The wisdom of the crowd in funding I nformation heterogeneity and social networks of crowdfunders

Crowdfunding has enabled large crowds to fund innovative projects. This type of funding might tap into the wisdom of crowds who were previously disconnected from the funding process. We distinguish between in-crowd and out-crowd funders (with and without ties to project creators) in order to test for heterogeneity in their information use. Based on the analysis of a large-scale survey amongst project funders, this paper shows that in-crowd investors rely more on information about the project creator than out-crowd investors. Out-crowd investors do not seem to attach more importance to information about the project itself than in-crowd investors, except in the case of donation-based crowdfunding. For financial-return crowdfunding, financial information becomes less important once a strong relationship with the project creator is established. Our study allows project creators to target information to specific audiences based on their relationship strength across different types of crowdfunding projects. the funders’ utility. We explain this from a motivation perspective.


Introduction
The funding of innovative start-ups has always been challenging due to a lack of track record, collateral and technological uncertainty (Engel & Stiebale 2014;Hall 2002;Giudici & Paleari 2000). More generally, small and medium sized firms face greater capital constraints than large firms, lacking access to market-based funding due to the high fixed costs associated with issuing equity and the unwillingness of institutional investors to take small holdings. This leaves start-ups highly dependent on bank credit, venture capital funds, angel investors and bootstrapping for their liquidity needs (Chittenden et al. 1996;Ebben and Johnson 2006;Giudici and Paleari 2000;Keasey and McGuinness 1990). Access to bank credit has become more transactional in recent decades with increased centralization and computerized assessment of creditworthiness (Bhidé 2010), and is often restricted due to a lack of profit and collateral. This shift severely affects innovative small firms due to their disproportionate reliance on soft information in the lending process (Brancati 2014;Cosci et al. 2016). Furthermore, the willingness of venture capitalists to fund start-ups is often limited to certain sectors (Huyghebaert et al. 2007) and there is evidence that the financial crisis has dampened their willingness to invest, particularly in follow-up rounds (Block and Sandner 2009;Cowling et al. 2016;Migendt et al. 2014). Structural financing constraints for small firms impede economic growth when firms downplay their growth strategy to match available funds (Beck and Demirguc-Kunt 2006; Binks and Ennew 1996;Chittenden et al. 1996;Rostamkalaei and Freel 2015).
The rise of crowdfunding over the past decade in part addresses this funding gap by offering entrepreneurs an alternative to traditional finance channels. Crowdfunding caters well to innovative, opaque, small firms and makes use of social networks in the funding process (Colombo et al. 2015;Vismara 2016). It builds on and expands beyond the traditional 'in-crowd' of family and friends by allowing both in-and out-crowd investors to provide finance through digital platforms (Bruton et al. 2015;Salomon 2016). Furthermore, it has lowered the transaction costs for entrepreneurs to collect small investment amounts from a dispersed set of investors and is becoming an increasingly sizable source of funding for start-ups and other bottom-up initiatives in the economy (Massolution 2015; Wardrop et al. 2015). However, it is unclear whether crowdfunding provides access to the wisdom of the crowd, or whether it opens up a wider audience of fools alongside the usual family and friends in-crowd.
In line with the growth of crowdfunding, academic research directed at understanding this phenomenon has emerged in recent years (Moritz and Block 2016). Much of this literature focuses on success factors driving crowdfunding campaigns, such as the role of early contributions (Agrawal et al. 2015;Cholakova and Clarysse 2015;Colombo et al. 2015). There is also considerable attention on the role of social networks in crowdfunding (Agrawal et al. 2015;Horvát et al. 2015;Hui et al. 2014) and on overcoming informational asymmetries (Ahlers et al. 2015;Lin et al. 2012;Vismara 2015). Lacking attention until now is the bridge between these two topics, namely how social networks affect the type of information used by investors in crowdfunding decision-making. Although there are suggestions regarding crowdfunding information mechanisms and the role of social networks (Ter Wal et al. 2016), there is little empirical evidence about the type of information that funders use to make investment decisions. Are crowdfunders well informed about the project they invest in, or are they jumping on a band-wagon set in motion by other investors in a campaign?
This study offers the first detailed empirical analysis on heterogeneity in information use by crowdfunders and how this is affected by their social networks. The ability to distinguish between investors based on their interpersonal ties to the entrepreneur offers insights into the application of theories about information asymmetries and social networks in funding decisions, and serves as input for public policy for entrepreneurship and finance. Our main research question is: How does the type of information used by crowdfunders vary with the strength of their ties to project creators?
This article is structured as follows: first, we review the relevant literature and introduce the theoretical framework. Next, we present the research design including our quantitative research approach and data. We then display the results which form the basis for the conclusions in the final section.

