offered, crowdfunding communities (i.e. backers) can potentially add value to the company (Schwienbacher & Larralde, 2010). Crowdfunding builds upon crowdsourcing activities, meaning that the community collaborates closely with the company. Kleemann et al. (2008) note that companies can benefit from cost-reduction due to the crowd. Backers participate in product design and improvement, which allows the company to save costs and time. Moreover, the crowd can help to promote the companies’ products, by word-of-mouth (Thies & Wessel, 2014). According to Brabham (2008), crowds can be very efficient in solving problems. Each individual of the crowd has its own expertise and skills to contribute to the company. The more diversity within the crowd, the more valuable the crowd is to the company. Since crowdfunding is often based on small investments of many backers, companies can make use of the wisdom of the crowd (Schwienbacher & Larralde, 2010).

If crowdfunding offers access to capital and various other benefits to companies, it might be that consumers’ perceptions towards crowdfunded products and services are

(5)

different than for products and services of regularly funded companies. Although crowdfunding is growing by popularity, there is little scholarly literature about what perceptions consumers have towards crowdfunded products. Consumers might perceive that crowdfunded products are of higher quality and are therefore willing to pay more for this particular product.

Schwienbacher and Larralde (2010) state that crowdfunding also functions as a tool to do market research. If there is little or no financial support for a particular product, it is presumable that there won’t be any demand for the product. Crowdfunding can therefore be a valuable tool for companies to estimate the market potential of a product. From a consumer perspective, the extent to which a company is able to raise funds using crowdfunding, might signal underlying product quality. So, if a company is able to raise a multiple of its funding goal, consumers might have a higher perceived product quality and will be willing to pay more for this product.

This thesis will examine consumers’ willingness-to-pay for products for which capital is raised using crowdfunding. Next to that, this paper will examine whether consumers are willing to pay more for crowdfunded products once there is more information provided about the amount of support for the crowdfunding project. It is possible that crowdfunding campaigns that were successfully completed will signal product quality and consumers are willing to pay more for the product. If this is the case, it might be beneficial to some companies to market their product as crowdfunded in order to influence the perceptions of consumers towards the product.

(6)

2. Literature review

This section will give an overview of the literature that is related to the subject of this research. Firstly, the nature of the new phenomenon crowdfunding will be discussed, thereafter an overview of the literature about signaling theory will be given and lastly the most relevant literature about herding behavior will be discussed.

2.1 Crowdfunding

Entrepreneurs of which their businesses are in the start-up phase often have difficulties to raise funds from traditional sources such as venture capitalists, business angels and banks. The new phenomenon crowdfunding is an alternative source to finance a project and has the potential to fill the finance gap for companies that are in an early stage of their life cycle (Mitra, 2012). Crowdfunding entails that many individuals finance one project, typically with relatively small contributions (Schwienbacher & Larralde, 2010). The concept of crowdfunding is close related to crowdsourcing, which basically entails that companies involve individuals (the crowd) by outsourcing specific tasks of the business activities with an open call over the Internet. So the crowd is besides an alternative source of finance also valuable to companies by contributing mostly voluntarily to production processes (Schwienbacher & Larralde, 2010).

Everyone with an idea and access to Internet is able to start a crowdfunding campaign. In general, individuals or companies submit their (business) plan on a crowdfunding platform. Project initiators are able to pitch their projects on the website of the crowdfunding platform in order to raise funds. The project initiators set, in accordance with the crowdfunding platform, the minimum amount of funding that is needed to realize the project, also referred to as the funding goal. After the funding goal is reached, the project initiator will receive the funds minus a fee for the crowdfunding platform. If the funding goal isn’t reached, the project initiator won’t receive the funds and the backers will be refunded (Belleflamme et al., 2014).

(7)

The returns for the backers vary greatly and therefore four main forms of crowdfunding can be distinguished: donation crowdfunding, reward/pre-purchase crowdfunding, lending crowdfunding and equity crowdfunding (Mitra, 2012).

Mostly the quality of a crowdfunding project can hardly be assessed in advance because there’re no prototypes of the product available and therefore backers have to look for signals of quality. Prior research already examined why some projects are more popular than others. Ward and Rachandran (2010) found that certain peer effects play an important role for backers. Backers base their funding decisions on actions of other backers and are strongly influenced by top-5 popularity lists in crowdfunding platforms. These findings suggest that backers show herding behavior and consumers might act in a similar way. The next paragraphs will elaborate on signaling theory and herding behavior.

2.2 Signaling theory

Whether a product is crowdfunded or not, can be a signal that affects consumers’ perceptions towards that product. Information about a product can function as a signal that can be used by consumers to assess the quality of a product. Signaling theory describes how consumer behavior is influenced by information about a product under specific conditions. Sometimes it can be very hard for a consumer to assess the quality of a product. Nelson (1970) distinguished two sorts of goods: search and experience goods. For search goods holds that consumers are able to know the quality of the product prior to purchase and use, whereas the quality of experience goods can only be assessed after the product is purchased and used. In case of experience goods, buyers and sellers act under conditions of asymmetric information. In most situations, the seller knows the actual quality of a product more or less. As mentioned before, buyers are often unable to know the quality of a product (or quality of certain properties of the product). Both low-quality sellers and high-quality sellers will try to

(8)

persuade consumers that they sell a high-quality product in order to increase demand and get a higher price for the product (Nelson, 1970).

