Social Media Marketing for Equity Crowdfunding: Which Posts Trigger Investment Decisions?

Based on 26,883 investment decisions, we examine the inŕuence of social media marketing on crowd participation in equity crowdfunding. We distinguish between different types of informative and persuasive posts on Facebook and Twitter. Informative posts provide investors with information about the crowdfunding campaign; persuasive posts do not, but rather aim to directly inŕuence an investor’s decision-making process. We őnd that both types of posts have a positive impact on the number of investments. However, persuasive posts also increase the amount of an investment if they contain a statement about the previous investment success of the campaign and signal the crowd that they are not investing alone.


Introduction
Signaling theory suggests that if information is asymmetrically distributed between entrepreneurs and investors, the entrepreneur can deliberately send effective signals to investors to ensure them that she is of high quality. To be effective, these signals need to be observable and costly, because otherwise the signal could be easily faked or imitated by low quality competitors. In equity crowdfunding, in which a large group of investors supports a startup over the Internet, information asymmetries are particularly high, since on-site screening and due diligence are rare (Hornuf et al., 2022). Thus, entrepreneurs often send effective signals via their project page on the equity crowdfunding platform, which acts as a gatekeeper and creates a trustworthy signaling environment (Block et al., 2018). However, the gatekeeper function largely falls away when startups seeking equity crowdfunding use social media platforms such as Facebook or Twitter to promote their crowdfunding campaigns.
Assuming that potential investors need credible and diagnostic information before investing in crowdfunding campaigns, the use of social media marketing seems less promising at őrst glance. However, startups also target different types of investors; for example, less sophisticated investors who may primarily use intuition for their investment decisions (Agarwal and Ambrose, 2018, Hornuf et al., 2022, Snow and Rasso, 2017. As the Twitter post by Fraisr above suggest, social media posts promoting crowdfunding campaigns do not always entail effective signals in the spirit of Spence (1973). Accordingly, recent research shows that even non-informative social media contentÐunder certain circum-stancesÐcan actually have a positive inŕuence on investment decisions (Bertrand et al., 2010, Madsen and Niessner, 2019, Tsai and Honka, 2021. We investigate which content startups post on social media platforms and whether these posts have an impact on crowd participation during an equity crowdfunding campaign. mative and a persuasive function (Santilli, 1983, Tsai andHonka, 2021). The informative function is to provide consumers with information about products so that they can make reasoned purchase decisions. The persuasive function, on the other hand, focuses on changing consumer preferences without necessarily providing decision-relevant information. Advertising is then less about an appeal to reason and more about an appeal to intuition (Bertrand et al., 2010).
In crowdfunding practice, startups use informative posts and persuasive posts when promoting their campaigns on social media platforms. The content of informative posts is comparable to the content typically published on crowdfunding platforms, such as the funding limit, the development of the campaign, or basic information about the startup's business model. Examples are: łNew sales expert on board: @fraisr strengthens its sales team with Ex-Daily Deal Key Account Manager!,ž łFundinglimit at Seedmatch increased to e 300.000.ž or łBetterTaxi was voted App of the Week at telefon.de.ž By contrast, persuasive posts do not contain any speciőc information about the crowdfunding campaign or the startup itself. They are primarily aimed at directly inŕuencing the cognitive stages an investor goes through during the decision processes, such as creating awareness, interest, desire, and purchase intentions for the startup, the campaign, and its products (Lavidge andSteiner, 1961, Tsai andHonka, 2021). The speciőc conőgurations of these posts can be quite diverse, ranging from a call to action, for example to invest in the startup or to purchase its products, to posts that are primarily intended to build a relationship with potential and existing investors. Examples are: łDo you want to participate in the success of interactive audio dramas? Then invest in Audiogent now!,ž łNow you can order your best movinary videos as DVD.