An Assessment of the Operational Effect of Non-Financial Information Exchange on Equity Crowdfunding Project Success

Abstract

Equity crowdfunding has supported numerous entrepreneurial projects and businesses lacking internal funding to commercialize their innovative products and grow them into businesses. Despite the above benefit, existing research has established challenges confronting the crowdfunding models. From the equity crowdfunding market perspective, information asymmetry leads to high product quality uncertainty and weakening potential investors’ willingness to invest. To solve the above challenges, many platforms have improved their operational designs and policies to enhance information exchange and accessibility by investors. Despite the improvements made by various platforms, limited knowledge exists on the effect of investor and sponsor participation, subsequent project updates, and the informative content of the updates on projects’ outcomes. In contributing to the above research gap, this paper uses data from Seedr’s equity crowdfunding platform to test hypotheses. The empirical research section used two sets of models: the first set involved cross-sectional data instrumental variables, and the second set utilized panel data to explore the longitudinal nature of the sample, providing additional support for the results of the cross-sectional data models. In the analysis of the topics of project updates, natural language processing techniques from machine learning were employed to extract content from unstructured data. The LDA (Latent Dirichlet Allocation) model was used as a topic modeling approach to identify the themes in the updated contents. The quantified results of the extracted topics were then incorporated into the empirical models. The results show that higher comments and update frequency will attract more subsequent potential investors to invest. If the project sponsor makes project modifications according to the wishes of the comments, the success rate of the projects will increase. Different update themes also have different impacts on project success. The findings are significant for entrepreneurs and investors involved in equity crowdfunding, providing valuable insights and empowering them to make informed decisions.

Share and Cite:

Liu, X. , Feng, Y. and Tuo, G. (2025) An Assessment of the Operational Effect of Non-Financial Information Exchange on Equity Crowdfunding Project Success. Journal of Data Analysis and Information Processing, 13, 156-181. doi: 10.4236/jdaip.2025.132010.

1. Introduction

The phenomenal growth rate of the phenomenal growth rate of crowdfunding has attracted much attention from the academic world. Experts and academics seem to agree that crowdfunding will become a popular way of financing in the future. [1] states global crowdfunding is expected to reach $96 billion by 2025.

The four types of crowdfunding are donation-based, loan-based model, rewards-based, and equity-based [2]. Much research still focuses on rewards-based crowdfunding and its influencing factors [2]-[4]. Equity crowdfunding was first proposed in the early 2000s in the United States. Also, the implementation and development of equity crowdfunding have been arduous, which has led to few academic discussions and studies on it.

Equity crowdfunding investors are usually motivated by the expected return on their investment. These investors face more information asymmetry [5]. The equity stake of investors makes it different from other crowdfunding models. Hence, research findings on the success factors of other models do not necessarily provide a helpful perspective on equity crowdfunding.

The current equity crowdfunding market and its operational process on the platforms have contributed to much information asymmetry [5] [6] and high uncertainty regarding project quality [5]. This uncertainty adversely affects potential investors’ willingness to pledge [7]. Platforms are improving information disclosure through website policies and designs. Continuous project updates by sponsors on the platform and investors leaving comments for responses later have improved information exchanges.

The above information exchange during funding is believed to improve transparency, increase disclosure, project initiator’s perceived effort, and investors’ trust [8]. However, the effect of the exchange of knowledge in the investor commenting process on funding success has received little attention in the existing crowdfunding literature [9]-[11], which has explored the role of project updates in filling the above research gap. [10] used linear regression to demonstrate the correlation between project updates and investor contributions. In a predictive analysis, [11] established a link between updated information and project fundraising success. Nevertheless, both studies treated project updates as exogenous variables, possibly leading to biased estimates. Hence, the following research question remains unanswered: Do investor and sponsor participation, subsequent project updates, and the informative content of the updates affect projects’ outcomes?

To answer the above research questions and contribute to knowledge, we investigate investors’ broader role in equity crowdfunding from the perspective of non-financial exchange mechanisms. We further construct a principal-agent relationship between the investors and the project sponsors, showing they can benefit from investors’ participation. Then, an initial set of cross-sectional instrumental variables approach is used to correct for the frequency and content of exchanged information endogenously. Also, we use panel data to mine the longitudinal nature of the sample, thus providing additional support to the results of the cross-sectional data models and enhancing the robustness of the final results.

The rest of the paper is organized as follows: Section 2 presents the literature review and the formulation of the hypotheses. Sections 3 and 4 describe the data source and models, respectively. Section 5 presents the result analysis, and Section 6 discusses the results, concludes the paper, and proposes future research.

2. Literature Review and Hypotheses Development

The text is based on the growing literature on crowdfunding in operations management. Early empirical work addressed project and sponsor characteristics associated with crowdfunding success [3] [12]. More recently, researchers in the operations management field have conducted theoretical analyses to address the best strategies to increase project success and revenue [13] [14] and the design of crowdfunding platforms’ mechanisms [15] [16]. However, few empirical papers have examined the impact of the communication features designed by equity crowdfunding platforms on investors’ pledging behaviors.

In the operations management literature on crowdfunding, most empirical studies focus on project success through cross-sectional data and ignore crowdfunding’s dynamic characteristics. To our knowledge, only two papers focus on crowdfunding dynamics. [16] studied the dynamics of backer behavior during fundraising and revealed that backers contribute more in the first and last weeks, whereas [17] found that investors with a public profile attract early and late investors to the project.

Unlike other research, this paper focuses on communication tools designed by equity crowdfunding platforms and used by project sponsors and investors. Thus, it investigates the impact of project updates and their informative content on project success.

3. Literature Review and Hypotheses Development

The text is based on the growing literature on crowdfunding in operations management. Early empirical work addressed project and sponsor characteristics associated with crowdfunding success [12] [14]. More recently, researchers in the operations management field have conducted theoretical analyses to address the best strategies to increase project success and revenue [13] [14] and the design of crowdfunding platforms’ mechanisms [15] [16]. However, few empirical papers have examined the impact of the communication features designed by equity crowdfunding platforms on investors’ pledging behaviors.

In the operations management literature on crowdfunding, most empirical studies focus on project success through cross-sectional data and ignore crowdfunding’s dynamic characteristics. To our knowledge, only two papers focus on crowdfunding dynamics. [16] studied the dynamics of backer behavior during fundraising and revealed that backers contribute more in the first and last weeks. [17] found that investors with a public profile attract early and late investors to the project.

Unlike other research, this paper focuses on communication tools designed by equity crowdfunding platforms and used by project sponsors and investors. Thus, it investigates the impact of project updates and their informative content on project success.

3.1. Hypotheses Development

This Section develops hypotheses regarding investor and sponsor participation, subsequent project updates, and the informative content of the updates on project success based on a closely related literature review.

3.1.1. Investor Participation and Project Success

Agency theory is one of the most essential theories in corporate operation, based on the game theory of asymmetric information. Its core task is to study how the principal can design the optimal contract to incentivize the agent under conflicting interests and asymmetric information [18] [19]. It further holds that due to the conflicting interests between the agent and principal, it is likely that the agent may pose an opportunistic behavior.