Literature review and theoretical framework Signalling in early-stage finance and information cascades
The way entrepreneurs obtain capital when forming a new firm has important implications for future performance (Bosma et al. 2004;Cassar 2004). Their search for external finance is characterised by agency problems between the entrepreneur and funder due to information asymmetries that lead to adverse selection and moral hazard (Denis 2004;Jensen and Meckling 1976;Parker 2009). This is especially the case for new firms that face high financing costs (Rostamkalaei and Freel 2015) driven by cumbersome information gathering, a lack of track record and, often, collateral (Blumberg and Letterie 2007;Cassar 2004).
Scholars suggest signaling can overcome these agency problems (Akerlof 1970;Amit et al. 1990;Gompers 1995;Myers and Majluf 1984;Stiglitz and Weiss 1981). Signaling can take place using different kinds of information, for example the availability of patents and prototypes, or the track record of entrepreneurial team (Audretsch et al. 2012;Becker-Blease and Sohl 2015;Busenitz et al. 2005;Gompers and Lerner 2001;Spence 1973). Many studies in the signaling literature establish a positive relationship between early-stage investments and firm success (Bernstein et al. 2015a;Bosma et al. 2004;Kerr et al. 2014;Kortum and Lerner 2000;Samila and Sorenson 2010)  Crowdfunding 1 , as a new form of seed finance, acts as a platform (agent) between investors and entrepreneurs (Bruton et al. 2015;Cumming, Pandes, et al. 2015;Harrison 2013;Salomon 2016 (Bruton et al. 2015). The quality of these signals as input into investment decisions is questionable since the crowd might not have expertise in production, marketing and competition, nor are they likely to invest in due diligence given high fixed costs (Belleflamme et al. 2013;Vismara 2016). As such, the wisdom of the crowd is not self-evident. On the one hand, the crowd could represent new customers, delivering knowledge about the market potential of an offering by signing up as  (Bikhchandani et al. 1992). Hornuf and Schwienbacher (2015) find that specific kinds of information, such as updates to investors, significantly drive investment as funders update their preferences in the light of project assessment. Moritz et al. (2015) examined investor communication in equity crowdfunding, highlighting that perceived sympathy, openness and trustworthiness in the relationship between venture and investor reduced perceived information asymmetries. They also found that third-party communication influences the decision making process of crowdfunders. Furthermore, allowing crowdfunders to adjust privacy settings regarding information about their contribution deters some investors but increases average contribution size (Burtch et al. 2015).
This suggests that some form of quality signalling between project creator and crowdfunder occurs which relates to the general notion of the 'wisdom of the crowd' in funding decisions (Mollick and Nanda 2015;Surowiecki 2005). But how does the crowd gather its 'wisdom'? Literature on investment processes suggest that this is facilitated by the social networks of both entrepreneur and investor (Alexy et al. 2012;Colombo et al. 2015;Ter Wal et al. 2016;Uzzi 1999).

Ties that bind, ties that blind: Social networks and information
Social networks strongly influence an entrepreneur's funding success as these provide access to resources such as finance, knowledge and partners (Davidsson and Honig 2003;Dubini and Aldrich 1991;Huang and Knight 2015;Kwon and Arenius 2010;Shane and Cable 2002). Social network theory provides a possible lens to study the role of information in the relationship between funder and venture (Granovetter 1973;Hoang and Antoncic 2003;Jack and Anderson 2002;Kwon and Arenius 2010;Uzzi 1999). Granovetter (1973Granovetter ( , p. 1361 defines the notion of 'strength' of interpersonal ties based on 'a combination of the amount of time, the emotional intensity, the intimacy and the reciprocal services which characterize the tie'.
Social networks, comprising both strong and weak ties, may affect the type of information used in a financing decision through three mechanisms.
First, the funder's motivation for investing, for example for financial return or to strengthen an existing relationship, will affect the information required (Belleflamme et al. 2014;Shane and Cable 2002). Second, the extent to which interpersonal ties develop and enforce common norms of behaviour will affect the perceived moral hazard of an investment (Bernstein et al. 2015a;Granovetter 2005;Uzzi 1999). This may make obtaining information about the entrepreneur more attractive than information about the project, its objectives, risk and finance. Third, the way in which quality signals are disseminated and received may vary Furthermore, the relationship between funders and project creators affects investment sequencing through information cascades. Individual funders possess different levels of information, hence some investors have an advantage over others (Cumming et al. 2015a;Hildebrand et al. 2016).
When professional investors with industry experience and track-record enter relatively early in a crowdfunding campaign, their public visibility attracts other investors (Vismara 2015), in a similar way as in other online market places (Dellarocas 2003;Lin et al. 2012). This suggests that the quality indication process with crowdfunding is staged, with an in-crowd to out-crowd sequence, using different types of information and levels of expertise to make a funding decision.