Whether it’s true or not, sellers will always claim that they sell high-quality products, because this is beneficial to them. Akerlof (1970) introduced the concept adverse selection. This problem occurs when consumers have to deal with asymmetric information and can’t verify the quality of a product. Each market consists of cherries (high-quality products) and lemons (low-quality products). Since consumers are not able to verify the quality of a product, they will be willing to pay the average price for each product in the market to prevent themselves of overpaying for a low-quality product. Sellers with high-quality products will be driven out the market because it’s not worthwhile to them to sell their product for the average market price. As a result, a market full of lemons is created due to adverse selection (Akerlof, 1970).

To prevent problems with adverse selection, consumers will be likely to search for information that helps to (partially) reveal the actual quality of the product. Sellers will provide signals to consumers prior to the purchase (Boulding & Kirmani, 1993). Examples of signals that are used by sellers are price, advertising and warranties (Milgrom & Robert, 1986; Kihlstrom & Riordan, 1984; Spence, 1997)

Since there are no prior studies that have examined how consumers perceive crowdfunded products, there is a need to understand what signal goes out from crowdfunded products.

2.3 Herding behavior

In modern economy, companies constantly try to persuade consumers to buy their product by sending signals to the consumer. This causes that consumers are daily overloaded with an enormous amount of signals that they can’t process and verify one by one. As before mentioned, consumers are sometimes unable verify the quality of certain products prior to

(9)

consumption (Nelson, 1970). In order to solve this problem, consumers tend to make use of evaluations, intentions and purchase behaviors of other consumers (Huang & Chen, 2006). Consumers are able to reduce search costs by basing their purchase decision on evaluations, intentions and purchase behaviors of others (Banerjee, 1992). Park and Lessig (1977) found that consumers use judgments of others as a sign of product quality. This type of behavior is also referred to as Herding Behavior.

A familiar situation in which herding behavior occurs, is when people are searching for a place to eat and pass two restaurants that are next to each other. In most cases people don’t know the quality of the restaurants, so they decide to pick the one with most seats occupied. Besides popularity in terms of sales, other possible indicators of product quality are consumer reviews/ratings and expert recommendations (Huang & Chen, 2006).

Deutsch and Gerard (1955) claim that others can influence the decision-making process of an individual in two different manners: informational influence and normative influence. Informational influence refers to the propensity of an individual to accept information received from others as an indicator of reality. Normative influence is related to the extent that an individual wants to conform to the expectations of others. In an online setting, consumers don’t have a need to conform to expectations of others, so their decision-making process is mainly affected by informational influence (Huang & Chen, 2006). Although herding behavior can be efficient to consumers, it can also lead to less optimal outcomes (Zhang and Liu, 2012). In the online setting consumers often just imitate each other instead of basing their decision on available information (Bonabeau, 2004). When individuals don’t inform themselves properly, but solely make decisions based on evaluations, intentions and purchase behavior of others this can lead to informational cascades. The imitation of others is an explanation for market bubbles and hypes (Avery & Zemsky, 1998). For example, Dholakia and Soltysinski (2001) found that buyers that participate in digital auctions are

(10)

likely to bid for services and products that have already received bids by others, despite the presence of products and services that are more attractive, but didn’t receive bids by others.

If a crowdfunding project is considered to be a success in terms of financial support, this might influence consumers’ perception towards crowdfunded products as a result of herding behavior. Popularity in terms of support for a crowdfunding project could possibly signal product quality to consumers.

(11)

3. Theoretical framework

The next section will describe the literature regarding consumers’ perceptions towards crowdfunding and the effects of herding behavior in this context.

3.1 Role of consumer and perceptions towards crowdfunding

The role of the consumer within the value creation process of the company has been changing over the last decades. In the past, the roles of the company and the consumer were clearly distinct: the company produced a product and the consumer consumed it. Nowadays consumers want to be actively involved in the production process of a company in order to fulfill their wants and needs (Von Hippel, 2005). As consumer preferences for products are heterogeneous, a few sizes doesn’t fit all. Consumers are better able to customize products to their own needs if they collaborate closely with the manufacturer. Pierce et al. (2002) found that consumers are willing to pay more for products if they are personally involved in the production process and therefore have a feeling of psychological ownership. Being involved in the production process increases the feeling of consumers as if a product is theirs. Moreover Schreier et al. (2012) have identified that not only participating consumers are willing to pay more for a product of a user-driven company, but also non-participating consumers seem to prefer these products. Non-participating consumers are more able to identify themselves with user-driven firms, because they feel that the company tries to fulfill their needs (Schreier et al., 2012).

As before mentioned, crowdfunding builds upon crowdsourcing activities and typically involves their consumers/users within the production process (Kleemann et al., 2008). If non-participating consumers are willing to pay more for user-driven products, it might be that their willingness-to-pay is also positively affected if they know that crowdfunding is applied to manufacture a product. Ho et al. (2014) found that consumers have different perceptions for crowdfunded products than for conventional products, which also affects their purchase intention. Consumers perceive that crowdfunded products than for

(12)

conventional products. Next to that, consumers are more likely to purchase a crowdfunded product. Based on foregoing theory, this paper will test the following hypotheses:

H1a: Consumers are willing to pay more for successfully crowdfunded products

H1b: Consumers perceive that successfully crowdfunded products have a higher quality

If these hypotheses are true, it might be beneficial for some companies to communicate that their product is established with the help of crowdfunding in order to increase consumers’ willingness-to-pay.