Similarly, in equity crowdfunding, investors provide funds to the project initiator with the expectation that the initiator will use the funds to develop a project that maximizes the investor’s utility. Subsequently, a principal-agent relationship is created between the two parties [20]. Also, information asymmetry between project initiators and investors would result in agency risks and opportunism. Their study [2] highlights that equity crowdfunding platforms amplify information asymmetry compared to other financing options for early-stage ventures where investors know much about the venture prospects. Investors on crowdfunding platforms, in turn, face two agency problems [19]. First, information asymmetry exists where investors rely on information provided by the sponsor to assess the project quality and the ability to achieve the expected performance. Second, the project description is not fixed until the crowdfunding project is closed. Hence, difficulty exists for an investor to monitor the project progress to ensure the project sponsor does not deviate from the investment interest.

The agency problem allows investors to participate in the project through information exchange to safeguard their investment interests [21] and influence the project description to suit their needs [22]. The agency issue is often raised in investment areas [21]. Unlike traditional investors, who mitigate agency risk through close monitoring of startups, equity crowdfunding investors battle risk due to a lack of information to assess project viability. The online information exchange during the project funding stage seeks to mitigate agency risk for equity crowdfunding investors.

A growing body of literature on crowdfunding focuses on the exchange of resources between backers and project sponsors [23], examining the contribution patterns of backers [3] [24] and the plausibility of group investment [25]. Still, exchanging non-financial resources to promote knowledge during investment has received little attention in the existing equity crowdfunding literature [9] [23]. This study constructs a principal-agent relationship between the investor and the project sponsor. The crowd investor is the principal, while the project sponsor becomes the agent. The principal entrusts their financial resource to the agent to work with for expected return. It shows that project sponsors can benefit from investors’ participation as their involvement in crowdfunding projects extends beyond the provision of financial resources. The non-financial resources come in the form of information to improve the project’s quality. The above position is based on customers’ involvement in product development and evaluation from their point of view, which increases the quality [26]. However, in equity crowdfunding, the role of investors in project development is unknown. This research argues that equity crowdfunding investors take on some value-added activities just like institutional investors do. The value-added role of equity crowdfunding investors is to help subsequent institutional investors alleviate the agency problem by providing information about the project’s quality [27].

Further, potential equity investors will refer to the information provided by pioneer investors and the level of participation to address the uncertainty of the project quality and bring a balanced investor voice into the project development [28]. The above process would constitute knowledge sourcing from the perspective of the project sponsor [29], where community members fund and influence the project’s development. This paper argues that project sponsors can improve disclosure and increase investor communication by posting updates, responses, and investor comments [30]. If the cumulative number of communication messages increases over time, potential investors receive more information about the project, reducing their uncertainty about its quality. The high frequency of these messages leads to greater transparency about the project and positively impacts potential investors’ willingness to invest.

As mentioned above, investor investment in equity crowdfunding projects creates a principal-agent relationship between the investor and the project sponsor. As the subject of such a relationship differs from existing literature, the investor relies on the project sponsor for non-financial resources to develop the project to maximize its effectiveness [31]. As a result, investors are incentivized to monitor the project’s progress, influence its development, and drive it in an optimal direction [2]. It further argues that projects with high investor engagement will receive more financing from equity crowdfunding than projects with low customer engagement. Based on the above position of the paper, the following hypothesis is put forth.

Hypothesis 1 (H1): Equity crowdfunding projects with higher investor participation attract more subsequent investors, increasing the funding success rate.

3.1.2. Project Modification and Project Success

Part of the argument for Hypothesis 1 is that investor engagement can increase financing because it allows investors to influence the project description, and investors will consider the potential future market for the project because it will affect their future financial returns. Research on product development suggests that consumers are experts in product application and value creation [32], improving product development performance and better service [33]. However, the subjective role of consumers may also hinder product and service development since potential application information from consumers to producers is sticky and costly. The above situation is evident in equity crowdfunding, where investors share sticky information because their investment in the project motivates them to influence its development (Hypothesis 1). By sharing project features, investors can reduce agency risk by directing the project development toward future customer needs.

When product developers are faced with many consumers acting only in their interests, it is difficult for them to know which ideas are likely to increase the success of their product and or result in high-cost niche products [34]. Following an idea that doesn’t make sense or is trying to address all consumer suggestions can result in a product that is not viable, too costly, or unattractive to consumers in a financial market. There is ample evidence that consumers’ unspecialized technical knowledge can lead products to adopt unpopular technologies. Similarly, under equity crowdfunding, the project sponsor faces a similar situation, in addition to the moral hazard of investors and the risk of group irrationality from those with solid influence [35]. However, due to future financial returns of interest, investors balance it against the project cost for market acceptance and economic gains, reducing the sponsor’s moral hazard and eventually achieving a better alignment of interest.

There are no sources in the current document.

In referring to the role of consumers in product development against the more rational investors in equity crowdfunding, this paper argues that the value of a project due to investor participation is realized when project sponsors modify their project description based on the information they receive from investors. Thus, investor engagement’s effect on crowdfunding success depends on whether or not the project sponsor modifies the project following the comments. Based on the paper’s position above, the second hypothesis follows.

Hypothesis 2 (H2): In equity crowdfunding, investor engagement’s effect on project success depends on whether the project sponsor modifies the project based on comments.

3.1.3. Frequency of Update and Project Success

The project initiator on the equity crowdfunding platform can disclose additional project-related information by releasing some communicative information during the project fundraising process. The project update section of the webpage serves as a communication tool on the crowdfunding platform, where the sponsor shares the latest project updates with expected feedback from investors. From the point of view of information disclosure, more updates can convey more current and project-related information to potential investors. The increase in information sharing perceived by potential investors will reduce information asymmetry, thus increasing investors’ willingness to invest [36].

Regarding project sponsors’ motivation, potential investors may perceive more significant effort and trust when a sponsor manages communication channels more effectively [10], subsequently attracting more investors. More updates may generate more project attention, transparency, and excitement, ultimately leading to better financing performance [31]. At the same time, more frequent updates can create a stronger connection between investors and the project, leading to more investment in the project [37]. [10] [11] also used project updates as an exogenous variable in their previous studies to discuss the link and correlation between it and fundraising success. However, using exogenous variables leads to estimation bias, which prevents us from determining whether the frequency of update releases positively affects the investors’ decision to pledge. In contrast, this paper uses instrumental variables to correct for endogeneity and determine the positive impact of project updates. The next hypothesis is formulated as follows.

Hypothesis 3 (H3): In equity crowdfunding, a higher frequency of project updates will attract more subsequent investors, which increases the success rate of crowdfunding projects.

3.1.4. Informative Update and Project Success

Research on information content analysis suggests that different types of content can impact users’ behavior differently due to the other values contained in their information. For example, [37] analyzed users’ posts on the Facebook business community and found that posts in different content categories affect consumer engagement differently. Marketing research has also demonstrated that communicating information affects consumers’ consumption decisions [38]. In the literature related to crowdfunding, some studies have shown that the content of communication between project promoters and investors affects crowdfunding outcomes [39]. However, there is a lack of qualitative measurements of the content of updates in equity crowdfunding projects to know how specific update topics affect investors’ investment decisions. Specifically, this paper argues that including updated information topics related to the project and those closely related to the project sponsor are more likely to mitigate information asymmetry. Also, increased information transparency that may enhance disclosure will encourage potential investors to decide to invest.

Alternatively, non-informative updates released solely to attract attention will likely be perceived as “distress signals” or objectionable to investors. This negative impact will discourage potential investors from investing, so updates that lack useful information will either be ineffective or discourage support from potential investors [40]. The effect of updates on investor behavior is content-driven. From the preceding discussion, the last hypothesis is formulated as follows.