In-crowd information needs
We define the in-crowd as those project funders who have strong or weak interpersonal ties with the project creator. On crowdfunding platforms, investors base their decisions on information provided by the project creator in the form of updates during the campaign and on the investment behavior and comments of other crowd investors (Hornuf and Schwienbacher 2015). In-crowd information requirements could be affected by the three mechanisms outlined above: funder motivation, project creator intentions and information flow.
Firstly, the in-crowd may have different motivations than wanting to contribute to a successful project, such as reinforcing their relationship with the project creator, social obligation or altruism (Belleflamme et al. 2014;Gartner et al. 2011;Klyver et al. 2016;Shane and Cable 2002).
This could make them less inclined to search for quality signals about the project itself, and focus more on information about the person behind the project. Secondly, we expect that funding decisions embedded within a social network will decrease fears of negative behaviour by the project creator (Bernstein et al. 2015a;Granovetter 1985;Uzzi 1999  Additionally, relationships may imply a longer term commitment to the entrepreneur and therefore a longer term perspective on the costs and benefits of investing in information gathering about the entrepreneur (Boot 2000;Brancati 2014;Scholtens 1999).

H1
In-crowd funders are more likely to rely on information about the person(s) behind the project than out-crowd funders.

Out-crowd information needs
We define the out-crowd as those project funders who have no personal ties to the project owner. We expect this to lead to different information needs through the same three mechanisms. First, without the funding decision embedded in a social relationship, the motivation is more likely to be based on expected results, such as financial return (Cholakova and Clarysse 2015), a finished product or societal impact rather than social capital (Apinunmahakul and Devlin 2008) or community benefits (Belleflamme et al. 2014). Information about the project, its objectives, finance and risk will be more relevant as it gives insight into the expected return of the project (Ahlers et al. 2015; Belleflamme et al. 2013Belleflamme et al. , 2014. Secondly, information gathering about the project team is unlikely to reduce moral hazard as there is no relationship to enforce social reward or punishment (Belleflamme et al. 2014;Vismara 2016). Third, as the outcrowd lacks direct insights from the project creator, they depend on information that reaches them through formal direct (project websites, newsletters) or indirect (media) channels (Hornuf and Schwienbacher 2015). Information about the project creator obtained through formal channels is often perceived as less trustworthy and more difficult to interpret as a quality signal than when obtained through interpersonal ties.
As such, it loses its advantage over more general information about the project and its objectives (Hornuf and Schwienbacher 2015;Vismara 2016). We expect out-crowd funders to be less motivated than in-crowd funders to gather information about the project team and to instead focus H2 Out-crowd funders are more likely to rely on information about the project and its objectives than in-crowd funders.
Furthermore, we expect out-crowd funders to rely more on information about financial planning and risk than in-crowd funders due to stronger instrumental (results-based) motivation and a lack of personal access to the project owner. A recent study on equity crowdfunding shows that the Besides proving a quality signal, information about financials and risk can also reduce the perceived risk of moral hazard by revealing the commitment level of the project creator, such as whether or not they provide personal collateral and/or invest their own resources (Blumberg and Letterie 2007). We therefore expect that out-crowd funders rely more on information about financial planning and risks than in-crowd funders, looking both for quality signals and to reduce perceived moral hazard risk.
H3 Out-crowd funders are more likely to rely on information about financial planning and risks than in-crowd funders.

Research design
In this paper, we seek to understand the effect of the strength of interpersonal ties on the information used by crowdfunders. In order to test the hypotheses formulated above, we constructed the analytical model presented in Figure 1. Most of the literature to date uses projectlevel investment data that includes varying degrees of information about the project and its creator, however, this type of data does not convey much information about the project funders themselves. To analyse the hypothesized relations, we used a large-scale survey of crowdfunders (Cholakova and Clarysse 2015;Mollick 2015). We note that this methodological approach is potentially vulnerable to common method bias (i.e. gathering all information for this analysis via one survey) which has been shown to affect survey data (Podsakoff et al. 2003). Whilst we could not conceptually identify any underlying factors that the predictor and criterion variables had in common, we adopted several measures to reduce potential bias. We started by minimizing item ambiguity which included avoiding vague concepts, complicated syntax and unfamiliar terms. We deliberately used simple, specific and concise questions to measure the constructs. The respondents were also guaranteed anonymity.