ž or łLottohelden.de wishes you a Merry #Christmas!ž Often persuasive posts contain so-called heuristic cues, which are intended to further trigger investment decisions (Grewal et al., 1996). Drawing on the information processing literature (e.g., Chaiken, 1980, Petty andCacioppo, 1986), these cues are based on the assumption that consumers are not always motivated to process information systematically, but instead rely on heuristics to make inferences about the attractiveness of advertised objects (Eisenbeiss et al., 2015). Heuristic decision behavior is especially likely when the decision period is limited (Coulter and Roggeveen, 2012), as is naturally the case with crowdfunding campaigns. A long stream of research in the literature has identiőed a number of heuristic cues that impact consumer value perceptions. Among the most popular ones that are particularly frequently used in social media marketing of crowdfunding campaigns are selling history and time limit.
The selling history represents a statement about previous demand, for example in terms of previous shares sold or the number of previous investors. In general, the marketing literature argues that consumers tend to use these cues as a heuristic to assess the quality or value of an object (Dean andLang, 2008, Hellofs andJacobson, 1999). Thus, the greater the previous demand, the more positive the assessment of an object. This is also why őrms like Costco, Amazon, Best Buy, and Circuit City make bestseller information available to consumers. Applied to equity crowdfunding, the presence of a selling history cue in a persuasive post may serve as a form of social validation for the crowdfunding campaign, which signals regular crowd investors an attractive investment opportunity (Coulter and Roggeveen, 2012). Examples of persuasive posts including a selling history cue are: ł422 #microinvestors convinced by @swabr. Now take the last opportunity for #crowdinvesting on @Companisto.ž or łOur #crowdinvesting on @Companisto: 326 people invest e 45,200 and 38 days remain. Secure shares now.ž A time limit represents a special type of a purchase constraint by limiting an offer's availability (Inman et al., 1997). Previous research argues that consumers use unavailability or limited availability as a heuristic cue that the object is scarce (Coulter and Roggeveen, 2012). Scarcity typically induces action by playing upon consumers' fear of missing out an attractive offer (Cialdini, 2007, Coulter andRoggeveen, 2012). Accordingly, previous studies show that time limits accelerate purchases, decrease the likelihood of searching for better alternatives, as well as lead to greater willingness to buy (Aggarwal and Vaidyanathan, 2003, Eisenbeiss et al., 2015, Inman et al., 1997. Moreover, psychological research suggests that łpeople őnd objects and opportunities more attractive to the degree that they are scarce, rare, or dwindling in availabilityž (Cialdini, 1999, p. 92).
Thus, applied to equity crowdfunding, the presence of a time limit in a persuasive post may trigger investment decisions by playing upon investors' fear of missing out an attractive investment opportunity. Examples are: łThe countdown is on -seven days left! Until then, you can still invest in our campaign on Companisto.ž or łToday last chance -Secure your shares now! Here you can őnd the auction.ž In summary, our theoretical considerations suggest that not only informative but also persuasive social media posts have an impact on investment decisions during an equity crowdfunding campaign. In the context of persuasive posts, the presence of selling history and time limit cues could be of particular importance. To investigate whether and to what extent our expectations can be conőrmed empirically will be determined by an empirical analysis of investment decisions on three large German equity crowdfunding platforms.