Hypothesis 4 (H4): In equity crowdfunding, posting informative updates on different topics attracts more subsequent investors, increasing crowdfunding projects’ success rate.

Figure 1 below displays the four hypotheses and predicts investor participation, project modification, update frequency, and information content will increase project success.

Figure 1. Hypothesis structure.

3. Variable Setting and Data

This study tests the hypotheses using data from Seedrs.com, a UK equity crowdfunding website. This paper uses a panel data modeling framework to analyze how dynamically published communication information during the project financing process moderates investor pledging behavior and affects the success of equity crowdfunding projects. This paper uses snowball sampling to collect 2534 projects funded on the Seedrs platform between May 12, 2022, and May 12, 2023. For each project, this paper observes its project attributes and all funding transaction details and tracks comments and updates during the fundraising period of each project. The data includes successful and failed projects.

3.1. Data

The sample contains daily observation data for 2,534 Seedrs.com projects that started between May 12, 2022, and May 12, 2023. Projects with the following features were excluded from the sample. First, 16 projects that lasted less than a month after the start of the data collection or started less than a month before the end of the data collection were excluded because these projects have too short a duration, which may introduce errors in the result. Second, 8 projects were suspended due to lawsuits, and 11 failed projects that could not observe the complete funding phase were removed. Third, 19 projects that lasted less than a week and made capturing meaningful operational effects from investor comments or project sponsor updates challenging were excluded. Also, 48 projects with more than 2 days of missing observations were removed. The final sample contained 2432 projects with an average of six equity crowdfunding projects per day; the average duration of equity crowdfunding was 32.60 days, with a median of 30 days.

Equity crowdfunding projects on Seedrs fall into 13 categories, such as art, fashion, gaming, and technology, allowing this research to look at investor behavior in various contexts to remove industry or market bias.

There are three key advantages of using data from Seedrs.com. First, projects on Seedrs vary by product, entrepreneur, and industry, allowing data to be collected on both successful and failed projects. Second, the content of individual investor reviews can be collected on the Seedrs website, which permits this paper to use investor reviews as a metric. Third, variations in project descriptions can be captured by manipulating the webpage’s backend, which often causes investors to ignore this information.

This project proposes to use two sets of models to validate the impact of investor participation in reviews on crowdfunding success. The first is a set of cross-sectional data models, while the second is the panel data model. The choice of the dataset is based on the premise that crowdfunding success is a dichotomous variable. Hence, the logit model is preferred to the linear probability model. Also, using cross-sectional data would help tap into the longitudinal nature of the sample, thus providing additional support for the results.

Further, this paper uses the control function method to control for endogeneity, which is preferred in logit models. To rule out reverse causality, the fixed effects model allows for the temporal separation of investor comments and subsequent investments. Moreover, to enhance the robustness of the final results, the fixed effects would allow for the removal of all constant influences from the estimation and the explicit control of social learning (e.g., herding effects). The remainder of this Section introduces the variables in more detail.

3.2. Dependent Variable

This paper uses the same explanatory variables in the instrumental variable (IV) and the panel models. The panel model uses all time-series observations. In contrast, the IV model uses the cumulative amount of equity crowdfunding projects during their success or failure.

Investor comments. To measure investor comments, this research counts the comments left by investors in the “Discussions” section of a project webpage, where investors can only comment on the projects they have invested in. For project i , C i,t is the cumulative number of investors’ comments up to the period t. In the panel model, we use this variable directly, while the cross-section IV model uses the cumulative number of investors’ comments up to the time when the project reaches its funding goal, defined as C i, t F , t F .

Project updates. To determine the number of updates posted by a project sponsor, this paper counts the number of updates displayed in the “Updates” section of the project webpage, where a sponsor can post updates for any information about the project. For project U i, t F . is the cumulative number of updates posted by the initiator up to period t; we use this variable directly in the panel model. In the cross-sectional data IV model, this paper uses the cumulative number of investor comments until the project reaches its funding goal, which is defined as U i, t F .

Project modifications. To test whether a project sponsor’s response to investor comments moderates the impact of comments on the success of a project, it counts the number of changes made to the project description by the sponsor. It identifies such changes by looking at the timestamps of updates in the HTML code of the project’s webpage. Each time the timestamp of an update moves forward, it represents a change in the project description. Notably, the timestamps are not visible to investors. This paper uses R i, t F defined as the cumulative number of revisions up to the period t F in the IV model, and the difference in the number of revisions R i,t from the previous period in the panel model.

Update content topics. To study the impact of updated content presented in H4, we will extract content from unstructured data using machine learning techniques. We use the LDA model, a popular method for extracting the topics designed in each update. The LDA model treats updated text documents as a mixture of several latent topics, each modeled as a distribution of high-frequency keywords extracted from all the textual documents [41] [42]. As a probabilistic generative model, LDA has received increasing attention from economics and business researchers in dealing with non-institutionalized text data [43] [44].

Following the standard steps of the LDA method, topic extraction is performed on the preprocessed comment text. Based on a set of keywords associated with each topic, we can derive the main content of all the topics. For the update, we obtained three extracted topics: the project success announcement, the details related to the project cooperation, and the project launch announcement. The keywords associated with each theme are provided in Table 1.

Table 1. Keywords corresponding to the topics.

Topic

Keywords

Success announcement

remarkable, outstanding, achieve, excellence, testament, commitment, team, completion, successful, property

Details related to cooperation

Game, energy, industry, player, share, delight, collaboration, financial, please

Launch announcement

leisure, sport, industry, embark, ready, journey, solution, experience, travel, exceptional

The LDA model in this paper generates probability distributions of each extracted topic in a given document, which are crucial pieces of information. Some quantifiable metrics are needed to validate the extracted content information, which can be used in subsequent analysis. Therefore, for the updated content, this paper constructs cumulative update topics ( CummulativeUpdateTopi c j , where j=1,2,3 ) by aggregating the specific topic probabilities of all updates related to the same project.

4. Instrumental Variable and Panel Models

In the first set of models, this paper uses an instrumental variable to control for the promoter’s personal competence and the project’s quality. It further tests the effect of investor comments on a project’s likelihood of success. The IV model is described in this Section.

4.1. Instrumental Variable Model

4.1.1. Selection of Instrumental Variable

This paper uses the mobile application published by Seedrs as a tool. The application facilitates investment and improves the interaction between the sponsor and the investors. The mobile app works independently; therefore, the change in the likelihood of success of a project caused by increased investor comments via the mobile app is exogenous. What is certain is that this kind of change in likelihood is not a derivative of unobservable project promoter competence or project quality. This paper uses a two-stage estimation to achieve this process [45]. The first stage uses the estimated exogenous growth of investor comments to estimate its impact on the success of a project. In the second stage, the results are independent of the project promoter’s ability and quality. The following paragraphs describe and evaluate the instrumental variable in detail.

This paper evaluates the quality of this instrumental variable in an uncontrolled model, and the results show that the introduction of the mobile app had a significant positive impact on investor comments ( β 1 =0.55 , p < 0.01). The first stage of F value is 24.86, much higher than the common threshold value of 10, indicating that the mobile app should be an effective instrumental variable. According to the control function approach, the second stage incorporates the residual values estimated in the first stage ( u ^ i ) , to control for unobservable confounders. The paper finds significant residual values in the second stage, which suggests that there is indeed an endogeneity problem and, therefore, justifies the use of this instrumental variable ( γ u =0.03 , p < 0.01) [45]. Despite using the mobile app on Seedrs, there is no evidence of violating the exclusivity restrictions that govern the selection of instrumental variables.