Data
We use data from a large-scale survey called the 'National Crowdfunding

Independent variables
To determine the influence of interpersonal ties on the information use of funders, we included the relationship to project creator as an independent variable ('What was your relationship with the project owner or business owner before making your financial contribution through crowdfunding?').
We combine the individual answer categories to create new variables measuring relationship strength, aggregating different types of relationship to strong ties, weak ties or no ties (a smiliar approach has been taken by Klyver et al. 2016). 'Strong ties' included family, friends, initiator of the project or employee (Kuppuswamy and Bayus 2015).
'Weak ties' consists of people who indicate that they know the person behind the project or are a friend of friend, a business relationship, customer, fan or visitor (Bruton et al. 2015

Control variables
To account for the effect of other characteristics of the funders, we include a number of control variables from the survey such as age, gender (Klyver et al. 2016), education level, (Mollick 2014), type of project invested in, amount funded, type of return (donation, in-kind, financial) (Vismara 2015(Vismara , 2016, motivation, investment of others and risk awareness. Following earlier work (Calic and Mosakowski 2016;Hörisch 2015), we distinguish between for-profit, social, cultural and ecological projects and coded all projects into these categories as follows: (1) For-profit, (2) Social, (3) Cultural, (4) Ecological. Multiple answers were not coded. We asked an external researcher to validate our coding and used this feedback to improve our coding process (Patton, 2002). If there was only a description of the specific project (without a name) we searched for a crowdfunding project which matched that description and the time period, and if we found a plausible match, we coded this project.
By including instrumental (vs. value-based) motivation as a control variable, we control for one of the mechanisms through which we expect relationship strength to influence the type of informational need. We do this in order to focus on the behavioral intention of the project funder and quality signals as key mechanisms to overcome informational asymmetries in our model (Vismara 2016). We use 'importance of security of getting a promised return' (securityreturn) as a proxy for instrumental motivation. Consistent with cognitive evaluation theory, the intrinsic motivation of lenders to provide capital is undermined when entrepreneurs focus on future extrinsic rewards associated with lending (Allison et al.

2015)
. We also control for the influence of others investing in the project (herding effect) (Bikhchandani et al. 1992;Vismara 2015) by including the variable 'knowing the financial contributions made by others' (knowingfincontriboth) in our analysis.
Finally, we control for risk awareness (professionalism) of crowdfunders, since we expect experienced investors to use more information than amateur investors. We use the statement 'I keep in mind the consideration that to invest through crowdfunding in a company can be a high risk investment' as a proxy for risk awareness.

Data analysis
Most variables were assessed on a 5-point Likert scale (Dillman 2000).
The level of ties (strong, weak and no ties) and type of crowdfunding (donation, reward and financial return) were entered as dummy variables, with reward-based crowdfunding being the reference case. Amount invested, gender, education and social media types have different scales.
The data analysis was conducted in several steps (Hair 2010 differ with respect to: motivation, objectives, amount invested, personal characteristics, etc., followed by a more structured correlation analysis (see Table A.2). Finally, as our dependent variable is of ordinal nature, we