3.2 Herding behavior in crowdfunding context

In the context of funding, there seems to be a pattern of herding behavior between the funders of crowdfunding projects (Agrawal et al., 2013; Ward & Rachandran, 2010). Zhang and Liu (2012) claim that once a project has already raised some capital, this will attract other backers more easily as it signals high quality of the project. Thus, potential backers are more likely to invest in that already raised a substantial amount of capital. This implies that funding success will only reflect underlying project quality if early backers do a good job assessing projects.

Herding behavior can be an efficient strategy in some situations, but it can also result in suboptimal outcomes. Zhang and Liu (2012) state that that the amount of raised capital can function as a credible signal of quality in an online lending setting. This is also referred to as rational herding. Furthermore, Burtch, Ghose, and Wattal (2013) show that in a crowdfunding setting, decisions of other can function as a signal of quality. Therefore crowdfunding also functions as a tool to reveal the market potential of a specific product. Crowdfunding platforms allow companies to raise more capital than the initial funding goal, which is also referred to as overfunding. If it’s correct that a large amount of accumulated capital by a

(13)

borrower, signals high product quality, it might also be that consumers are willing to pay more for the product. Therefore the following hypotheses are stated:

H2a: Once consumers know that a crowdfunding project is overfunded, they will be willing to pay more for the product.

H2b: Once consumers know the amount of backers that supported a crowdfunding project, they will be willing to pay more for the product.

H2c: Once consumers know that a crowdfunding project is overfunded, they will perceive that the product is of a higher quality.

H2d: Once consumers know the amount of backers that supported a crowdfunding project, they will perceive that the product is of a higher quality.

(14)

4. Methodology

This section will describe the set-up of the experiment and the techniques that are used in order to answer the research questions. Next to that, some general descriptive statistics of the sample will be given.

4.1 Research design

This study chooses to measure consumers’ willingness-to-pay in an experimental setting. An experimental design enables to manipulate certain aspects that might directly affect the willingness-to-pay for crowdfunded products. The data of the experiment will be collected via Alleeup (Alleeup.com). Alleeup is a platform that enables start-ups and designers to sell their product or service in an online auction setting. Furthermore, Alleeup is suitable as a tool to do (market) research.

The auctioned product is a t-shirt of with Edgar, which is financed with the help of crowdfunding platform WayV. The target of the project was €3750. During the crowdfunding campaign, 117 backers supported the project, which resulted in €3870 of capital. The t-shirt is available for males and females.

The experiment consists of two different stages. Firstly the participants of the auction have to place a bid on the product after they were randomly assigned to one of the four treatments. Each treatment provides different information to the participants. The data that is collected through the bids provide an insight in what effect each treatment has on the willingness-to-pay for products and services. The following treatments were shown to the participants (for more detailed information see appendix C):

Treatment 1: Only the basic information about the product

Treatment 2: Basic information + mentioned that the product was crowdfunded

Treatment 3: Basic information + mentioned that the product was crowdfunded + mentioned the funding goal and the amount of funding raised.

(15)

Treatment 4: Basic information + mentioned that the product was crowdfunded + mentioned the funding goal and the amount of funding raised + mentioned how many backers supported the crowdfunding campaign.

The first treatment contains no signal that the product was crowdfunded, each subsequently treatment contains more information about the crowdfunding project. As all factors, except for the different treatments, are held constant, variances in the willingness-to-pay and perceived quality for the t-shirt are directly attributable to the different signals provided in the advertisements.

After the participants have placed their bid they will continue to a short questionnaire, which contains questions about how they perceive the quality of the product, their experiences with crowdfunding and some demographic questions (for more detailed information see appendix B).

4.2 Research methodology

Various methods are used to measure ones willingness-to-pay (Vickrey, 1961; Jones, 1975; Kalish & Nelson, 1991). This thesis will use an experimental research design to measure consumers’ willingness-to-pay for crowdfunded products. The true monetary valuation of a consumer for a product can be revealed with an auction. The specific auction that is used in this thesis is similar to the Becker-DeGroot-Marshak (BDM) auction (Becker et al., 1961). This auction allows different participants to bid on a specific product, without knowing the bids of other participants. The participants are allowed to make a bid between zero and infinite. In this experiment, a sales price, which is unknown to the bidders, is set in advance. If the bid is higher than the sales price, the participant receives the product and has to pay an amount equal to the sales price. Obviously, when a bid is lower than the sales price, the participant doesn’t receive the product and won’t have to pay (Breidert et al., 2006).

(16)

According to Wertenbroch and Skiera (2002), the BDM auction is a valid instrument to measure consumers’ willingness-to-pay. Wertenbroch and Skiera (2002) asked people after a BDM auction if they were satisfied with the outcome. Both buyers and non-buyers indicated that they were very satisfied with the outcome of the experiment. Unlike the well-known Vickrey auction, the BDM auction doesn’t suffer from the overbidding bias, because its’ purchasing context is more close to reality compared to the more artificial purchase context of a Vickrey auction. In the setting of a Vickrey auction, participants tend to overbid because of the competitive environment in which the bidders have to act (Kagel, Harstadt, & Levin, 1987).

The dependent variables will be the consumers’ willingness-to-pay and perceived quality for the crowdfunded t-shirt. The two main independent variables will be whether the advertisement mentions that the product is established using crowdfunding and the amount of support in terms of funding and backers.

4.3 Sample

After filtering out the data of test-respondents and respondents that didn’t fully answer the questionnaire, the remaining sample contained 157 participants. The male/female proportion is equal to 44.6% males and 55.4% females. This is relatively close to the Dutch population distribution of 50.5% females and 49.5% males (Centraal Bureau Statistiek, 2015).