4.1.2. Description of Control Variables

This paper includes several control variables in the model to account for potential confounders, particularly project sponsor competence, project quality, and complexity.

The paper controls the following variables. First, video production is costly and influences an investor’s pledge decision [46]. Therefore, the number of videos in a project description allows us to control the project sponsor’s ability and the project’s attractiveness [47]. Second, we control whether a project has a separate website. Creating and maintaining a separate website requires resources and signifies a high-quality project. Third, the paper controls the length of the project description based on the number of words and the number of images uploaded.

The description communicates the project idea and provides “first-hand information to the public” [48]. These variables are controlled because investors rely on the description and the images to assess the project quality [49]. Uploading product prototypes augments the textual information [50], proves initiator credibility, and enhances investment decisions. Also, whether or not a project sponsor launches a pitch is a variable that must be controlled. A pitch helps potential investors understand the project and evaluate it comprehensively. A project initiated by a team is also a control variable. A team is more resourced due to the diverse resources needed to begin a quality project than an individual [51]. Reward is also one of the control variables since it can influence investment decisions and as a symbol of project quality [52]. Therefore, we expect that the reward setting will positively impact the attraction of investors.

Moreover, the natural logarithm of the project crowdfunding target amount (in pounds) is controlled. The project goal is the sponsor’s actual expectation of the minimum feasible project budget, which allows us to control the size and complexity of the project [53]. Additionally, the length of time is a common measure of project complexity [53]; hence, it is also controlled.

We control whether projects have partnerships with other organizations and are backed by venture capital, which is a good control indicator of project resources, development experience, and financing networks of sponsors. The existence of venture capital is bound to be followed by lead and follow-on investments. Each investment round would attract many venture capitalists and potential investors to invest in projects. Furthermore, we control for project categories. Project frequency, funding deal volume, and investor investment amount vary by category. Lastly, the paper controls whether a project involves e-commerce or not. In an era of e-commerce development, people tend to expect more from the e-commerce economy, and the difference between the e-commerce and brick-and-mortar economy is a factor that influences the decisions of potential investors.

4.1.3. Model Description

Equity crowdfunding project success is a dichotomous variable; therefore, the logit model is preferred to the linear probability model [54]. To control for endogeneity, we use the control function method in two-stage regressions [46], the preferred method for dealing with endogeneity problems in logit models. In the first stage, we regress the number of investor reviews on exogenous and instrumental variables ( Z i ) in an ordinary least squares model, followed by the residuals from the first-stage estimation ( u ^ i ) in the second stage, logit regression. The second-stage estimates are consistent estimators. In the case of logit models, the obtained estimates will be smaller than the actual effect size. Since the obtained standard errors may not be strictly valid, we use a nonparametric bootstrap in the second stage to obtain corrected standard errors [54]. We obtained H1 first and second-stage regressions (IV model 1):

C i, t F = β 0 + β 1 Z i + β X X i + u i

S i = γ 0 + γ 1 C i, t F + γ u u ^ i + γ X X i + v i ,

Positioning Figures and Tables: Figures/tables where C i, t F is the number of investor comments, Z i is an instrumental variable, and X i is a vector of a set of control variables. Next, we describe how to test whether project sponsors making project changes based on the content of comments moderates the effect on project success (H2) in a cross-sectional IV model. In the initial model (IV Model 1), where project success is the dependent variable, and the number of investor comments is the independent variable, we are interested in R i, t F =δ C i, t F and S i = γ R R i, t F . We are still using the cell phone application as an instrumental variable, but this time, it serves as the mediating variable R i, t F . Because project sponsors cannot use the mobile app to modify their project descriptions, there is no direct causal relationship between the introduction of the mobile app and project content modification. Therefore, project content modification is irrelevant to the instrumental variable and will only be affected by increased investor comments . Any moderating effect of project content modification on success is denoted as γ R , which is a consequence of . Thus, we can obtain a model for H2 in the first and second stages (IV Model 2):

R i, t F = β 0 + β 1 Z i + β X X i + u i

S i = γ 0 + γ 1 R i, t F + γ U u ^ i + γ X X i + v i

For testing H3, we use the same idea as for testing H1, and we can get the model of H3 in the first and second stages (IV model 3):

U i, t F = β 0 + β 1 Z i + β X X i + u i

S i = γ 0 + γ 1 U i, t F + γ U u ^ i + γ X X i + v i

To investigate the content effect of updated information, this paper directly observes the significant relationship between the final success of a project and a specific topic by adding the variable CummulativeUpdateTopi c j , (where j=1,2,3 ) to the model (IV model 4):

S i = β 0 + β 1:3 ( CummulativeUpdateTopic 1 i :CummulativeUpdateTopic 3 i )+ β X X i

The parameter we are interested in testing H4 is β will be supported when estimating β changes concerning sign and significance level.

4.2. Panel Model

A panel model will be used to explore the longitudinal nature of the sample to provide additional support for the initial results of the IV model. It would further help to dynamically analyze the impact of investor comments and sponsors’ updates on project success. The panel model enhances the robustness of the results in three ways. First, the temporal separation of investor comments and subsequent investments allows us to rule out reverse causality. Second, the project fixed effects eliminate all constant influences present in the estimation, allowing us to control for project quality and project sponsor competence differently than instrumental variables. Third, it will enable us to control social learning effectively. The panel data contains unbalanced data at 24-hour observation intervals from when the project is posted on Seedrs to when it ends. Since some variables require two lags, the first two observations for each project are removed, and the dependent variable in all panel models is the amount of new financing received since the previous period ( F i,t ).

Model Description

The fixed effects models for H1 and H3 are as follows:

F i,t = β 1 C i,t1 + β 2 U i,t1 + β X X i,t1 + β T1 TIM E i,t + β T2 TIM E i,t 2 + a i + u i,t

where C i,t1 is the one-period lag of the cumulative amount of investor comments and U i,t1 is the one-period lag of the cumulative amount of project sponsor updates. X i,t1 is the control variable for all one-period lags, and the fixed effect is a i . We can test H1 and H3 by estimating β 1 and β 2 . To test H2, we apply a two-step approach to analyze the interaction between investor comments and project modifications. First, we decompose C i,t1 into its two-period lag and its difference with the two-period lagged variable:

C i,t1 = C i,t1 + C i,t2

We then interact with program content modifications ( R i,t1 ) with a two-period lag of the number of investor comments, C i,t2 , which effectively replaces C i,t1 as the primary explanatory variable:

F i,t = β 1 C i,t1 + β 2 C i,t2 + β 3 R i,t1 + β 4 C i,t2 × R i,t1 + β X X i,t1 + β T1 TIM E i,t + β T2 TIM E i,t 2 + a i + u i,t

The logic is that investor comments are needed to affect financing through project content changes, so we test the effect of lagged two-period volumes on financing now by lagging one-period project content changes ( β 4 ); cumulative investor comments (t − 2) → the number of project content changes (t − 1) → the volume of financing (t), and also due to the separation in time, the number of project content changes in period t − 1 does not affect the number of investor comments in period t − 2. To investigate the content effect of update topics, this paper again expands Model 1 by including the variable CummulativeUpdateTopi c j (where j=1,2,3 ):

F i,t = β 1:3 ( CummulativeUpdateTopic 1 i,t :CummulativeUpdateTopic 3 i,t ) + β X X i,t1 + β T1 TIM E i,t + β T2 TIM E i,t 2 + a i + u i,t

Similarly, the parameter we are concerned with for testing H4 is β 1:3 , which will be supported when the value of β 1:3 is significant.