Determinants of information use of crowdfunders
Our models (1-6, see Table 1) allow analysis of the importance of several types of information used by crowdfunders according to relationship strength between funder and project owner. We enter both strong and weak ties into the regression as dummy variables, using no ties as a reference case. Our results show that relationship strength has significant effects on the importance of different types of information.
First, our regression model shows that funders with strong or weak ties attach significantly higher importance to information about the project creator and their previous projects than funders with no ties. This supports our hypothesis 1 (H1). We differentiate this result across crowdfunding types in two steps. As a first step, in our regression model we add dummy variables for both financial return and donation crowdfunding, using reward-based crowdfunding as a base case (this is the largest sample). We find significantly higher information is required about the person and their previous projects for both financial return and donation crowdfunding compared to the reference reward crowdfunding case (independent of ties). As a second step, to analyse the effect of ties on information needs within each type of crowdfunding, we computed the full model again specifically for the subsets of donation-based, rewardbased and financial-return crowdfunding respectively (Tables A.3 in information needs about the project creator compared to those without ties. We therefore conclude that relationship strength drives an increased need for information about the project team, in particular for reward and financial (debt and equity) crowdfunding. Second, only for donation crowdfunding do we find evidence that outcrowd funders rely more on information about the project and its objectives compared to in-crowd funders (an effect in line with hypothesis 20 2). This is driven by the significantly lower need for information about the project and its objectives in donation crowdfunding by funders with weak ties, who rely less on this information than those with strong ties or no ties (a U-shaped relationship between project information need and the strength of ties).
Overall, and for reward and financial return crowdfunding individually, we find no evidence that out-crowd funders rely more on information about the project than in-crowd funders. We therefore reject our second hypothesis (H2) both for our aggregated model and for reward and financial return crowdfunding; a higher information need about the project and its objectives only holds for those funders with no ties participating in donation crowdfunding, in relation to funders with weak ties in donation crowdfunding.
Third, we find evidence that in financial return crowdfunding, out-crowd funders rely more on information about financial planning and risk than incrowd funders. This result is driven mostly by funders with strong ties, who indicate a significantly lower information need for financial planning and risk than funders with no ties. This decreased information need is not observed for funders with weak ties. For donation and reward crowdfunding, and in our model that includes all types of crowdfunding, we find no significant differences in information needs about financial planning and risks for any strength of ties. Hence our hypothesis 3 (H3) is supported for financial return (debt and equity) crowdfunding and rejected for reward and donation crowdfunding.

Types of crowdfunding projects
We also investigated the influence of different types of projects on the use of information about the project, entrepreneur and financial planning and risks by funders with different strength of ties. We carried out this analysis by adding project type dummies to the full model (profit, social, ecological and cultural). First, we find no influence of project type on the information need about the project owner. Second, we find that the importance of information about the project and its objectives varies with the project type. In for-profit campaigns funders attach less importance to information about the project and its objectives. In campaigns with an ecological purpose, this effect is reversed. These effects are consistent across all relationship types. In donation-based and reward-based crowdfunding, the coefficients for both for-profit and ecological projects are higher. The importance of information about the project and its objectives is high in donation-based crowdfunding for ecological projects.
Also, in the presence of strong ties, the negative coefficient for for-profit projects disappears. Third, the importance of information about finance and risks does not vary with the type of project in our full model that includes all crowdfunding types. Interestingly, within reward-based crowdfunding (our largest subset), funders of cultural and for-profit projects attach less importance to information about finance and risks than those funding social and ecological projects. This could indicate that these projects display higher informational asymmetries related to their social and ecological goals versus cultural and for profit projects.

Financial-return
crowdfunding exhibits no significantly different information use based on the type of project, except for a decreased information use about the owner and her track record.

Control variables
As for our control variables, age and security of a promised return (which we interpret as instrumental motivation) show a statistically significant positive relationship to nearly all information variables in our full model.

Age is only insignificant for information needs about finance and risk.
When we split up the data into different types of crowdfunding, age loses most of its significance. The positive significant relationship between instrumental motivation and information needs remains consistent in all types of crowdfunding, except for information about finance and risks in financial-return crowdfunding. This is probably due to lack of variation within this category (financial return funders are likely to be instrumentally motivated). We find a strong positive moderating relationship for donation-and financial return crowdfunding regarding information about the entrepreneur and track-record as well as information about financials and risk in financial-return crowdfunding. As expected, the size of the investment (amount) drives the importance of information about financial planning and risks. Risk awareness is not significantly correlated with the importance of information in general.
Knowing the financial contribution of others increases the importance of information about financial planning and risks, indicating some additionality between knowing the contribution of others and information gathering for particularly out-crowd, instrumentally motivated fundersthe contribution of others increases the chance that the project will be fully funded, and therefore increases the expected payoff of time taken to gather financial and risk information.

Robustness checks
In order to check the robustness of our findings, we checked for multicollinearity i.e. the correlation among explanatory variables. Investigating the variance inflation factors (VIFs) reveals no multicollinearity, given the mean VIF of 1.5 in models including all types of crowdfunding and 1.6, 1.5, 1.7 in models using donation-based, reward based and financial-return crowdfunding respectively (see Kutner et al. 2005). We also divided relationship dependent variables into in-crowd and out-crowd and . We find that this hypothesis holds for financial crowdfunding. In our full model, we find no support for this notion and also find that funders in general -with or without ties -attach a lower importance to this type of information.
Crowdfunding decision making can thus be characterised as relationshipdriven (Bernstein et al. 2015b;Colombo et al. 2015). In this regard crowdfunders, when aggregated across all types, apparently behave differently to professional (VC) investors who rely also on financial due diligence and an alignment of goals between venture and investor (Audretsch et al. 2012;Bernstein et al. 2015b;Busenitz et al. 2005). This study also reveals interesting differences regarding the use of information