Table 1. Gender distribution

Gender

Population Sample Male 49.5% 44.6% Female 50.5% 55.4%

The age distribution of the sample is far from similar to the age distribution of the Dutch population (Centraal Bureau Statistiek, 2015). The age distribution is skewed to the right and most participants are between 20 and 40 years old. This overrepresentation is due to the data

(17)

collection procedure. A convenience sampling strategy was chosen, which means that the students, who were responsible for recruiting the respondents, mainly asked their friends and relatives to participate in the experiment. Additionally, some respondents were bought via Euroclix, an online respondent panel. The fact that the sample isn’t representative for the Dutch population might have a negative effect on the external validity of the research (Doane & Seward, 2005). Therefore this has to be taken in consideration when the results of this research are interpreted.

Table 2. Age distribution population and sample

Age <20 years 20-40 years 40-65 years >65 years Population 22.9% 31.8% 28.0% 17.3% Sample 1.9% 83.4% 13.4% 1.3%

With respect to education level, 88.3% of the sample has obtained a Bachelor’s degree or higher, which means that the participants of the experiment are more educated compared to the average individual in the Dutch population (Centraal Bureau Statistiek, 2013). This is also due to the earlier mentioned convenience sampling strategy. The non-representative sample has to be taken into account when the statistical test results are interpreted.

Table 3. Distribution of respondents over treatments

Treatment 1 Treatment 2 Treatment 3 Treatment 4

n 30 45 37 45

Percentage 19.1% 28.7% 23.6% 28.7%

As table 3 shows, the sample sizes for each treatment slightly vary. Although it would be better if the respondents were allocated over the four treatments, no problems for the statistical tests are expected.

(18)

5. Results

This section will present the statistical results that were extracted from the data. Firstly, some descriptive statistics related to normality, homogeneity of variance and scale reliability will be given for the bid amount and the perceived quality of the auctioned product. Thereafter the hypotheses of this research will be tested.

5.1 Descriptive statistics for willingness-to-pay

This section will describe the nature of the data for the willingness-to-pay. Normality and equality in variance will be checked for the bid amount of the four treatments. The mean bid value of treatment 1, 2, 3 and 4 are respectively €8.85, €8.19, €10.65 and €8.32. The frequency distribution histograms of the bid value (Appendix A) shows that all treatments are somewhat skewed to the right, which is supported by table 4. This is probably due to the large share of zero bids. To check for normality, firstly skewness and kurtosis are tested. Since all values for skewness and kurtosis are between -1 and 1, the data seems to fit normality standards, but after checking for normality with the Kolmogorov-Smirnov test and Shapiro-Wilk test, problems are encountered (table 5). All treatments are found to be significant (p<0.05), which means that normality for the bid amount can’t be assumed for any treatment. Table 4. Descriptive statistics of the bid amount with respect to the treatments

Treatment Mean Std. Dev. Min. Value

Max. Value Skewness Kurtosis Treatment 1 8.85 6.83 0 25 0.780 0.369 Treatment 2 8.19 7.15 0 25 0.564 -0.749 Treatment 3 10.65 7.77 0 31 0.661 -0.300 Treatment 4 8.32 7.56 0 25 0.439 -0.832

(19)

Table 5. Kolmogorov-Smirnov test and Shapiro-Wilk Test for normality bid amount

Kolmogorov-Smirnov test Shapiro-Wilk test Treatment Statistic df Sig. Statistic df Sig. Treatment 1 0.166 30 p < 0.05 0.922 30 p < 0.05 Treatment 2 0.144 45 p < 0.05 0.909 45 p < 0.01 Treatment 3 0.172 37 p < 0.01 0.929 37 p < 0.05 Treatment 4 0.167 45 p < 0.01 0.890 45 p < 0.001

In order to satisfy the normality conditions, another normality test was executed with only the top 50 percent of the bids included and next to that the natural logarithm of the bid amount was used. By only using the top 50 percent of the bids, zero bids are excluded from the data, which might help to meet the normality requirements. The results in table 7 show that normality for the data still can’t be assumed. According to the Kolmogorov-Smirnov test normality can only be assumed for treatment 3. The Shapiro-Wilk test shows that only treatment 3 is found to be normally distributed after using the top 50 percent of the bids transformed in its’ natural logarithm. Although the measures undertaken do slightly help to overcome problems with normality, it will decrease the power of the test results, because the sample size will decrease to n=90 (Doane & Seward, 2005). So for the remaining statistical tests the natural logarithm of the bid amount data is used (table 8). The non-normal distribution of the natural logarithm of the bid amount has to be taken in consideration when the results of this research are interpreted.

Table 6. Descriptive statistics of the natural logarithm of the bid amount with respect to the treatments

Max. Value Skewness Kurtosis Treatment 1 1.96 0.94 0 3.26 -0.897 -0.072 Treatment 2 1.78 1.07 0 3.26 -0.537 -1.050 Treatment 3 2.19 0.81 0 3.46 -0.605 -0.087 Treatment 4 1.69 1.22 0 3.26 -0.443 -1.527

(20)

Table 7. Kolmogorov-Smirnov test and Shapiro-Wilk test for normaility of the natural logarithm of the top 50 percentile of the bid amount

Table 8. Kolmogorov-Smirnov test and Shapiro-Wilk test for normality of the natural logarithm of the bid amount

Next to normality, in order to run a one-way ANOVA properly, it is assumed that the variances of the different samples are equal (Doane & Seward, 2005). In order to check this, a Levene’s test is carried out. The test shows that the variances are significantly different (F(3, 153)=5.552, p<0.001) This has to be taken in consideration when the results of the statistical test are interpreted.