5. Result Analysis

5.1. Instrumental Variable Model Result Analysis

Table 2 demonstrates the statistical description of the IV model, and Table 3 shows its correlation matrix, which, according to the report, shows no multicollinearity problems.

Table 2. Descriptive statistics of the IV model.

Variables

Average

Standard error

Minimum

Maximum

Project success

0.875

0.331

0.000

1.000

Mobile app

0.395

0.489

0.000

1.000

Investor participation

35.559

36.355

0.000

220.000

Sponsor updates

124.439

111.806

0.000

1291.000

Project modifications

124.801

107.970

0.000

883.000

Target amount (ln)

44.132

26.875

11.290

265.429

Project duration

32.60

91.783

22.000

1092.000

Text length

2046.748

1296.667

341.000

13,028.000

Number of pictures

14.014

13.125

0.000

147.000

Rewards

0.441

0.497

0.000

1.000

Team

0.658

0.475

0.000

1.000

Venture capital

0.329

0.470

0.000

1.000

Partnership

0.362

0.481

0.000

1.000

Pitch

0.441

0.497

0.000

1.000

e-commerce

0.803

0.398

0.000

1.000

Video

5.298

6.167

0.000

67.000

Separate website

0.586

0.493

0.000

1.000

Follow-on

0.599

0.490

0.000

1.000

IV Model 1 in Table 4 shows the results related to H1, where we find that the number of investor comments (i.e., participation) has a significant positive effect on project success (H1) after controlling unobservable confounders in the second stage ( γ 1 =0.260 , p < 0.01). Each additional comment before the project’s success increases the likelihood of its eventual success by 26%.

To test whether project sponsors made project changes based on the content of the (H2), we regressed project content modification on the instrumental variable. First, we assessed the quality of the instrumental variable without control variables and found a significant positive effect on modification ( β 1 =0.64 , p < 0.01) and

Table 3. Correlation matrix of the IV model.

Table 4. H1 and H2-Stage 1 and Stage 2 regressions of IV model.

Variables

IV Model 1

IV Model 2

Stage 1

(investor comments)

Stage 2

(project success)

Stage 1

(project modification)

Stage 2

(project success)

Mobile app

33.418*** (28.509)

0.64*** (0.04)

Investor comments

0.260*** (10.050)

Project modification

0.330*** (4.076)

Target amount (ln)

−0.040* (−1.834)

−0.014** (−2.256)

0.267** (2.272)

−0.051*** (−2.370)

Project duration

−0.023*** (−4.827)

−0.000 (−0.189)

0.353*** (10.996)

−0.086*** (−3.025)

Text length

−0.000*** (−0.774)

−0.000 (−0.782)

0.015*** (5.373)

0.003** (−2.241)

Number of pictures

0.004 (0.103)

0.019** (2.246)

0.917*** (4.943)

−0.226*** (−3.010)

Rewards

0.098 (0.101)

0.358 (1.268)

15.056*** (3.709)

−3.997*** (−3.443)

Team

5.923*** (6.983)

0.828*** (3.144)

4.098 (0.992)

−0.191 (−0.528)

Venture capital

1.451 (1.444)

−0.233 (−0.930)

3.341 (0.778)

−0.891** (−2.318)

Partnership

−2.856*** (−3.534)

−0.194 (−0.708)

17.011*** (4.059)

−0.954 (−0.639)

Pitch

4.069*** (5.218)

0.991*** (3.698)

14.145*** (3.693)

−1.967* (−1.905)

e-commerce

−11.864*** (−7.372)

−2.486*** (−3.057)

5.379 (1.009)

0.273 (0.371)

Video

2.161*** (10.373)

0.880*** (7.224)

1.830*** (3.963)

−0.181 (−1.151)

Separate website

5.791*** (6.528)

−1.006*** (−2.865)

27.261*** (4.572)

−3.572* (−1.835)

Follow-on

0.030 (0.038)

−0.999*** (−2.920)

−8.283 (−1.507)

1.659** (2.208)

Residual

0.161*** (5.216)

−0.185** (−2.348)

Intercept

17.193*** (9.239)

0.572 (0.604)

−19.692** (−2.312)

0.211 (0.113)

F-statistic

R2

276.61

0.600

6026.15

0.615

38.94

3042.30

0.910

0.851

Statistic

2432

2432

2432

2432

p < 0.10; **p < 0.05; ***p < 0.01.

an F-statistic of 38.94 in the first stage, which is well above the common threshold of 10, suggesting that it is a valid instrumental variable. We found significant residuals in the second stage γ U =0.185 , p < 0.05), which suggests that there is indeed an endogeneity problem, thus justifying the use of an instrumental variable. In the IV model 2, with the inclusion of control variables, the effect of project modifications on project success is positive and significant ( γ 1 =0.330 , p < 0.01). This means that when unobservable confounders are controlled for validating H2, project initiators who make project changes based on the comments’ content increase the likelihood of project success by 33%.

Table 5 reports the analysis results of the impact of specific informative update themes on project success. For the project sponsor’s informative, updated theme variables, the estimates for CumulativeUpdateTopic1 (i.e., project success announcement), CumulativeUpdateTopic2 (i.e., detailed information related to the project’s collaboration), and CumulativeUpdateTopic3 (i.e., project launch announcement) are positive and significant (p < 0.01).

Table 5. H4-Thematic effects of update.

Dependent variable = project success

Topic label

Estimated value

Standard error

Update topic 1

Project success announcement

235.538***

6.553

Update topic 2

Project cooperation information

175.205***

5.811

Update topic 3

Project launch announcement

100.343***

8.414

Target amount (ln)

0.074**

2.290

Project duration

−0.021*

−1.910

Text length

0.000

0.883

Number of pictures

−0.016

−0.400

Rewards

−29.263***

−6.910

Team

−12.253***

−6.599

Venture capital

−0.392

−0.286

Partnership

25.918***

6.746

Pitch

−0.272

−0.347

e-commerce

−4.879***

−3.040

Video

−0.162

−1.278

Team

−12.253***

−6.599

Venture capital

−0.392

−0.286

Partnership

25.918***

6.746

Pitch

−0.272

−0.347

e-commerce

−4.879***

−3.040

Video

−0.162

−1.278

Separate website

20.749***

7.565

Follow-on

9.122***

2.605

Residual

−51.560***

−6.611

Intercept

38.26

R2

0.929

Statistic

2432

p < 0.10; **p < 0.05; ***p < 0.01.

5.2. Panel Model Result Analysis

Table 6 presents the descriptive statistics of the panel model and reports that there are no multicollinearity problems. The dependent variable is the amount of new funds received since the previous period ( F i,t ), and Table 7 demonstrates the main estimation results. All standard errors are clustered at the project level and robust to heteroskedasticity and autocorrelation.

Table 6. Descriptive statistics for panel data.