Data
We analyze 26,883 investment decisions resembling a funding volume of e 18.56 million, which we hand-collected from November 6, 2011, to August 28, 2014, from the equity crowdfunding platforms Companisto, Innovestment, and Seedmatch. In line with Block et al. (2018) and Hornuf and Schwienbacher (2018), we aggregate the investments that a startup received on a single day. We thus get a panel data structure with the duration of the campaign in days as the time dimension, while the cross-sectional dimension relates to the campaigns. We further collected 2,583 Twitter and 1,816 Facebook posts for the startup in our sample during and seven days before the start of the respective campaign. 1 1 For simplicity, we pool the information from Facebook and Twitter posts in our empirical analysis.

Variables
Following Block et al. (2018), we use three distinct but complementary dependent variables to operationalize investment decisions in a relatively broad way: the number of investments per day as a measure of crowd participation (Investments(#)), as well as the funding volume on a given day (Amount(e )) and the average investment amount per day (Avg. amount(e )) to capture the magnitude of an investment.
To investigate which social media posts affect investment decisions, we őrst classify posts into informative or persuasive posts based on their content. For a more nuanced analysis, we further distinguish between different manifestations of informative and persuasive posts. Within the informative posts, we differentiate between the categories Startup, External Certiőcation, and Campaign Development. 2 While all informative posts provide information that investors might consider relevant for decision-making, they evidently differ in the nature of the information. Within the persuasive posts, we distinguish between Product Advertising, Investment Advertising, and Follower Communication. While all persuasive posts aim to inŕuence the investment decisions, the main differences among them are in the way the inŕuence is exercised.  Figure 1 shows the distribution of all posts in our dataset across the six categories. In our dataset, more than 82 % of all posts on social media contain persuasive content. For each of the six categories, we calculate the percentage of posts per startup that fall into this category on a given day.
For example, if a startup publishes ten posts in one day and two of them fall into the investment advertising category, the Investment Advertising variable takes the value of 0.2.
In addition, we document whether startups use selling history and time limit cues in their investment advertising posts. We őnd that around 15.8 % of all investment advertising 2 Note that we also searched for other categories of informative posts as classiőed by Block et al. (2018). However, these categories only play a minor role in social media posts. Therefore, we summarized the categories Team (2.2 %), Business Model (0.3 %), Product Development (0.6 %), Cooperation Projects (1.0 %), New Funding (0.05 %), and Business Development (0.9 %) under the label Startup.

Results
We present our regression results in Table 2. Columns (1) to (2) show the results of őxed-effects negative binomial regressions with the number of investments as dependent variable. We report incident rate ratios, which can be interpreted as multiplicative effects.
Columns (3) to (6) show the results from OLS őxed effects regressions with the funding volume on a given day as well as the average investment amount per day as dependent variable, which have been log-transformed respectively.  (1) to (2) is the number of investments, in columns (3) to (4) the Ln(Amount) of investments, and in columns (5) to (6) the Ln(Avg. amount) of investments in a speciőc campaign and day. The data takes panel data structure. The method of estimation in columns (1) to (2) is the negative binomial őxed effects panel estimator and in columns (3) to (6) the OLS őxed effects panel estimator. In all models, we include the same control variables as in Table 2. Standard errors are shown in parentheses. Signiőcance levels: * p < 0.1, ** p < 0.05, *** p < 0.01.
Our őndings indicate that informative social media posts positively affect the number of One reason for this is that potential investors might unintentionally obtain information about the equity crowdfunding campaign on social media. In contrast, they only receive updates on crowdfunding platforms if they actively search for it on the platforms' websites.
When analyzing the impact of persuasive posts, we őnd that the two content categories, Online Appendix The number of investments made by crowd investors on day t in campaign i.

Amount (e )
The amount in Euros invested by crowd investors on day t in campaign i.

Ln(Amount (e ))
The natural logarithm of the amount in EUR invested by crowd investors on day t in campaign i. Ln(Avg. amount (e )) The natural logarithm of the average amount in EUR invested by crowd investors on day t in campaign i.

Duration
The number of days elapsed from the start until the end of a campaign.
Post funded Dummy variable equal to 1 if the campaign has surpassed the Funding Goal, and 0 otherwise.

Funding goal
The minimum funding goal as deőned by the startup and portal at t = 0.

Active campaigns
The total number of campaigns across three major and one minor German equity crowdfunding portal (Seedmatch, Companisto, Innovestment, and United Equity) accepting investments on day t.