5.2 Descriptive statistics for Perceived Quality

This section will describe the nature of the data for the perceived quality of the consumers for the auctioned product. Firstly the reliability of the measurement scale for perceived quality will be checked. Thereafter normality and equality in variance will be checked for the perceived quality.

Kolmogorov-Smirnov test Shapiro-Wilk test Treatment Statistic df Sig. Statistic df Sig. Treatment 1 0.252 15 p < 0.05 0.857 15 p < 0.05 Treatment 2 0.227 21 p < 0.01 0.903 21 p < 0.05 Treatment 3 0.155 20 p = 0.200 0.957 20 p = 0.482 Treatment 4 0.201 23 p < 0.05 0.911 23 p < 0.05

Kolmogorov-Smirnov test Shapiro-Wilk test Treatment Statistic df Sig. Statistic df Sig. Treatment 1 0.199 30 p <0.01 0.896 30 p <0.01 Treatment 2 0.164 45 p <0.01 0.888 45 p <0.001 Treatment 3 0.144 37 p=0.051 0.950 37 p=0.097 Treatment 4 0.208 45 p <0.001 0.818 45 p <0.001

(21)

In the questionnaire the respondents were asked to assess the quality of the t-shirt. To measure the Perceived Quality of the t-shirt respondents had to fill in three questions about fabric quality, design quality and general quality on a 5-point Likert scale. In order to measure the Perceived Quality of the t-shirt, one scale was created. To test the reliability of the scale, Cronbach’s Alpha is computed (table 7). The Cronbach’s Alpha for the 3-item scale is found to be 0.706. After controlling whether it was able to increase the reliability of the scale by deleting one of the items, it was found that the measurement scale for perceived quality would be more reliable if the item design quality was deleted. Hereby Cronbach’s Alpha increases to 0.755, which is more than the minimum of 0.7 and therefore the scale for perceived quality is considered to be reliable.

Table 9. Descriptive statistics of the perceived quality with respect to the treatments

Max. Value Skewness Kurtosis Treatment 1 3.40 0.65 2 4 -0.617 -0.685 Treatment 2 3.29 0.71 1.50 4.50 -0.398 -0.445 Treatment 3 3.43 0.60 2.00 4.50 -0.531 -0.189 Treatment 4 3.41 0.67 2.00 5.00 -0.138 -0.021

The mean for the perceived quality is quite similar for the different treatments, ranging from 3.29 to 3.43 (table 8). All values for skewness and kurtosis for the different treatments are between -1 and 1, so at first sight, no problems with normality are assumed. However, the Kolmogorov-Smirnov test and the Shapiro-Wilk test show very different results (table 9). According to both tests, all treatments are non-normally distributed. Therefore statistical results have to be interpreted carefully.

(22)

Table 10. Kolmogorov-Smirnov test and Shapiro-Wilk test for normality of the perceived quality

To check whether the variances for perceived quality are equal for the different treatments, a Levene’s test is used. The test (F(3, 152)= 0.516, p=0.679) shows that the variance can be assumed to be equal.

5.3 Hypotheses testing

To test the hypotheses that are related to the willingness-to-pay, the natural logarithm of the bid amount is used in order to come closer to a normal distribution. The first hypothesis (H1a) stated: Consumers are willing to pay more for crowdfunded products. In order to test the first hypothesis, a one-way ANOVA was performed. The one-way ANOVA is used to examine whether there is a significant difference between the mean of treatment 1 compared to treatment 2. Surprisingly the data showed that the treatment where it was mentioned that the product was crowdfunded (M=1.78, SD=1.07) was valued lower than the baseline treatment (M=1.96, SD=0.94). However no significant effect (F(1, 73)= 0.567, p=0.454) was found. The large amount of zero bids, assumes that a lot of respondents weren’t interested in the product and therefore probably didn’t take a good look at the advertisement. It might be that the effect of signaling that the product was crowdfunded couldn’t be measured for that reason. Hence, a one-way ANOVA was run with only the 50 percentile of the biddings. The

Kolmogorov-Smirnov test Shapiro-Wilk test Treatment Statistic df Sig. Statistic df Sig. Treatment 1 0.289 30 p <0.001 0.803 30 p <0.000 Treatment 2 0.180 44 p <0.001 0.928 44 p <0.01 Treatment 3 0.232 37 p <0.000 0.867 37 p <0.001 Treatment 4 0.175 45 p <0.001 0.927 45 p <0.01

(23)

50 percentile of the biddings also didn’t show significant results (p=0.467). Next to that, controlling for gender and education (Bachelors degree or higher) also didn’t yield significant results.

The second hypothesis (H1b) states: Consumers perceive that products have a higher

quality once they know it is successfully crowdfunded. To test this hypothesis, an independent

samples t-test was used. The results showed that treatment 1 (M=3.40, SD=0.65) and treatment 2 (M=3.29, SD=0.71) aren’t significantly different (F(1, 72)= 0.415, p=0.522), so no effect was found. Controlling for gender or education (Bachelors degree or higher) also didn’t yield any significant results.

To test whether the willingness-to-pay was affected by herding behavior of the respondents, the mean of the bid amount of the natural logarithm of treatment 3 (M=2.19, SD=0.81) and 4 (M=1.69, SD=1.22) was compared to the mean of treatment 2 (M=1.78, SD=1.07). A one-way ANOVA (F(1, 80)= 3.623, p<0.10) shows that there is a significant difference for the means of the bid amount of the natural logarithm. So, the bid amount is significantly higher when the funding goal and the amount of funding raised were mentioned. This corresponds to hypothesis H2a.