Variables

Average

Standard error

Minimum

Maximum

Cumulative investor comments

14.534

24.957

0

220

Cumulative sponsor updates

0.286

0.902

0

20

New project modifications

0.91

3.31

0

41

Cumulative investors

145.196

294.456

0

2990

New investors

9.827

60.534

−25

2116

Cumulative financing

27,368.305

350,131.704

−100,000

16,564,190

New financing

146.55

1801.56

−83,913

923,169

Average financing

87,729.866

486,863.224

0

16,564,190

Cumulative Facebook sharing

78.56

504.54

0

522

To test H1, this paper regresses the amount of new funding received since the previous period on the number of investor comments in the lagged period. The value of the β1 coefficient in Model 1 is 0.115, which is significant (p < 0.01) supports H1. One investor comment immediately generates £0.115 of funding for a project, and this cumulative effect increases every 24 hours for a project duration of 32 days on average. One comment posted by an investor halfway through generates £1.84 of additional funding, while an investor comment posted at the start of the same project generates £3.68. For the most extended duration (60 - 61 days), the cumulative effect increases to £6.9 (0.115 * 60). These are averages, so some reviews may result in more or less investment. Also, to test H3, the amount of new funds received since the previous period is regressed on the amount of project promoter updates lagged by one period in Model 1. The value of the coefficient of β 2 in Model 1 is 0.434, which is positive and significant p < 0.01, and supports H3. Similarly, a single update from a promoter immediately generates £0.434, with a cumulative effect, and increases every 24 hours within an average project duration of 32 days. While £13.888 is generated at the start of a project, additional funding of £6.944 is generated halfway when a promoter posts one update. For the longest (60 - 61 days) project duration, the cumulative effect increases to £26.04 (0.434 * 60).

To test the mediating effect of the number of changes in program content (H2), the one-period lagged cumulative number of investor comments is split into the two-period lagged and compared with the previous results (Model 2A). The effect of the two-period lagged cumulative number of investor comments is slightly weaker than that of the one-period lagged cumulative number ( β 2 =0.045 , p < 0.01). We then incorporate the interaction between the number of program content revisions and the number of cumulative investor comments with a two-period lag in the faceted model 2B. Now, the effect of a two-period lagged cumulative number of investor comments is not significant, and the interaction between investor comments and project content revisions is highly significant and very similar to the estimated effect of a two-period lagged number of investor comments in Model 2A β 4 =0.043 , p < 0.01. Thus, the direct effect of the two-period lagged cumulative number of investor comments is fully absorbed by its interaction with the number of program content revisions supporting H2. Table 7 presents the main result.

Table 7. Results of the panel data model.

Variables

Model 1

Model 2A

Model 2B

Investor comments (t − 1)

0.115*** (8.169)

Sponsor updates (t − 1)

0.434*** (6.276)

New investor comments (t − 1)

0.5235*** (7.233)

0.277*** (3.977)

Investor comments (t − 2)

0.045*** (4.119)

0.010 (0.915)

New project modifications (t1)

0.128*** (8.539)

Investor comments (t − 2) × new project modifications (t − 1)

Cumulative investors (t − 1)

New investors (t − 1)

Average financing (t − 1)

0.0355*** (19.597)

−0.0005*** (−6.954)

0.000 (1.576)

0.026*** (15.366)

−0.000*** (−2.990)

0.0005*** (3.174)

0.043*** (4.257)

0.0165*** (9.843)

0.000 (0.528)

0.0005** (2.109)

Cumulative Facebook sharing

Project duration

Project duration2

Fixed effects

−2.53* (1.31)

−0.4595*** (−7.569)

0.006*** (4.180)

Yes

−2.53* (1.31)

−0.1855*** (−3.279)

0.01 (0.514)

Yes

−2.54* (1.32

0.005 (0.089)

−0.0025* (−1.735)

Yes

Intercept

0.6045*** (2.809)

0.3295* (1.647)

0.058 (0.315)

F-statistic

R2

172.68

0.57

163.55

0.57

130.81

0.58

Statistic

5624

5642

5642

p < 0.10; **p < 0.05; ***p < 0.01.

Table 8 reports the results of the panel data analysis for specific update content and presents the same results as the cross-sectional data. Cumulative updates on topics 1, 2, and 3 are positive and significant. Consistent with the hypotheses of this paper, Cumulative update topic 2 was found to be effective in encouraging subsequent investors to invest, as were the project launch and project success announcements.

Table 9 summarizes the hypotheses’ results for both the Panel model and the instrumental variable model.

Table 8. Panel data thematic effects of updates.

Dependent variable = project success

Topic label

Estimated value

Standard error

Update topic 1

Project success announcement

235.538***

6.553

Update topic 2

Project cooperation information

175.205***

5.811

Update topic 3

Project launch announcement

100.343***

8.414

Target amount (ln)

0.074**

2.290

Project duration

−0.021*

−1.910

Text length

0.000

0.883

Number of pictures

−0.016

−0.400

Rewards

−29.263***

−6.910

Team

−12.253***

−6.599

Venture capital

−0.392

−0.286

Partnership

25.918***

6.746

Pitch

−0.272

−0.347

e-commerce

−4.879***

−3.040

Video

−0.162

−1.278

Separate website

20.749***

7.565

Follow-on

9.122***

2.605

Residual

−51.560***

−6.611

Intercept

38.26

R2

0.929

p < 0.10; **p < 0.05; ***p < 0.01.

Table 9. Results of the panel data model.

Hypothesis No.

Hypothesis

Model

Outcome

H1

Equity crowdfunding projects with higher investor participation attract more subsequent investors, increasing the funding success rate.

Panel and Instrumental Variable Models Results

HI is positive and significant

H2

In equity crowdfunding, investor engagement’s effect on project success depends on whether the project sponsor modifies the project based on comments.

Panel and Instrumental Variable Models Results

H2 is positive and significant

H3

In equity crowdfunding, a higher frequency of project updates will attract more subsequent investors, which increases the success rate of crowdfunding projects.

Instrumental Variable Model Result

H3 is positive and significant

H4

In equity crowdfunding, posting informative updates on different topics attracts more subsequent investors, increasing crowdfunding projects’ success rate.

Panel and Instrumental Variable Models Results

H4 is positive and significant

6. Discussions, Limitations, and Future Research Direction

6.1. Discussions

This paper investigates the impact of information exchange in equity crowdfunding. Specifically, the effects of comments and updates and their frequency and content on investment decisions by potential investors have been explored. Theoretically, this paper’s findings add to the growing literature on crowdfunding information asymmetry by examining the impact of the value of information derived from investor and promoter behavior on the success of projects. The interactive nature of the funding process reduces the information asymmetry. The principal-agent relationship is affirmed by the project initiators’ role in modifying contents based on investors’ input. Under this model, agency problems exist between the investors and the sponsors due to dependence and conflicting goals [55]. The exchange of information provides a solution to the agency problem, which further helps investors to determine the project quality.

The results show that comments and updates influence investment decisions, confirming the informational value of investors’ and sponsors’ behaviors. The herding effect is evident due to the quality signals investors pick from previous ones for investment decisions. The finding further suggests that investors react negatively to project updates that ask for more investment. An adverse reaction may occur because the project’s progress is displayed in real time on the funding webpage. It further proves that investors are impacted differently when updates are measured regarding contents and that the finer details from other sources signal project quality. It also suggests that obtaining ideas from investors and modifying the project could increase its chances of receiving funding. Also, investor comments’ positive effect on project success aligns with previous study findings that show that customers’ influence on product development impacts project success and reduces agency costs for prospective investors [56].

Moreover, this study proves that different topics that impact investors’ decisions determine the degree of investor participation in the initial offering and later attract subsequent investors. The above is consistent with [57], who point out that crowdfunding is not only about funding but also the quality of the information a project owner provides on the product. In addition, this study uses a panel data model to attenuate the effect of the herd on and correct the endogenous bias of updates and comments. The platform designs the exchange of information function, which project sponsors and investors use to reduce the investment risk associated with herding.