Information heterogeneity across types of crowdfunding
We find more support for our hypotheses when we separate distinct types of crowdfunding. The mechanisms through which we expect social networks to affect informational needs (motivation, intention of the project owner and quality of the project) seem to lead to different information needs for donation, reward, and financial (debt and equity) Interestingly, donation-based funders show significantly higher levels of information need about the person behind the project than reward-based funders, at similar levels as financial return crowdfunding. This is counter to expectations of Belleflamme et al. (2014) and Ahlers (2015), who argue that in donation-based crowdfunding the degree of asymmetric information is of little importance because other intangible factors increase the funders' utility. We explain this from a motivation perspective.
Donation crowdfunding can be likened to philanthropy, where 'returns' can be in the form of 'warm glow' (Andreoni 1990), societal impact (Maas and Liket 2010) or community benefits (Belleflamme et al. 2014). Donation funders interested in the (social, cultural or ecological) impact of their donation are more likely to be motivated to look for quality signals, indicating that their money will be well spent, before pledging their funds.
When we look at the effect of strength of ties on information needs in donation-based crowdfunding, we find no increased demand for information on either the project creator or financial planning and risks.
However, for out-crowd donation-based crowdfunding we find a significantly higher information need about the project and its objectives than for in-crowd funders. This is driven by a negative effect of weak ties in particular. This lower interest of weak tie funders in information about the project may point to a (weak) relationship motivation to donate instead of interest in the project and its impact. This is in contrast to, on the one hand, strong tie funders who may display interest in the project due to their strong relationship and, on the other hand, due to out-crowd funders who donate primarily out of interest in the project, without a social relationship.
We also find that, compared to reward-based and donation crowdfunding, financial return funders, with and without ties, are significantly more interested in information about financial planning and risks. The risk profile of reward-based crowdfunding is lower than debt or equity crowdfunding since they can be seen as early adopting consumers (Hornuf and Schwienbacher 2015;Vismara 2016) and their return does not depend on the long-term profitability of the enterprise, only on the ability to deliver the promised product. Non-delivery rates on the largest rewardbased platform Kickstarter are approximately 9% (Mollick 2015), which points to a much lower risk than average venture failure rates (Aldrich and Ruef 2006). Within the subset of financial return crowdfunding, we find that out-crowd funders have a higher need for information about finance and risk than in-crowd funders. Our results indicate that a strong relationship appears to substitute financial due diligence and complements the importance of teams quality signals as financial return funders with strong ties are less interested in information about finance and risk (Ahlers et al. 2015;Bernstein et al. 2015b;Uzzi 1999).

Conclusions and implications Conclusions
Our study offers the first detailed analysis of the heterogeneity in information use by crowdfunders, and more particular how information use is affected by social networks within different types of crowdfunding.
This paper highlights the heterogeneity in information use by crowdfunders that are differently connected to the project creator.
Funders from the in-crowd attach more importance to information about the project creator, as expected, but funders from the out-crowd do not rely more on information about the project, except for donation-based crowdfunding. Our findings suggest a trade-off between strong ties and the importance of information about financial planning and risks in the context of financial return (equity and debt) crowdfunding. In general, this information is perceived as less important and is not influenced by social network ties between crowdfunder and project for donation and reward crowdfunding. Donation and financial return crowdfunders attach more importance to the information about the person behind the project which reflects a relationship-based funding approach, whereas reward-based crowdfunders care significantly less about the project creator as they focus on the product as specific output with lower information asymmetry issues. Additionally, the information use of crowdfunders is influenced by the type of project they invest in. For-profit project funders need less information about a project and its objectives whereas ecological projects exhibit a higher need for this type of information.

Implications
Our research has important implications for project developers and platform managers. Based on the results of our research, platform managers and project owners can customize their campaign directly to the group of funders they would like to attract, based on their relationship strength and also on insights from our control variables (age, gender, education, instrumental motivation or financial means). More importantly, it is possible to deploy a tailored and staged in-crowd/out-crowd process of crowdfunding (see Figure 2). We indicate 'average' information use when coefficients are small or not significant. and financial planning/risks (for financial return crowdfunding), and summarise personal information about the project creator.