In order to test the effect of mentioning the amount of backers on the willingness-to-pay, the mean of the natural logarithm of the bid amount of treatment 3 (M=2.19, SD=0.81) is compared with treatment 4 (M=1.69, SD=1.22). A one-way ANOVA shows that the mean of treatment 3 is significantly higher than the mean of treatment 4 (F(1, 80)= 4.555, p<0.05). So, the opposite effect of hypothesis H2b is found, respondents valued the product lower if the amount of backers was mentioned.

To test whether the perceived quality was affected by herding behavior of the respondents, the means of the perceived quality of treatment 2 (M=3.29, SD=0.71), treatment 3 (M=3.43, SD=0.60) and treatment 4 (M=3.41, SD=0.67) were compared to each other. A

(24)

one-way ANOVA (F(1, 79)= 0.858, p=0.357) shows that there is no significant difference for the means of perceived quality of the treatment 2 and 3. Next to that, no significant difference (F(1, 80)= 0.023, p=0.881) was found for the means of treatment 3 and 4. Hence perceived quality is probably not affected by herding behavior, so both hypotheses H2c and H2d can be rejected.

(25)

6. Discussion

This section will firstly discuss the general findings and will compare the results to the literature related to this topic. Furthermore the theoretical and managerial implications will be discussed. In addition, the limitations of the research will be discussed and some suggestions for future research are presented.

6.1 General findings

The aim of this study was to examine whether signaling that a product is crowdfunded influences consumers’ willingness-to-pay and perceived quality. In addition this research examines whether signaling the popularity of the crowdfunding project affects the willingness-to-pay and the perceived quality of the consumer for the product.

The first hypothesis (H1a) stated: consumers are willing to pay more for successfully

crowdfunded products. According to the consulted literature, consumers are willing to pay

more for products when they’re personally involved in the business activities of the company (Pierce et. al, 2002). Schreier et al. (2012) found that also non-participating consumers are willing to pay more for products in which the user/consumer is involved. As crowdfunding is often associated with involvement of the consumer, a positive effect on the willingness-to-pay was expected. After running several statistical tests no significant differences are found for the different treatments. The second hypothesis (H1b) states: Consumers perceive that

products have a higher quality once they know it’s successfully crowdfunded. Similarly to the

results for willingness-to-pay, no difference for perceived quality towards crowdfunded products was found. Fuchs et al. (2013) claim that signaling a user-driven design might have a negative influence on the willingness-to-pay for luxury fashion brands. Although the auctioned t-shirt in this experiment can’t be entirely labeled as a luxury fashion brand, the same effect might also hold for regular fashion brands.

Park and Lessig (1977) claim that consumers often demonstrate herding behavior when making a purchase decision. Consumers base their purchase decision on evaluations,

(26)

intentions and purchase behavior of other consumers. Hypothesis H2a states: Once consumers

know that a crowdfunding project is overfunded, they will be willing to pay more for the product. This hypothesis is supported by the statistics in this research. The author expected

that adding an extra signal that displays the popularity of the crowdfunding campaign would lead to a higher willingness-to-pay for the crowdfunded product. Therefore hypothesis H2b was stated: Once consumers know the amount of backers that supported a crowdfunding

project, they will be willing to pay more for the product. Surprisingly the opposite effect was

found. A possible explanation for this finding is that the support for the auctioned product was relatively small, containing only 117 backers. Hence, the respondents might have perceived that the crowdfunding campaign was unsuccessful and therefore valued the crowdfunded product lower. Initially, a variable was included to measure the degree to which the respondents perceived whether the crowdfunding campaign was a success. Unluckily, due to a mistake in the data collection procedure it wasn’t realized.

Already having concluded that hypothesis H2a is supported by this research and as willingness-to-pay and perceived quality are two related constructs, it would be obvious that hypothesis H2c is also true: Once consumers know that a crowdfunding project is overfunded,

they will perceive that the product is of a higher quality. However no significant effect was

found. A possible explanation might be that the surplus that consumers are willing to pay for crowdfunded products isn’t driven by an expected surplus in quality, but because of a coolness factor of crowdfunded products.

The data also didn’t support hypothesis H2d: Once consumers know the amount of

backers that supported a crowdfunding project, they will perceive that the product is of a higher quality.

6.2 Theoretical and managerial implications

(27)

this topic. The existing crowdfunding literature is predominantly about why and how crowdfunding can be a useful source of finance. Very little is known about the perceptions of the consumer towards crowdfunded products. This study tries to reveal whether consumers valuate crowdfunded products differently than ordinary products. Contrary to what was expected, there is no evidence that crowdfunded products are valued more. However, consumers seem to demonstrate certain herding behavior in the crowdfunding context. Consumers valuate crowdfunded products higher when the amount of funding is signaled. In contrast, mentioning the amount of backers might have a negative influence on the willingness-to-pay.

Knowing that in 2025 the crowdfunding market is estimated to become twice the size of the venture capital industry today (The World Bank, 2013), there is a need for a better understanding of the concept of crowdfunding. The outcomes of this study provide valuable insights for managers that sell crowdfunded products. This research implies that it would be beneficial to communicate the financial support that has been obtained during the crowdfunding campaign to the consumer. This will increase consumers’ willingness-to-pay for the product and hence profits will increase.