From a managerial perspective, this study provides some practical lessons for platform managers and project initiators in the online equity crowdfunding market. For project initiators, the crowdsourcing behavior of potential investors and the behavioral mechanisms behind it are strategically vital for day-to-day project operations. Specifically, when using these communication tools designed by crowdfunding platforms, project initiators can actively post timely updates and respond to investor comments throughout the project fundraising period. Besides, more comments and updates help attract subsequent potential investors, who view such as a positive project signal. Similarly, project sponsors can adopt meaningful suggestions in the comments to improve their project based on the comments to avoid the negative implications that the content of such an update would bring to subsequent investors. Features such as real-time chat may allow project promoters to exchange information with investors and reduce the number of negative consumer postings on the seller’s page, as found in previous studies.

6.2. Conclusion, Limitations, and Future Research Direction

There are some limitations associated with this study. First, this paper analyzes examples of equity crowdfunding platforms with an “all-or-nothing (AON)” mechanism without considering the “Keep-It-All” (KIA) mechanism. Also, due to data limitations, the study only focuses on the observational learning effect from investors’ sharing and searching behaviors on the Facebook platform. It also excludes the effect of project initiators attraction of potential investors through marketing campaigns on other social media platforms. The model in this paper can continue to be expanded in future research when this type of information is available. Finally, additional attitude-based information could be collected to improve understanding of investment behavior in the equity crowdfunding market. Also, future studies should consider more equity crowdfunding platforms with the ‘‘Keep-It-All’’ model to support the generalizability of the findings. In conclusion, how to design more efficient service operation strategies in equity crowdfunding is a direction that deserves more research.

Conflicts of Interest

The authors declare no conflicts of interest regarding the publication of this paper.