Competing investments
The total number of investments made on day t across all campaigns ran on three major and one minor German equity crowdfunding portal (Seedmatch, Companisto, Innovestment, and United Equity) that where not attracted by campaign i.
The natural logarithm of the total amount of money in EUR invested by the crowd until the previous day in campaign i. # Investments 0→t−1 The total number of investments made by the crowd until the previous day in a particular campaign. #Posts The number of social media posts by the startup on day t in campaign i.

Startup
The share of posts containing information about the startup in relation to the total number of posts in campaign i on day t. These post include information about the entrepreneurial team, the business model, product development, new collaborations, new funding or about the business development.

External Certiőcation
The share of posts including external certiőcation in relation to the total number of posts in campaign i on day t. External certiőcation includes, for example, expert opinions, success stories, news about awards received, patent applications, patent approvals as well as press and media coverage about the start-up.

Campaign Development
The share of posts including information about the crowdfunding campaign and its development in relation to the total number of posts in campaign i on day t.

Product Advertising
The share of posts including product advertising and promotions in relation to the total number of posts in campaign i on day t.

Investment Advertising
The share of posts including investment advertising in relation to the total number of posts in campaign i on day t.

Follower Communication
The share of posts containing other social media content such as invitations for personal meetings, sharing information on related topics or updated proőle pictures in relation to the total number of posts in campaign i on day t.

Time Limit
The share of investment advertising posts that include time limit cues in relation to the total number of investment advertising posts in campaign i on day t. Time limit cues contain information that the purchase of shares is limited in time.

Selling History
The share of investment advertising posts that include selling history cues in relation to the total number of investment advertising posts in campaign i on day t. Selling history cues contain information about previous demand (e.g., shares sold) in the crowdfunding campaign The posts include information about campaign updates or announcements that the funding limit has been changed.
a) łYeah! Thanks to you we have reached the minimum threshold!ž b) łFundinglimit at Seedmatch increased to e 300.000 ś join us and invest in the future of nutsž c) łThe funding threshold for the @OvulaRing has just been reached: http://t.co/F0WPJdXqdr Congratulations to the team!ž Product Advertising: The posts contain product advertising and promotions. a) łBook a tour with us until 31.03. and save up to 50 %! More info on this at...ž b) łNow you can order your best movinary videos as DVD.ž c) łAre you looking for sportswear that combines the special with the functional? Then take a look at the offers from...ž Investment Advertising: The posts include investment advertising without providing speciőc information on the status of the corporate funding campaign. a) łNow is the time! Our crowdfunding is still open for your investment!ž b) łTime to join the other investors! Read more on...ž c) łDo you want to participate in the success of interactive audio dramas? Then invest in Audiogent now!ž Follower Communication: The share of posts containing other social media content such as invitations for personal meetings, sharing information on related topics or updated proőle pictures in relation to the total number of posts in campaign i on day t. a) łIn this sense we wish you a good start into the week and a nice evening! ;)ž b) łThe German women's handball team secured an important victory for the 2014 European Championship qualiőcation! Despite...ž c) łLottohelden.de wishes you a Merry #Christmas!ž Time Limit: Investment advertising post contains information that the purchase of shares is limited in time.
a) Today last chance -Secure your shares now! Here you can őnd the auction: http://t.co/2csGeOVW. Have a nice sunday :) b) The countdown is on -seven days left! Until then, you can still invest in our campaign on Companisto. Never before have so many people participated in a cultural medium via crowdinvesting! c) Only 4 days left...! Take your chance and participate in the success of #OvulaRing via #Crowdfunding.
Selling History: Investment advertising post includes information about previous demand (e.g., shares sold) in the crowdfunding campaign.
a) Crowdfunding already reached e 150,000 ! Thank you for your investment in KERNenergie ! b) Our #crowdinvesting on @Companisto: 326 people invest e 45,200 and 38 days remain. Secure shares now. c) 422 #microinvestors convinced by @swabr. Now take the last opportunity for #crowdinvesting on @Companisto.