6.3 Limitations and suggestions for future research

In order to put the results of this research in the right context, the limitations have to be discussed. First of all, the gathered sample is not representative for the Dutch population. The age distribution of the sample is very different from the age distribution of the population. The age category 20-40 years is disproportionally overrepresented. Furthermore the sample is overqualified compared to the national average in terms of education. The non-representative sample limits this study to generalize for the entire population. Hence external validity of the results can be considered to be low (Doane & Seward, 2005). The non-representative sample is due to a convenience sampling strategy. The vast majority of the sample consists of friends

(28)

and relatives of the author. Future research could focus on a sample that is more representative for the Dutch population.

The data that is used to measure willingness-to-pay contains a lot of zero bids or bids that are close to zero. Therefore the data is positively skewed and after running Kolmogorov-Smirnov and Shapiro-Wilk tests, the data was found to be non-normally distributed. Only using the top-50 percentile of the bids and transforming the data to its’ natural logarithm didn’t help to overcome problems with non-normality. Hence, for the statistical tests all data was used and transformed to the natural logarithm. Filtering the lower bids out, would lead to a small sample size and therefore a lower power of the tests. The problems with normality, should be taken in consideration when the results are interpreted. In order to get normally distributed data, future research could work with a larger sample size. Another violation of an assumption of a one-way ANOVA is equality of variance. This might result in incorrectly significant differences in the means. Future research should aim to have equal sample sizes, as this would mitigate problems with inequality of variances. In addition, a bigger sample size might also help to overcome this problem (Doane & Seward, 2005).

The data for consumers’ perceived quality towards the crowdfunded product is similarly to the data for the willingness-to-pay non-normally distributed. The same measures that are mentioned before, hold to overcome the problems with normality.

Due to a mistake in the data collection procedure, this research failed to measure the degree to which the respondents perceived that the crowdfunding project was a success. If future research would include this variable, it would be able to get a better understanding of the herding effects on the willingness-to-pay and perceived quality of the consumer.

(29)

7. Conclusion

The aim of this study was to examine whether signaling that a product is successfully crowdfunded influences consumers’ willingness-to-pay and perceived quality. In addition, this research examines whether signaling the popularity of the crowdfunding project affects the willingness-to-pay and the perceived quality of the consumer for the product. It was hypothesized that consumers are willing to pay more for successfully crowdfunded products. Moreover, it was hypothesized that consumers perceive that a product is of a higher quality once they know it’s crowdfunded. Finally, it was tested whether signaling the popularity of a crowdfunding project in terms of amount of backers and funding has an effect on consumers’ perceived quality and willingness-to-pay for a product. The findings of this research show that there is no increase in willingness-to-pay and perceived quality for the product once it’s signaled that a product is crowdfunded. Next to that, it is found that consumers’ perceived quality is not affected by herding behavior. However, consumers are willing to pay more once they know that the crowdfunding project is overfunded. A contrary effect is found for willingness-to-pay once consumers know how much backers supported the crowdfunding project.

(30)

References

Agrawal, A. K., Catalini, C., & Goldfarb, A. (2013). Some simple economics of

crowdfunding. National Bureau of Economic Research.

Akerlof, G. A. (1970). The market for" lemons": Quality uncertainty and the market mechanism. The quarterly journal of economics, 488-500.

Avery, C., & Zemsky, P. (1998). Multidimensional uncertainty and herd behavior in financial markets. American economic review, 724-748

Banerjee, A. V. (1992). A simple model of herd behavior. The Quarterly Journal of

Economics, 797-817.

Becker, G. M., DeGroot, M. H., & Marschak, J. (1964). Measuring utility by a single‐ response sequential method. Behavioral science, 9(3), 226-232.

Belleflamme, P., Lambert, T., & Schwienbacher, A. (2010). Crowdfunding: An industrial organization perspective. In Prepared for the workshop Digital Business Models:

Understanding Strategies’, held in Paris on June (pp. 25-26).

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

Bonabeau, E. (2004). The perils of the imitation age. Harvard business review, 82(6), 45-54. 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.

Breidert, C., Hahsler, M., & Reutterer, T. (2006). A review of methods for measuring willingness-to-pay. Innovative Marketing, 2(4), 8-32.

Burtch, G., Ghose, A., & Wattal, S. (2013). An empirical examination of the antecedents and consequences of contribution patterns in crowd-funded markets. Information Systems

Research, 24(3), 499-519.

Centraal Bureau Statistiek (2013). Opleidingsniveau. URL:

http://statline.cbs.nl/StatWeb/publication/?VW=T&DM=SLNL&PA=71822NED&D1=0&D2

=a&D3=0&D4=1,3-9&D5=0-2,4,8-10&D6=0&D7=l&HD=081216-1451&HDR=T,G6,G2,G5,G3&STB=G1,G4. Retrieved at June 16, 2015. Centraal Bureau Statistiek (2015). Bevolkingscijfers. URL:

http://statline.cbs.nl/StatWeb/publication/?VW=T&DM=SLNL&PA=03759ned&LA=NL. Retrieved at June 16, 2015.

Dholakia, U. M., & Soltysinski, K. (2001). Coveted or overlooked? The psychology of bidding for comparable listings in digital auctions. Marketing Letters, 12(3), 225-237. Doane, D. P., & Seward, L. E. (2005). Applied statistics in business and economics. USA:

(31)

Franke, N., Schreier, M., & Kaiser, U. (2010). The “I designed it myself” effect in mass customization. Management Science, 56(1), 125-140.