References

[1] Massolution (2015) 2015CF the Crowdfunding Industry Report.
[2] Agrawal, A., Catalini, C. and Goldfarb, A. (2014) Some Simple Economics of Crowdfunding. Innovation Policy and the Economy, 14, 63-97.
https://doi.org/10.1086/674021
[3] Mollick, E. (2014) The Dynamics of Crowdfunding: An Exploratory Study. Journal of Business Venturing, 29, 1-16.
https://doi.org/10.1016/j.jbusvent.2013.06.005
[4] Ahlers, G.K.C., Cumming, D.J., Guenther, C. and Schweizer, D. (2015) Signaling in Equity Crowdfunding. Entrepreneurship: Theory and Practice, 39, 955-980.
[5] Petitjean, M. (2018) What Explains the Success of Reward-Based Crowdfunding Campaigns as They Unfold? Evidence from the French Crowdfunding Platform Kisskissbankbank. Finance Research Letters, 26, 9-14.
https://doi.org/10.1016/j.frl.2017.11.005
[6] Cornelius, P.B. and Gokpinar, B. (2020) The Role of Customer Investor Involvement in Crowdfunding Success. Management Science, 66, 452-472.
https://doi.org/10.1287/mnsc.2018.3211
[7] Belleflamme, P., Lambert, T. and Schwienbacher, A. (2014) Crowdfunding: Tapping the Right Crowd. Journal of Business Venturing, 29, 585-609.
https://doi.org/10.1016/j.jbusvent.2013.07.003
[8] Granados, N., Gupta, A. and Kauffman, R.J. (2010) Research Commentary—Information Transparency in Business-to-Consumer Markets: Concepts, Framework, and Research Agenda. Information Systems Research, 21, 207-226.
https://doi.org/10.1287/isre.1090.0249
[9] Fleming, L. and Sorenson, O. (2016) Financing by and for the Masses: An Introduction to the Special Issue on Crowdfunding. California Management Review, 58, 5-19.
https://doi.org/10.1525/cmr.2016.58.2.5
[10] Mejia, J., Urrea, G. and Pedraza‐Martinez, A.J. (2019) Operational Transparency on Crowdfunding Platforms: Effect on Donations for Emergency Response. Production and Operations Management, 28, 1773-1791.
https://doi.org/10.1111/poms.13014
[11] Xu, A., Yang, X., Rao, H., Fu, W., Huang, S. and Bailey, B.P. (2014) Show Me the Money! An Analysis of Project Updates during Crowdfunding Campaigns. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, Toronto, 26 April-1 May 2014, 591-600.
https://doi.org/10.1145/2556288.2557045
[12] Lukkarinen, A., Teich, J.E., Wallenius, H. and Wallenius, J. (2016) Success Drivers of Online Equity Crowdfunding Campaigns. Decision Support Systems, 87, 26-38.
https://doi.org/10.1016/j.dss.2016.04.006
[13] Chakraborty, S. and Swinney, R. (2016) Signaling to the Crowd: Private Quality Information and Rewards-Based Crowdfunding. SSRN Electronic Journal.
https://doi.org/10.2139/ssrn.2885457
[14] Hu, M., Li, X. and Shi, M. (2015) Product and Pricing Decisions in Crowdfunding. Marketing Science, 34, 331-345.
https://doi.org/10.1287/mksc.2014.0900
[15] Belavina, E., Marinesi, S. and Tsoukalas, G. (2020) Rethinking Crowdfunding Platform Design: Mechanisms to Deter Misconduct and Improve Efficiency. Management Science, 66, 4980-4997.
https://doi.org/10.1287/mnsc.2019.3482
[16] Kuppuswamy, V. and Bayus, B.L. (2018) Crowdfunding Creative Ideas: The Dynamics of Project Backers. In: Cumming, D. and Hornuf, L., Eds., The Economics of Crowdfunding, Springer, 151-182.
https://doi.org/10.1007/978-3-319-66119-3_8
[17] Vismara, S. (2018) Information Cascades among Investors in Equity Crowdfunding. Entrepreneurship Theory and Practice, 42, 467-497.
https://doi.org/10.1111/etap.12261
[18] Fama, E.F. and Jensen, M.C. (1983) Separation of Ownership and Control. The Journal of Law and Economics, 26, 301-325.
https://doi.org/10.1086/467037
[19] Eisenhardt, K.M. (1989) Agency Theory: An Assessment and Review. The Academy of Management Review, 14, 57-74.
https://doi.org/10.2307/258191
[20] Jensen, M.C. and Meckling, W.H. (1976) Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure. Journal of Financial Economics, 3, 305-360.
https://doi.org/10.1016/0304-405x(76)90026-x
[21] Gompers, P.A. (1995) Optimal Investment, Monitoring, and the Staging of Venture Capital. The Journal of Finance, 50, 1461-1489.
https://doi.org/10.1111/j.1540-6261.1995.tb05185.x
[22] Larsson, R. and Bowen, D.E. (1989) Organization and Customer: Managing Design and Coordination of Services. The Academy of Management Review, 14, 213-233.
https://doi.org/10.2307/258417
[23] Kuppuswamy, V. and Bayus, B.L. (2015) A Review of Crowdfunding Research and Findings. SSRN Electronic Journal.
https://doi.org/10.2139/ssrn.2685739
[24] Burtch, G., Ghose, A. and Wattal, S. (2013) An Empirical Examination of the Antecedents and Consequences of Contribution Patterns in Crowd-Funded Markets. Information Systems Research, 24, 499-519.
https://doi.org/10.1287/isre.1120.0468
[25] Mollick, E. and Nanda, R. (2016) Wisdom or Madness? Comparing Crowds with Expert Evaluation in Funding the Arts. Management Science, 62, 1533-1553.
https://doi.org/10.1287/mnsc.2015.2207
[26] von Hippel, E. (1994) “Sticky Information” and the Locus of Problem Solving: Implications for Innovation. Management Science, 40, 429-439.
https://doi.org/10.1287/mnsc.40.4.429
[27] Hsu, D.H. (2004) What Do Entrepreneurs Pay for Venture Capital Affiliation? The Journal of Finance, 59, 1805-1844.
https://doi.org/10.1111/j.1540-6261.2004.00680.x
[28] Griffin, A. and Hauser, J.R. (1993) The Voice of the Customer. Marketing Science, 12, 1-27.
https://doi.org/10.1287/mksc.12.1.1
[29] Bayus, B.L. (2013) Crowdsourcing New Product Ideas over Time: An Analysis of the Dell Ideastorm Community. Management Science, 59, 226-244.
https://doi.org/10.1287/mnsc.1120.1599
[30] Karmarkar, U. (2015) OM Forum—The Service and Information Economy: Research Opportunities. Manufacturing & Service Operations Management, 17, 136-141.
https://doi.org/10.1287/msom.2015.0525
[31] Gremler, D.D. and Gwinner, K.P. (2000) Customer-Employee Rapport in Service Relationships. Journal of Service Research, 3, 82-104.
https://doi.org/10.1177/109467050031006
[32] Priem, R.L., Li, S. and Carr, J.C. (2011) Insights and New Directions from Demand-Side Approaches to Technology Innovation, Entrepreneurship, and Strategic Management Research. Journal of Management, 38, 346-374.
https://doi.org/10.1177/0149206311429614
[33] Lilien, G.L., Morrison, P.D., Searls, K., Sonnack, M. and Hippel, E.V. (2002) Performance Assessment of the Lead User Idea-Generation Process for New Product Development. Management Science, 48, 1042-1059.
https://doi.org/10.1287/mnsc.48.8.1042.171
[34] Nijssen, E.J., Hillebrand, B., de Jong, J.P.J. and Kemp, R.G.M. (2012) Strategic Value Assessment and Explorative Learning Opportunities with Customers. Journal of Product Innovation Management, 29, 91-102.
https://doi.org/10.1111/j.1540-5885.2012.00960.x
[35] Isenberg, D. (2012) The Road to Crowdfunding Hell.
https://hbr.org/2012/04/the-road-to-crowdfunding-hell
[36] Zheng, H., Li, D., Wu, J. and Xu, Y. (2014) The Role of Multidimensional Social Capital in Crowdfunding: A Comparative Study in China and US. Information & Management, 51, 488-496.
https://doi.org/10.1016/j.im.2014.03.003
[37] Yang, M., Ren, Y. and Adomavicius, G. (2019) Understanding User-Generated Content and Customer Engagement on Facebook Business Pages. Information Systems Research, 30, 839-855.
https://doi.org/10.1287/isre.2019.0834
[38] Ho, Y., Wu, J. and Tan, Y. (2017) Disconfirmation Effect on Online Rating Behavior: A Structural Model. Information Systems Research, 28, 626-642.
https://doi.org/10.1287/isre.2017.0694
[39] Xu, J.J. and Chau, M. (2018) Cheap Talk? The Impact of Lender-Borrower Communication on Peer-to-Peer Lending Outcomes. Journal of Management Information Systems, 35, 53-85.
https://doi.org/10.1080/07421222.2018.1440776
[40] Wang, R.Y. and Strong, D.M. (1996) Beyond Accuracy: What Data Quality Means to Data Consumers. Journal of Management Information Systems, 12, 5-33.
https://doi.org/10.1080/07421222.1996.11518099
[41] Blei, D.M. (2012) Probabilistic Topic Models. Communications of the ACM, 55, 77-84.
https://doi.org/10.1145/2133806.2133826
[42] Blei, D.M., Ng, A.Y, and Jordan, M.I. (2003) Latent Dirichlet Allocation. Machine Learning Research, 3, 993-1022.
[43] Büschken, J. and Allenby, G.M. (2016) Sentence-Based Text Analysis for Customer Reviews. Marketing Science, 35, 953-975.
https://doi.org/10.1287/mksc.2016.0993
[44] Humphreys, A. and Wang, R.J. (2017) Automated Text Analysis for Consumer Research. Journal of Consumer Research, 44, 1274-1306.
https://doi.org/10.1093/jcr/ucx104
[45] Wooldridge, J.M. (2003) Econometric Analysis of Cross Section Panel. Journal of Economics, XXIII, 752.
[46] Kickstarter (2014) One Question, Six Creators: What Makes a Good Project Page?
https://www.kickstarter.com/blog/one-question-six-creators-what-makes-a-good-project-page
[47] Colombo, M.G., Franzoni, C. and Rossi-Lamastra, C. (2015) Internal Social Capital and the Attraction of Early Contributions in Crowdfunding. Entrepreneurship Theory and Practice, 39, 75-100.
https://doi.org/10.1111/etap.12118
[48] Parhankangas, A. and Renko, M. (2017) Linguistic Style and Crowdfunding Success among Social and Commercial Entrepreneurs. Journal of Business Venturing, 32, 215-236.
https://doi.org/10.1016/j.jbusvent.2016.11.001
[49] Burtch, G., Ghose, A. and Wattal, S. (2016) Secret Admirers: An Empirical Examination of Information Hiding and Contribution Dynamics in Online Crowdfunding. Information Systems Research, 27, 478-496.
https://doi.org/10.1287/isre.2016.0642
[50] Koch, J.A. and Siering, M. (2015) Crowdfunding Success Factors: The Characteristics of Successfully Funded Projects on Crowdfunding Platforms. Proceedings of the Twenty-Third European Conference on Information Systems (ECIS), Münster, 26-29 May 2015, 1-15.
[51] West, M.A. and Anderson, N.R. (1996) Innovation in Top Management Teams. Journal of Applied Psychology, 81, 680-693.
https://doi.org/10.1037//0021-9010.81.6.680
[52] Kunz, M.M., Bretschneider, U., Erler, M. and Leimeister, J.M. (2016) An Empirical Investigation of Signaling in Reward-Based Crowdfunding. Electronic Commerce Research, 17, 425-461.
https://doi.org/10.1007/s10660-016-9249-0
[53] Müller, R. and Turner, J.R. (2007) Matching the Project Manager’s Leadership Style to Project Type. International Journal of Project Management, 25, 21-32.
https://doi.org/10.1016/j.ijproman.2006.04.003
[54] Pindyck, R.S. and Rubinfeld, D.L. (1991) Econometric Models and Economic Forecasts. 3rd Edition, McGraw-Hill.
[55] Jensen, M.C. and Meckling, W.H. (1976) Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure. Journal of Financial Economics, 3, 305-360.
https://doi.org/10.1016/0304-405x(76)90026-x
[56] Keongtae, K. and Visawanathan, S. (2019) The Experts in the Crowd: The Role of Experienced Investors in a Crowdfunding Market. MIS Quarterly, 43, 347-372.
https://doi.org/10.25300/misq/2019/13758
[57] Belleflame, P., Lambert, T. and Schwienbacher, A. (2010) Crowdfunding: An Industrial Organization Perspective. Digital Business Models: Understanding Strategies, Paris, 25-26 June 2010, 1-30.

Copyright © 2025 by authors and Scientific Research Publishing Inc.

Creative Commons License

This work and the related PDF file are licensed under a Creative Commons Attribution 4.0 International License.