Fuchs, C., Prandelli, E., Schreier, M., & Dahl, D. W. (2013). All that is users might not be gold: How labeling products as user designed backfires in the context of luxury fashion brands. Journal of Marketing, 77(5), 75-91.

Gerber, E. M., & Hui, J. (2013). Crowdfunding: Motivations and deterrents for participation.

ACM Transactions on Computer-Human Interaction (TOCHI), 20(6), 34.

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

Ho, H. Y., Lin, P. C., & Lu, M. H. (2014). Effects of Online Crowdfunding on Consumers’ Perceived Value and Purchase Intention. Anthropologist, 17(3), 837-844.

Huang, J. H., & Chen, Y. F. (2006). Herding in Online Product Choice. Psychology &

Marketing, 23(5), 413-428.

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

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

Jones, D. F. (1975). A survey technique to measure demand under various pricing strategies.

The Journal of Marketing, 75-77.

Kagel, J. H., Harstad, R. M., & Levin, D. (1987). Information impact and allocation rules in auctions with affiliated private values: A laboratory study. Econometrica: Journal of the

Econometric Society, 1275-1304.

Kalish, S., & Nelson, P. (1991). A comparison of ranking, rating and reservation price measurement in conjoint analysis. Marketing Letters, 2(4), 327-335.

Kihlstrom, R. E., & Riordan, M. H. (1984). Advertising as a Signal. The Journal of Political

Economy, 427-450.

Kleemann, F., Voss, G. & Rieder, K. (2008). “Un(der)paid Innovators: The Commercial Utilization of Consumer Work through Crowdsourcing.” Science, Technology & Innovation

Studies 4:1, 5-26.

Li, G., Li, G., & Kambele, Z. (2012). Luxury fashion brand consumers in China: Perceived value, fashion lifestyle, and willingness to pay. Journal of Business Research, 65(10), 1516-1522.

Milgrom, P., & Roberts, J. (1986). Price and advertising signals of product quality. The

Journal of Political Economy, 796-821.

Mitra, D. (2012). The role of crowdfunding in entrepreneurial finance. Delhi Business

(32)

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

Business Venturing, 29(1), 1-16.

Nelson, P. (1970). Information and consumer behavior. The Journal of Political Economy, 311-329.

Park, C. W., & Lessig, V. P. (1977). Students and housewives: Differences in susceptibility to reference group influence. Journal of consumer Research, 102-110.

Pierce, J. L., Kostova, T., & Dirks, K. T. (2003). The state of psychological ownership: Integrating and extending a century of research. Review of general psychology, 7(1), 84. Schreier, M., Fuchs, C., & Dahl, D. W. (2012). The innovation effect of user design: Exploring consumers' innovation perceptions of firms selling products designed by users.

Journal of Marketing, 76(5), 18-32.

Schwienbacher, A., & Larralde, B. (2010). Crowdfunding of small entrepreneurial ventures.

Handbook of Entrepreneurial Finance, Oxford University Press, Forthcoming.

Spence, M. (1977). Consumer misperceptions, product failure and producer liability. The

Review of Economic Studies, 561-572.

The World Bank (2013). Crowdfunding’s potential for the developing world.

Thies, F., & Wessel, M. (2014). The Circular Effects of Popularity Information and Electronic Word-of-Mouth on Consumer Decision-Making: Evidence from a Crowdfunding Platform. Vickrey, W. (1961). Counterspeculation, auctions, and competitive sealed tenders. The

Journal of finance, 16(1), 8-37.

Von Hippel, E. (2005). Democratizing innovation: The evolving phenomenon of user innovation. Journal für Betriebswirtschaft, 55(1), 63-78

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

Crowds at NIPS2010.

Wertenbroch, K., & Skiera, B. (2002). Measuring consumers’ willingness to pay at the point of purchase. Journal of Marketing Research, 39(2), 228-241.

Zhang, J., and P. Liu (2012): “Rational Herding in Microloan Markets,” Management Science, 58(5), 892–912.

(33)

Appendix A

(34)

(35)

(36)

(37)

Appendix B

The questionnaire contained the following questions:

I think the quality of the fabric quality of the t-shirt is high

5-point Likert scale: totally disagree - disagree – neutral – agree – totally agree I think the quality of the design of the t-shirt is high

5-point Likert scale: totally disagree - disagree – neutral – agree – totally agree I think the general quality of the t-shirt is high

5-point Likert scale: totally disagree - disagree – neutral – agree – totally agree I think the crowdfunding campaign was a success

5-point Likert scale: totally disagree - disagree – neutral – agree – totally agree I’m familiar with the concept crowdfunding

(38)

Appendix C

The following advertisements were shown to the respondents.

Treatment 1 (baseline)

Text: Garandeer jezelf van een vrolijke dag door het dragen van een Edgar shirt.

Treatment 2

Text: Garandeer jezelf van een vrolijke dag door het dragen van een Edgar shirt. Dit shirt is middels een crowdfundingcampagne gefinancierd.

(39)

Treatment 3

Text: Garandeer jezelf van een vrolijke dag door het dragen van een Edgar shirt. Dit shirt is middels een crowdfundingcampagne gefinancierd. Er werd een bedrag van €3870 opgehaald waardoor de financieringsdoelstelling van €3750 werd behaald.

Treatment 4

Text: Garandeer jezelf van een vrolijke dag door het dragen van een Edgar shirt. Met de steun van 117 mensen werd een bedrag van €3870 opgehaald waardoor de financieringsdoelstelling van €3750 werd behaald.

Consumers’ perceptions towards crowdfunded products